document resume - eric · 2014-06-02 · document resume. se 060 701. guo, chorng-jee, ed....

ED 426 847

AUTHORTITLE

INSTITUTIONISSNPUB DATENOTEAVAILABLE FROM

PUB TYPE

JOURNAL CIT

EDRS PRICEDESCRIPTORS

IDENTIFIERS

ABSTRACT

DOCUMENT RESUME

SE 060 701

Guo, Chorng-Jee, Ed.Proceedings of the National Science Council, Republic ofChina. Part D: Mathematics, Science, and TechnologyEducation, 1997.Taiwan National Science Council, Taipei.ISSN-1017-71241997-05-00200p.; Published three times per year.National Science Council, Room 1701, 106 Ho-Ping E. Road,Sec. 2, Taipei 10636, Taiwan, R.O.C. List of contentsavailable at http://www.nsc.gov.tw/scieduCollected Works Proceedings (021) Collected WorksSerials (022)Republic of China. National Science Council. Proceedings.Part D: Mathematics, Science, and Technology Education; v7n1-3 1997MF01/PC08 Plus Postage.Elementary Secondary Education; Foreign Countries; HigherEducation; *Mathematics Education; Preservice TeacherEducation; Professional Development; Science Curriculum;Student Evaluation; *Technology EducationTaiwan

This proceedings is devoted to the publication of researchpapers in mathematics, science, and technology education, coveringdomain/content areas such as learning and the learner, curriculum andmaterials, instruction, assessment and evaluation, history and philosophy ofscience, and teacher preparation and professional development. Papers in thisvolume include: (1) On the Development of Professional Development Programfor Secondary School Science and Mathematics Teachers (Chorng-Jee Guo); (2)

Effects of the Tansui River Education Program on High School Students'Environmental Attitudes and Knowledge (Shun-Mei Wang); (3) TaiwaneseElementary Students' Perceptions of Dinosaurs (Chi-Chin Chin); (4) TheRelationship between Biology Cognitive Preferences and Science Process Skills(Yeong-Jing Cheng, James A. Shymansky, Chiou-Chwen Huang, and Bih-Ju Liaw);(5) Investigation of the Structure and Dimensionality of ILPS Tests Items(Miao-Hsiang Lin, Yeong-Jing Cheng, Song-Ling Mao, Hong-Ming Guo, Tai-ShanFang, Jen-Hong Lin, and Jin-Tun Lin); (6) Linking STS Teacher Development andCertification (Cheng-Hsia Wang); (7) Continually Expanding ContentRepresentations: A Case Study of a Junior High School Biology Teacher(Sheau-Wen Lin); (8) Expert Opinions Concerning a Taxonomic Structure for theCurricular Organization of Biotechnology (Jerming Tseng); (9) A Study of theConcept "Living Organism" and Living Organism Classification in AboriginalChildren (Shih-Huei Chen and Chih-Hsiung Ku); (10) An InternationalInvestigation of Preservice Science Teachers' Pedagogical and Subject MatterKnowledge Structures (Norman G. Lederman and Huey-Por Chang); (11) EmpiricalReview of Unidimensionality Measures for the Item Response Theory(Miao-Hsiang Ling); (12) Development of a Grade Eight Taiwanese PhysicalScience Teacher's Pedagogical Content Knowledge Development (Hsiao-Lin Tuanand Rong-Chen Kaou); (13) Identification of the Essential Elements andDevelopment of a Related Graphic Representation of Basic Concepts in

+++++ ED426847 Has Multi-page SFR---Level=1 +++++Environmental Education in Taiwan (Ju Chou); and (14) A Beginning BiologyTeacher's Professional Development--A Case Study (Jong-Hsiang Yang and I-LinWu). (Author/DDR)

********************************************************************************* Reproductions supplied by EDRS are the best that can be made *

* from the original document. *

********************************************************************************

hffq

D

e apfiT maane admg2

at i emakcs, Science, Wino

ITEDGMT ©ff CHEm

ogf KItioation

0004t24i

@ 0 kiC D0 ¶

IARM6r1 igg?

CONTENTS

On the Development of Professional Development Program for Secondary SchoolScience and Mathematics Teachers (Invited Review Paper)

Chorng-Jee Guo 1

Effects of the Tansui River Education Program on High School Students' EnvironmentalAttitudes and Knowledge

Shun-Mei Wang 10

Taiwanese Elementary Students' Perceptions of Dinosaurs24Chi-Chin Chin

The Relationship between Biology Cognitive Preferences and Science Process SkillsYeong-Jing Cheng, James A. Shymansky, Chiou-Chwen Huang, and Bih-Ju Liaw 38

Investigation of the Structure and Dimensionality of ILPS Tests ItemsMiao-Hsiang Lin, Yeong-Jing Cheng, Song-Ling Mao, Hong-Ming Guo, Tai-Shan Fang,Jen-Hong Lin, and Jin-Tun Lin

PERMISSION TO REPRODUCE AND

DISSEMINATE THIS MATERIAL HAS

B N GRANTED BY

TO THE EDUCATIONALRESOURCES

INFORMATION CENTER (ERIC)

1

46

U.S. DEPARTMENT OF EDUCATIONOffice of Educational Research and Improvement

EDUCATIONAL RESOURCES INFORMATIONCENTER (ERIC)

This document has been reproduced as1ceived from the person or organizationoriginating it.

0 Minor changes have been made toimprove reproduction quality.

Points of view or opinions stated in thisdocument do not necessarily representofficial OERI position or policy.

PUDLOOHED OENAHRNALLV DV

HWEDng McgFD@G, Coam[10VAPEO, TONAN

REPU[31U1© ©NM 2

EDITING: Editorial Board of the National Science CouncilRepublic of China

EDITOR IN CHIEF: Chorng-Jee GuoEDITORS:Cheng-Hsia Wang Ovid J.L. TzengFou-Lai Lin Tze-Li KangYung-Hwa Cheng Yeong-Jing Cheng

ADVISORY BOARD:James BensenBrmidji State University,Bemidji, MN, U.S.A.

Barry J. FraserCurtin University of TechnologyPerth, Australia

Louis M. GomezNorthwestern UniversityEvanston, IL, U.S.A.

Vincent N. LunettaPenn State UniversityUniversity Park, PA, U.S.A.

Tien-Ying LeeJon-Chao HongTak-Wai Chan

Alan J. BishopMonash University,Victoria, Australia

Dorothy L. GabelIndiana UniversityBloomington, IN, U.S.A.

Robert W. HoweOhio State UniversityCamas, WA, U.S.A.

James ShymanskyThe University of IowaIowa City, IA, U.S.A.

Bao-Tyan HwangJong-Hsiang Yang

Reinders DuitUniversity of KielKiel, Germany

James J. GallagherMichigan State UniversityEast Lansing, MI, U.S.A.

Norman G. LedermanOregon State UniversityCorvallis, OR, U.S.A.

Marc SpoeldersUniversity GentGent, Belgium

Proceedings of the National Science Council, Part D: Mathematics, Science, and Technology Educationis published three times per year by the National Science Council, Republic of China

EDITORIAL OFFICE: Room 1701, 106 Ho-Ping E. Rd., Sec. 2, Taipei 10636, Taiwan, R.O.C.Tel. +886 2 737-7594; Fax. +886 2 737-7248; E-mail: [email protected]., [email protected]

SEND SUBSCRIPTION INQUIRIES TO: Room 1701, 106 Ho-Ping E. Rd., Sec. 2, Taipei 10636,Taiwan, R.O.C. Tel. +886 2 737-7594; Fax. +886 2 737-7248; E-mail: [email protected].,[email protected]

SUBSCRIPTION RATES: NT$ 160.00 for one year in R.O.C.; US$8.0 for one year outside the R.O.C.

POSTAGE: (for each copy)Surface mail: free of chargeAir mail: US$ 5.00 America, Europe and Africa

US$ 4.00 Hong Kong and MacaoUS$ 4.00 Asia and Pacific Area

Payment must be made in US dollars by a check payable to National Scinece Council.

The Proceedings (D) is indexed in ERIC.

Printed in the Republic of China.

Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, isgranted by the National Science Council, provided that the fee of $5.00 per article copy is paid directly to CopyrightClearance Center, 222 Rosewood Drive, Danvers MA 01923, USA.

The list of contents is available at: http://www.nsc.gov.tw/sciedu

44"

Information to ContributorsThe Proceedings of the National Science Council include four parts: (A) Physical Science and Engineering, (B) Life Sciences, (C)

Humanities and Social Sciences, and (D) Mathematics, Science, and Technology Education. Part D is devoted to the publication ofresearch papers in Mathematics, Science, and Technology Education, covering domain/content areas such as learning and learner,curriculum and materials, instruction (including computer assisted instruction), assessment and evaluation, history and philosophy ofscience, teacher preparation and professional development, etc. Research papers of similar interests in other disciplines including:environmental, special, health, medical, and information education, etc. may also be submitted. Complete papers describing the resultsof original research that have not been and will not be published elsewhere will be considered for acceptance. Review papers providingsystematic evaluative reviews or interpretations of major substantive or methodological issues in certain disciplines or fields may also besubmitted. Manuscripts are welcome from any country, but must be written in English.

I. SUBMISSION OF MANUSCRIPTS:Manuscripts should be sent in triplicate to:

Proceedings of the NSC, R.O.C., Part DRoom 1701, 17F, 106 Ho-ping East Road, Sec. 2Taipei, Taiwan 10636, Republic of ChinaTEL: (02) 737-7594FAX: (02) 737-7248Email: [email protected]

[email protected]

II. MANUSCRIPT PREPARATION:Manuscripts should be double-spaced, typewritten or printed on only one side of each page, and paginated throughout (including

abstract, figure/table captions, and reference section). Good office photocopies are acceptable. Original complete papers should notexceed thirty pages, review articles should not exceed forty pages. Figures and tables should be kept to a minimum. The compositionof manuscripts should be the following (as applicable): title page, abstract, body of paper, appendices, figure captions, figures, and tables.A page in Chinese at the end should contain the title, author's name, affiliation, mailing address, and abstract.

I. Title page should contain the following information:(1) Title

Title should be brief, specific, and contain informative words. It should not contain formulas or other symbols.(2) Full name of the author(s)(3) Affiliation

The department, institution and country of each author should be furnished. If there are several authors with different affiliations,authors should be paired with their respective institutes by means of superscript symbols after the author's names in this order*, **, ***, etc. Unless indicated, a single star refers to the author to whom correspondence should be addressed.

(4) Mailing address, phone/fax number, and email address of the author to whom correspondence should be addressed.(5) Running head, with no more than 45 letters including spaces.

2. The second page should contain the English abstract and key words. The abstract should be no longer than 500 words. It should statebriefly the reason for the work, its significant results, and conclusions. Three to ten key words should be supplied after the abstract.

3. UnitThe metric or international standard system should be used for recording length, area, volume and weight. When used in conjunctionwith numbers, units should be abbreviated and punctuated (e.g. 3 ml, 5 gm, 7 cm).

4. FootnotesFootnotes should be kept to a minimum and indicated in the text by superscript numbers.

5. NomenclatureInternational standards on nomenclature should be used.

6. AbbreviationsStandard abbreviations for certain substances and units of measure do not need to be defined. Non-standard abbreviations should bespelled out on first usage. Both standard and nonstandard abbreviations should also conform to international standards.

7. Figures and Graphs(1) There should be only one figure per sheet, and its position should be indicated in the text by typing Fig. on a separate line.(2) Figures and graphs should be of the smallest possible printed size that does not hinder clarity.(3) High-contrast computer-generated artwork, drawings in black ink on white paper or tracing paper, or high quality photographic

prints of these are all acceptable.(4) All illustrations should be identified by marking the figure number and author's name on the reverse side. Captions should be

typed on a separate sheet, and not included on the figures.(5) Symbols and lettering should be done in black ink and must be proportional to the size of the illustration. The size of the lettering

should be chosen so that its smallest elements (superscripts or subscripts) will be readable when reduced.(6) Magnification must be indicated with a scale if possible.

(7) If the original drawing (or photograph) is submitted, the second copy of the drawing need not be original; clear photoduplicatedcopies will also be accepted.

8. Tables(1) Tables should be used only if they will present information more effectively than text.(2) Each table should be typed on a separate sheet, briefly captioned, and properly numbered. A table's position in the text should

be indicated by typing Table on a separate line. Explanatory'matter should be in footnotes, not as a part of the legend.(3) Short or abbreviated column heads should be used, and nonstandard abbreviations should be defined in the legend if not defined

in the text.(4) Indicate units of measure in parentheses in the heading of each column. Do not change the unit of measure within a column.

9. AcknowledgmentsAcknowledgments should be stated using a minimum of words and given as a paragraph at the end of the text. Acknowledgmentsto individuals usually precede those for grant support. Names of grant sources should be spelled out.

10. References(1) References should be cited in the text by the last name of the author and year. If there is more than one reference to the same

author(s) and the same year of publication, the references should be differentiated by a, b, etc. e.g., (Jones, 1996); (Jones & Smith,1996); (Jones, Smith, Chang, & Lee, 1992); (Reed, 1994a), (Reed, 1994b); reported by Smith, Jones, & Chen (1995) [firstcitation]; as described by Smith et al. (1995) [subsequent citations].

(2) References should be listed in alphabetical order according to the last name of author(s), patentee(s), or editor(s). Give completeinformation with the title of the paper, report, patent, etc. Do not number references.

(3) References should be arranged as follows:(i) For journals*

Galotti, K. M., & Komatsu, L. K. (1989). Correlates of syllogistic reasoning in middle childhood and early adolescence.Journal of Youth and Adolescence, 18, 85-96.

Johnson-Laird, P. N., Byrne, M. J., & Tabossi, P. (1989). Reasoning by model: The case of multiple quantification.Psychological Review, 96, 658-673.

(ii) For books*. .

Kuhn, T. S. (1970). The structure of scientific revolutions (2nd ed.). Chicago, IL: University of Chicago Press.Gentner, D., & Gentner, D. R. (1983). Flowing waters or teeming crowds: Mental models of electricity. In: D. Gentner

& A.L. Stevens (Eds.), Mental models (Vol. 5, pp. 865-903). Hillsdale, NJ: Lawrence Erlbaum.(iii) For theses*

Brush, S. (1993). A case study of learning chemistry in a college physical science course developed for prospectiveelementary teachers. Unpublished doctoral dissertation, Florida State University, Tallahassee, FL.

(iv) For conference papers*George, S. E., & Grace, J. R. (1980). Entrainment of particles from a pilot scale fluidized bed. Paper presented at the

meeting of the Third International Conference on Fluidization, Henniker, NH.Edmondson, K. M., & Novak, J. D. (1992). Toward an authentic understanding of subject matter. In: S. Hills (Ed.),

Proceedings of the Second International Conference on the History and Philosophy of Science and Science Teaching(pp. 253-263). Kingston, Ontario: Queen's University.

(v) For technical reports*Tiller, F. M., & Leu, W. F. (1984). Solid-liquid separation for liquefied coal industries: Final report for RP- 1411-1 (EPRI

AP-3599). Palo Alto, CA: Electric Power Research Institute.(vi) For patents*

Verschuur, E. (1978). Agglomerating coal slurry particles. U.S. Patent 4, 126, 426.

III. PROOFS AND REPRINTSI. Authors will receive galley-proofs of their papers to check for errors. Proofs should be checked against the manuscript and return

as soon as possible. Keep alternations to a minimum, and if any, mark them on both galley and manuscript.2. Fifty reprints of each paper will be supplied to contributors free of charge. Additional copies may be supplied at the author's own

expense.

IV. PAGE CHARGEThere are no page charge for publication in the Proceedings of the NSC.

V. COPYRIGHTUpon acceptance of a manuscript for publication in the Proceedings of the NSC, the author will be asked to sign an agreement

transferring copyright to the Publisher, who reserves the copyright. No published material may be reproduced or published elsewherewithout the written permission of the Publisher and the author.

Revised December, 1996

Proc. Natl. Sci. Counc. ROC(D)Vol. 7, No. 1, 1997. pp. 1-9

(Invited Review Paper)

On the Development of Professional DevelopmentProgram for Secondary School Science and Mathematics

Teachers

CHORNG-JEE GUO

Graduate Institute of Science EducationNational Changhua University of Education

Changhua, Taiwan, R.O.C.

ABSTRACT

A professional development program was developed for in-service science and mathematicsteachers attending summer school on campus. It was planned, tried out, evaluated, and revised as anaction research. The primary goal of the program was for the participating teachers to obtain bettertheoretical understanding and practical experience about what counted as good teaching and learningof science and mathematics at the secondary school level. The program extended over a period of fouryears, and sufficient time and opportunities were provided for the in-service teachers to practise andreflect on what they had learned in the summer courses. Although contemporary theoretical viewpoints,especially constructivist perspectives, on the nature of science, the nature of science teaching and learning,and on science and mathematics teacher education had great impact on the development of this program,it was also strongly influenced by the experiences and findings that the author obtained from carryingout a series of related research studies in recent years with the collaboration of colleagues and graduateassistants. A review of these research studies is given, providing personal accounts of the interests andthoughts that led to the development of the professional development program. Following a summaryof the rationales for the program's development, some of its key features are described. Based on informalassessment of the results obtained in its implementation, the practicability of the program is discussed.Potential impacts and implications of the program, especially for subsequent studies on science andmathematics teacher education, are also discussed.

Key Words: professional development, science and mathematics teachers, constructionist perspectives

I. Introduction

Research findings from studies of students'conceptions and learning of science (Driver & Erickson,1983; Gilbert & Watts, 1983; Novak, 1988; Osborne& Freyberg, 1985) have led to extensive researchefforts aiming at improving the learning and teachingof science from constructivist perspectives (Hand &Prain, 1995; Fensham, 1988; Tobin, 1993; Ramsden,1988), which emphasize the active roles students playin learning, the importance of students' prior knowl-edge, the existence of higher order learning skills, andthe viewpoint that learning amounts to conceptualchange (Resnnick, 1983; Shuell, 1986; Wittrock, 1985).

These perspectives are of particular significance toattempts at improving the teaching and learning ofscience in Taiwan, because traditionally science in-struction at the secondary school level in Taiwan isfor many students, and teachers alike, a preparationto perform well on tests. It is common to find studentsspending a great deal of time memorizing factualknowledge and solving contrived problems. Manystudents either become bored of learning science orstudy science without meaningful understanding. Thesituation for mathematics education is very much thesame. Constructivism appears to provide a helpfulframework for evaluating current practice and makingreforms in science and mathematics education.

C.J. Guo

In order to improve the quality of science andmathematics instruction at the secondary school lev-els, it is desirable that science and mathematics teach-ers have a good understanding of the main ideas ofconstructivism, be familiar with a variety of teachingstrategies that are consistent with constructivist pointsof view, and above all, be willing and skillful in puttingconstructivist ideas into practice. However, sincemost science and mathematics teachers were preparedunder an educational system which emphasized thetransmission of knowledge and competition amongstudents, they tend to hold beliefs that run againstconstructivist ideas on the teaching and learning ofscience and mathematics. Clearly, it is a great chal-lenge to educate science and mathematics teacherswho can teach successfully in a constructivist manner.It was with this in mind, and for other reasons to beexplained in the following, that the author decided in1992 to initiate an action research project focussed onthe development of a professional developmentprogram for in-service science and mathematics teach-ers.

A summer school offering graduate courses to in-service science and mathematics teachers was alreadyin existence on our campus for a number of years by1992. The in-service teacher education program at thattime consisted of a total of twenty courses offered overfour consecutive summers, covering areas such assubject matter content, theoretical foundations ofscience education, research methods, instructionalstrategies, and so on. Constructivism and other con-temporary viewpoints on philosophy of science, cog-nitive psychology, and various instructional strategiesfor teaching science and mathematics were includedin a few courses mentioned above. Although theintroduction of constructivist perspectives on theteaching and learning of science and mathematics wasreceived by the in-service teachers with some interestduring the summer sessions, most of them tended toteach the same as before when they returned to theirschools. Nevertheless, in recent years an increasingnumber of teachers have expressed strong interest inimproving their teaching in accordance withconstructivist viewpoints.

At the same time, the author had carried out aseries of related studies, including the assessment ofstudents' science process skills involving the use ofscience concepts, students' learning styles in studyingscience, students ' misconceptions and alternativeframeworks, instructional strategies facilitating stu-dents' conceptual change about force & motion, heat& temperature, elements & compounds, and chemicalreactions, etc. As a result of these research efforts,

the author and the collaborating faculty members hadgained valuable experience in diagnosing students'conceptual understanding, in developing teaching strat-egies to facilitate students' conceptual change, andhad explored the feasibility of setting up a new pro-fessional development program for science and math-ematics teachers with constructivist perspectives asguiding principles. We felt strongly that we had theobligation and research foundation to make some majorchanges to the original program, and to develop a newone ,with a consistent set of goals, rationales, andstrategic plans.

In order to show how the goals and rationales ofthe new professional development program were setup in accordance with our previous research findings,a review of three studies which the authors had con-ducted in recent years is given in the following section.It is interesting to note that a number of researchgroups worldwide had moved along the same line. Infact, in a similar layout, the papers included in a recentbook edited by Treagust, Duit, & Fraser (1996) wereorganized into the following three parts: investigatingstudent understanding of science and mathematics,improving curriculum and teaching in science andmathematics, and implementing teacher changes inscience and mathematics.

II. Review of Three Prior Research Studies

Research findings of three prior research projectsare given below. Original papers in English areavailable upon request.

1. Alternative Frameworks of Motion, Force,Heat, and Temperature (Guo, 1993)

The students investigated ranged from grade fiveto grade eight. Researchers developed paper-and-pencil tests used to probe students' conceptual under-standing of a range of phenomena presented in prob-lem situations. Analysis of students' responses yieldedpreliminary results on students' alternative concep-tions. In each study, a number of students (approxi-mately 1.0) having alternative conceptions which wereshared by many students were selected for in-depthinterviews. Students' alternative conceptions deter-mined from both written tests and interviews werecross-checked and synthesized in order to identifystudents' alternative frameworks, which were definedas the underlying belief systems capable of explainingmeaningfully the occurrence of these students' alter-native conceptions. A summary of the results obtainedis given as follows:

2

7

Professional Development Program

A. Alternative Frameworks of Motion and Force forStudents in Grades 5 and 6 in Study A.

(a). Self-Centered Reference Frame

Students tended to take themselves as the refer-ence point for describing the motion of objects, withthe ground as an absolute reference frame.

(b). Position-Oriented Framework

The velocity of an object and the force it expe-rienced were determined by simply considering itsposition relative to the other object.

(c). Motion-Oriented Framework

A moving object was thought to carry with it aforce (or other quantity deemed to be the cause forproducing the effect of motion), and that the faster anobject moved the greater the force.

(d). Animate Framework

Force was thought to be produced only by livingthings that were able to move by themselves, such asanimals and cars. Passive forces such as friction andgravity were often neglected or considered to benonexistent.

B. Alternative Frameworks of Motion for Students inGrades 7 and 8 in Study B.

(a). Motion/Speed-Oriented Framework

Besides ideas similar to those held by the elemen-tary students, the junior high students were found topossess the following alternative conceptions:

(1) A moving object was though to experience aforce acting along the direction of motion. Anobject which was free from external force oughtto be at rest.

(2) An object experiencing a constant external forcewould move with constant velocity on a per-fectly smooth surface. An increase in velocityof an object moving in one dimension must beaccompanied by an increase in the force whichwas acting on the object. A decrease of the latterwas often thought to result in a decrease in thespeed of the object.

(3) In the case of an elastic collision, a billiard ballwas thought to experience no force upon impactwith a solid wall, because the speed of the ballremained unchanged.

(b). Weight-Oriented Framework

For an object initially at rest to start moving, itmust be subjected to a force which is greater than the

A

weight of the object.

C. Alternative Frameworks of Heat and Temperaturefor Students in Grade 6 in Study C.

(a). Substantive Framework

For many students, heat was thought of as a sortof gas, foam, or vapor, there being hot and cold onesin a certain proportion for a body at a given tempera-ture. The conduction of heat was thought of as therising or diffusion of the hot stuff (molecules, orparticles) from one end of a rod to the other, or fromone thing to another.

(b). Equal/Undifferentiated Framework

Heat and temperature, often referred to as "hot-ness," appeared to be the same for many students. Thetemperature of a body was considered to be equal tothe "heat" it contained.

D. Alternative Frameworks of Heat and Temperaturefor Students in Grades 7 and 8 in Study D.

(a). Substantive Framework

Alternative conceptions related to the substan-tive framework included:

(1) Heat was a sort of gas, like water vapor orsmoke, and consisted of hot and cold gases.

(2) Matter conducted heat by means of "diffusion"of temperature, heat, or molecules.

(b). Equal/Proportional/Undifferentiated Framework

In addition to the alternative conceptions held byelementary students, junior high students were foundto have the following alternative conceptions:

(1) Heat was a kind of force or energy, which wasproduced by temperature and was therefore pro-portional to it.

(2) The temperature of an object increased withcontinued heating, even during phase transition.

(3) The heat "absorbed" or "released" by an objectwas considered to be influenced by factors in-cluding mass and temperature change only.

(4) Heat was confused with or used interchangeablywith temperature.

(c). Characteristic Temperature Framework

Temperature of an object was considered as itsintrinsic property, with the effects of environmentneglected.

(d). Passive Role Framework

Such a framework consisted of the idea that heat"absorbed" by water was stored within the "density"

3

C.J. Guo

of water (meaning the vacant space in between themolecules), leaving water molecules unaffected. Or,as boiling water was kept heated, water temperatureremained at 100 °C and the absorbed heat would"diffuse" into the air, leaving molecules in the waterunaffected.

2. Improving Science Teaching and Learning:A Holistic Study Based on ConstructivistPerspectives (Guo, Chiang, Tuan, & Chang,1993)

The purpose of this study was to initiate an effortaimied at improving the teaching and learning of sci-ence at junior high school level, based on the view-point that teachers and students construct meaning inthe process of classroom instruction. From the stan-dard textbook currently used in Taiwan, four instruc-tional units including heat and temperature, motionand force, elements and compounds, and chemicalreactions were selected as topics for study. Prevalentalternative conceptions on these topics were identifiedbefore suitable instructional plans from a constructivistview were developed, tried out, and evaluated. Theinvolvement of science teachers in developing the testinstruments for probing students' alternative concep-tions and in developing suitable lesson plans was builtinto this study, so that both teachers' development ofa constructivist view of science teaching and students'concept learning results could be studied in a holisticway. Major findings include the following:

(1) Although the science teachers were willing toaccept constructivist views of science teachingand learning, their preconceptions about sci-ence teaching were resistant to change. Forinstance, the belief that a clear presentation ofwell organized materials to the students was ofutmost importance was still held by the scienceteachers at the end of the first year after theyhad participated in the project.

(2) The science teachers showed great interest inparticipating in the research activities. Thedevelopment of test items aimed at probingstudents' understanding of science concepts wasa valuable experience for them.

(3) Students' alternative conceptions identified inthis study on motion and force, heat and tem-perature, and the molecular model of matterwere very similar to those reported in the lit-erature. It was noted that many students whohad received formal instruction on the selectedtopics still held alternative conceptions similarto those of students who had not yet studied the

same topics.(4) It appeared difficult in the beginning for science

teachers to develop their own instructional plansaimed at bringing about conceptual change instudents.

3. Practicability of Constructivist Approachesin Science Teaching - A Case Study of SixScience Teachers in Taiwan (Guo, Chiang,Chen, & Wang, 1995)

This study was an integral part of the precedingproject as it entered into the third (final) year. Effortswere made to answer the following questions concern-ing the science teachers participating in the researchteam:

(1) To what extent do science teachers have a graspof the basic ideas of constructivism and theirimplications for science learning and teaching.How do the science teachers teach accordingly?

(2) What kind of experience would be helpful forscience teachers to develop their belief inconstructivism and skill in teaching science usingstrategies which are consistent withconstructivist perspectives?

(3) What are the key factors affecting the practi-cability and effectiveness of including con-structivist ideas in science teaching, particu-larly considering the existing educational andsocial-economic environment in Taiwan?

The study was carried out with the cooperationof four science educators, two junior faculty members,two graduate students, and six science teachers. Themembers of the research team were equally dividedboth in terms of sex and fields of specialty physicsand chemistry. The four science educators, havingmany years of research experience in science educa-tion, were also actively involved in the in-servicescience teacher education program. Six experiencedlocal science teachers were selected to participate inthis study. They had been exposed to constructivismthrough a series of lectures in the in-service scienceteacher education program. Four of them had partici-pated in our research group for two years, while theother two had just joined us at the start of this study.

A qualitative case study approach was used togather data from classroom observations, interviews,and relevant documents. The data obtained for eachscience teacher was organized to form a case record,which were then compared, consolidated, and inter-preted in order to come up with the following generalconclusions.

(1) The science teachers tended to take constructivist

4

9


ideas towards teaching and learning mainly asnew teaching approaches towards concept learn-ing and conceptual change, and focussed onideas which were consistent with their ownteaching beliefs, styles, and practices. Althoughthey were able to come up with plans andactivities which were conducive to constructivistteaching and thus providing more opportunitiesfor students to think, question, and discuss, theirmajor concerns towards and beliefs in scienceteaching were essentially the same as before.To be sure, some changes in teaching behaviorwere observed. For instance, they tended to askmore questions which were important for a mean-ingful understanding of some key concepts, andoften requested students share examples fromdaily experiences in order to illustrate certainconcepts and principles under study. Neverthe-less, teachers' desire to cover scheduled topicsand students' desire to do well in written ex-aminations often forced them to return to theuse of traditional, direct teaching methods, whichthey felt to be straightforward and less timeconsuming.

(2) It was noted in this study that regularly sched-uled meetings and practical experience in con-ceptual change teaching strategies, includingprobing students' misconceptions and design-ing instructional plans, were of great signifi-cance to the implementation of a constructivistteaching approach by science teachers. Inaddition to providing opportunities for self andgroup reflection, they seemed to provide avehicle for the science teachers to develop abetter understanding of constructivism, and tobuild confidence in its relevance to scienceeducation.

(3) It was also noted that there were other importantfactors influencing the practicability and effec-tiveness of constructivist teaching, such as stu-dents' study skills and learning strategies,appropriate ways to evaluate students' achieve-ment, and school and parents' support for andconfidence in the science teachers. The scienceteachers in this study were experiencedscience teachers who knew how to guide stu-dents to study differently from what they wereused to, and to convince students, parents, andschool administrators that the long-termeffects of a constructivist teaching approachought to be very beneficial. Nevertheless, theensuing challenges and pressure were verynoticeable.

5

III. Program Development Ration-ales and Implementation Strat-egies

Since professional development is tied to salaryraises and retirement pensions, the in-service teachersattending the summer school on campus were admittedthrough a competitive selection process favoring ex-perienced teachers. Typically, they had at least fiveyears' teaching experience and were selected eitherbecause of involvement in some sort of administrativework or because their students performed well inacademic activities such as science exhibitions, sci-ence contests, and so on. They tended to be highlymotivated teachers, and were usually considered to begood teachers in their schools. Their teaching view-points and techniques are, however, very traditionaland monotonous. In view of recent calls worldwidefor reform of science and mathematics education byemphasizing constructivist approaches to teaching andlearning, the interaction of science/technology/soci-ety, improving students' scientific literacy, etc., it isvery desirable that the in-service teacher's' viewpointson the teaching and learning of science and mathemat-ics should be widened and brought up to date. There-fore, the primary goal of the professional developmentprogram developed in this study was to provide learn-ing environments and activities in such a way that theattending teachers would be able to construct for them-selves a better understanding of contructivist view-points on science and mathematics teaching and learn-ing, and would be able to put them into praciice intheir own classes. Furthermore, they were expectedto play leadership roles in encouraging and workingwith their peers to improve science and mathematicsteaching at the classroom and school levels.

Based on previous research findings, our ration-ales for developing a new professional developmentprogram to obtain the above goals were as follows:

(1) Teachers' conceptual understanding of andbelief in what counts as good science and math-ematics teaching are hard to change. It takestime and effort for this to occur.

(2) There are cognitive and affective/social factorswhich must be taken into account when oneattempts to induce change in teachers' beliefs,conceptual frameworks, teaching behaviors,etc. The former has to do with the teachers'own evaluation of the intelligibility, plausibil-ity, and fruitfulness of constructivism, as sug-gested by the conceptual change model of Posner,Strike, Hewson, and Gertzog (1982). While thelatter involves motivation, self-efficacy, peer

10

C.J. Guo

pressure, societal expectations, administrators'support, and so on.

(3) In order for the teachers to be able to teach ina constructivist manner, they must be taught inthe same way. That is, ideally, teachers mustbe taught the way they are expected to teach.

(4) In order to have a good grasp of the essence ofconstructivist approaches to teaching, teachersneed to put what they have learned intopractice with careful planning and constantreflection. Simply listening to a few lectureson constructivism is not going to cause anynoticeable change. Reflection in action andknowledge from action are keys to a betterunderstanding of constructivism.

(5) In addition to content and pedagogical knowl-edge, it is very important for teachers to be ableto assess students' prior knowledge, alternativeconceptions, development in conceptual under-standing, science process skills, and learningstyles. Teachers should be able to help studentsmeaningfully learn science and mathematics,based on the information gathered through theabove assessments. In short, a teacher's peda-gogical content knowledge is important.

Consistent with the above goals and rationales,a number of strategies were adopted in the implemen-tation of the program.

(1) The program consisted of a total of twenty sum-mer courses, which were designed to help thein-service teachers carry out action researchaiming at improving science or mathematicsinstruction in their own classes when they re-turned to school to teach. This should lead toputting theory into practice, and a spreading ofprofessional development efforts from summersessions to all year round. The teachers wereencouraged to carry out action researchcollaboratively with their peers from the sum-mer school, or with other teachers in their ownschools.

(2) From choosing a research topic to the publish-ing of research results, the entire process wasdivided into four stages, with summer coursesproviding needed guidance in each stage overa period of four years. For instance, coursesin the first summer focused on the nature ofscience, the goals of science education, con-structivism, conceptual change teaching strat-egies, and so on. Courses in the second summerfocused on research designs and data collectionmethods (including assessment of students'learning styles, misconceptions, and skills in

problem solving), teaching techniques, class-room management, and educational technology.Courses in the third year included analyses ofboth quantitative and qualitative data, the writ-ing of research reports, and so on. Other coursessuch as educational administration, STS, inte-grated science, and so on, were included in thefourth summer.

(3) At least two-thirds of the summer courses weretaught by faculty members participating in theresearch group. These courses were taught ina manner consistent with constructivist ideasabout teaching and learning. Various instruc-tional strategies such as learning cycles, Gowin' sV-diagrams, problem-centered activities, andso on were used. Collaborative projects andgroup discussions were commonplace, and thein-service teachers were given the shared rightto decide course contents, evaluation methods,and so on. Some courses were taught by a teamof two to four professors, providing a workingmodel of collaboration.

(4) An action research project was built into thedevelopment of the professional developmentprogram, in order to evaluate its practicality andeffectiveness. A number of in-service teachers,typically four to eight in each year, were chosento participate in our research group. Theymaintained close contact with the universityfaculty. Their classroom teaching and researchactivities were studied by other members in theresearch group.

(5) Various kinds of supports were provided to thein-service teachers, particularly to those par-ticipating in the research group, so that theywould be able to carry out their own researchefforts in improving teaching and learning, whileat the same time learn to assume leadershiproles in science and mathematics educationreforms. This support included correspondenceby phone and mail, visits to their schools andprincipals, helping them to hold workshops onconstructivist ideas and practices in teachingscience and mathematics, and circulation ofbulletins on constructivism and constructivistapproaches in teaching. The use of the Internethas also become popular lately.

The development of this professional develop-ment program was based on our understanding ofconstructivist approaches to the teaching and learningof science and mathematics, and on our researchexperiences obtained from previous related studies. Itwas developed from a vision of science and mathemat-

6


ics education for the 21st century, and an understand-ing of existing educational and socio-economic con-ditions in Taiwan. Although it was not modelled afterany particular program, the underlying themes seemedto be close to those described by Doyle (1990) as the"innovator paradigm" and the "reflective professionalparadigm". With reference to these themes, teachereducation should be a source of renewal and innova-tion for schools, and should foster the reflective ca-pacities of observation, analysis, interpretation, anddecision making. The practicality and effectivenessof the program, to be discussed in the following section,should be viewed with such an orientation too.

IV. Practicality and Effectiveness ofthe ProgramAs Doyle (1990) pointed out that, with the the-

matic shift in teacher education research, quality controlshould be based not on a system of external controlbut on the available knowledge about content andabout practices and their implementation in the class-room. Instead of asking whether a practice appearsto be effective according to external criteria used, theaccountability question becomes, is the appropriateknowledge being used in this situation? As an actionresearch, qualitative data including video tapes ofclassroom observations, interviews, teachers' instruc-tional plans, students' journals, and so on were col-lected, analyzed, and cross-checked in an attempt toanswer questions about the practicability and effec-tiveness of the program. As the in-service teacherswho first entered this program in 1994 are still in theirthird year, a complete evaluation of the entire programwill not be available until the end of 1998. However,from informal assessments of the results obtained .sofar, we have noticed that most of the teachers wereable to do what they were expected to with varyingdegrees of success.

Evaluation of the practicability and effectivenessof the program in the context of current educationalpractices in Taiwan is greatly influenced by the factthat competition to enter better senior high schools anduniversities has become a major public concern. Al-though measures have been taken by the Ministry ofEducation to lessen such competition and its negativeeffects on education, no real progress has beenachieved so far. Many schools still secretly dividetheir students into classes consisting of students con-sidered to be of different abilities, .paying greaterattention to teaching the brighter students who appearto have better chances of doing well in the entranceexaminations. There were many anecdotes told by the

13 A

in-service teachers about their successes in teachingclasses consisting of whatused to be considered lowerability student, using constructivist approaches. Stu-dents who were proviously deemed as stupid, naughty,or reckless were found to participate and often do wellin discussions and hands-on activities. The realchallenge lies, however, in teaching classes withhigher ability students. Parents and school adminis-trators were reluctant for teachers to try anything newon them, because they were worried that it would leadto harmful results on students' achievement tests. Thispressure was so high that many teachers dared notdeviate from their current teaching practices in theseclasses. However, approximately 10% of the in-ser-vice teachers were convinced of the merits of aconstructivist approache after taking the first summercourses and were confident enough to put them intopractice during the following school year. Theybought back encouraging experiences to share withtheir classmates in the summer of the next year. Withsuch input and further summer courses designed toenhance better understanding of constructivism and avariety of instructional skills, most teachers becamemotivated and confident enough to make certainchanges in their teaching following the second sum-mer. In a sense, our program appeared to be practicaland effective. Nevertheless, some of the in-serviceteachers still found themselves facing the dilemma ofwhether to teach in a constructivist manner consis-tently or just to do it once in a while as a supplementto traditional teaching method focusing mostly onmemorization and solving problems designed forpaper-and-pencil tests. It would be interesting to seehow teachers' decisions on such matters evolve astime goes on, and what are the important factorsinfluencing their decision making. These are, in fact,the research questions of the thesis of a Ph.D. can-didate surpervised by the author.

V. Concluding Remarks

The goals, rationales, implementation strategies,and practicality of a professional development pro-gram developed for in-service science and mathemat-ics teachers were described in the above sections. Themain purpose was to summarize the thoughts andefforts that went into developing the program, basedon the author' s understanding of contemporary theo-retical perspectives on science education, the per-ceived educational needs of science and mathematicsteaching in Taiwan, and the research findings of re-lated studies headed by the author in recent years. Itwas meant to serve as an illustration of how one might

7

12

C.J. Guo

combine research with practice in an effort to makea contribution to the reform of science and mathemat-ics education. In this respect, it appeared to be ex-cusable to not include comprehensive reviews ofrelated literature and more detailed information of theprogram in this article. It must also be pointed outthat although constructivism was refered to as a guidingprinciple in developing the program, it was by nomeans considered as a panacea for solving all prob-lems in science education. In the professional devel-opment program, the in-service science teachers wereprovided on opportunity to reconstruct their view-points of the nature of science and mathematics, thenature of the learners, and the nature of the teachingof science and mathematics. They were not forced intobelieving or doing anything that did not make senseto them.

Although determining the effectiveness of theprogram and its potential impacts on the in-serviceteachers and their students must await further studiesin the near future, it is perhaps worthwhile mentioningsome of its implications for other reform efforts inscience and mathematics education. In carrying outthe development of the professional developmentprogram as an action research, the author and thecollaborating faculty members made frequent contactwith the in-service teachers, their school principals,and other eddeaticnal administrators, and were ableto bring to their attention that meaningful understand-ing and scientific literacy are important goals, and thatconstructiVist teaching approaches have much to offerin reaching these goals. In addition, a few principalswere so impressed with the improved performance oftheir science or mathematics teachers from the pro-gram that they wished we could also help their scienceand mathematics teachers, and even those in othersubject matter areas to teach in a constructivist man-ner. This is certainly something worthwhile for ourresearch agenda in the near future.

Another potential impact of the program has todo with the preparation of presevice science andmathematics teachers. With the advent of the 21stcentury, calls for reform of education have been putforward all over the world in recent years. Taiwanis no exception. Some steps have been taken by theMinistry of Education in an effort to bring aboutchanges in existing educational practices. For in-stance, with the recent legislation regarding the diver-sification of teacher education programs, the respon-sibility for the preparation of teachers will no longerlie on normal colleges and normal universities. In-stead, colleges and universities offering teacher edu-cation programs will be eligible to prepare prospective

teachers, who will be engaged in a practicum for oneyear before they are hired by schools and local edu-cational authorities. It is obvious that the in-servicescience and mathematics tearchers participating in theprofessional development program will be able to playvery important roles in serving as mentors for thepreservice teachers in their practicums, and helpingin the selection of qualified new teachers at the localschool level. In fact, concerns in this regard had ledthe author to carry out a project entitled "A Study onEducational Practicum and Certification ofJunior HighMathematics and Science Teachers" beginning inApril 1995. This consisted of several sub-projectsheaded by collaborating colleagues, with researchquestions focussed on the description of the personaltraits of student teachers and their environmental fac-tors, the development of models of practice teaching,the identification of desired teacher competencies andthe observation of student teachers' actual perfor-mance in classroom situations, etc. Three commonthemes, namely constructivism, pedagogical contentknowledge, and portfolio development provide unify-ing threads which run through all the sub-projects.Evidently, there are valuable experiences and researchfindings from the professional development programproject that these studies can build on.

Aknowledgement

The works reported were supported, in part, by the NationalScience Council under grant numbers NSC 84-2513-5018-004, NSC85-2517-S-018-001, and NSC 86-2511-S-018-012-SP. Thanks tothe collaborating colleagues, graduate assistants, and inserviceteachers participated in the research team, particularly to Prof. Wu-Shiung Chiang for co-directing these research projects.

References

Doyle, W. (1990). Themes in teacher education research. In: W.R.Houston (Ed.), Handbook of research in teacher education.New York, NY: Macmillan.

Driver, R., & Erickson, G. (1983). Theories in action: Sometheoretical and empirical issues in the study of students' con-ceptual frameworks in science. Studies in Science Education,10, 37-60.

Fensham, P. (Ed.). (1988). Development and dilemmas in scienceeducation. London: Falmer Press.

Gilbert, J.K., & Watts, D.M. (1983). Concepts, misconceptions andalternative conceptions: Changing perspectives in scienceeducation. Studies in Science Education, 10, 61-98.

Guo, C.J., Chiang, W.H., Tuan, H.L., & Chang, H.P. (1993).Improving science teaching and learning: A holistic studybased on constructivist perspectives. Paper presented in theposter session at the annual meeting of the National Associationfor Research in Science Teaching, Atlanta, GA.

Guo, C.J. (1993). Alternative frameworks of motion, force, heat,

8

t3


and temperature: A summary of studies for students in Taiwan.Paper presented at the annual meeting of the National Asso-ciation for Research in Science Teaching, Atlanta, GA.

Guo, C.J., Chiang, W.H., Chen, M.L., & Wang, C.Y. (1995). Prac-ticability of constructivist approaches in science teaching - Acase Study of six science teachers in Taiwan. Paper presentedat the annual meeting of the National Association for Researchin Science Teaching, San Francisco, CA.

Hand, B., & Prain, V. (Eds.). (1995). Teaching and learning inscience The constructivist classroom. Sydney: HarcourtBrace.

Novak, J.D. (1988). Learning science and the science of learning.Studies in Science Education, 15, 77-101.

OsborneR., & Freyberg, P. (Eds.). (1985). Learning in science:The implications of children's science. London: HeinemannPress.

Posner, G.J., Strike, K.A., Hewson, P.W., & Gertzog, W.A. (1982).Accommodation of a scientific conception: Toward a theory of

conceptual change. Science Education, 66(2), 211-228.Ramsden, P. (Ed.). (1988). Improving learning. London: Kogan

Page Ltd.Resnick, L.B. (1983). Toward a cognitive theory of instruction.

In: S.G. Paris, G.M. Olson, & H.W. Stevenson (Eds.), Learningand motivation in the classroom. Hillsdale, NJ: LawrenceErlbaum ASsociates.

Shuell, T.J. (1986). Cognitive conceptions of learning. Review ofEducational Research, 56(4), 411-436.

Tobin, K. (Ed.). (1993). The practice of constructivism in scienceeducation. Washington, D.C.: AAAS Press.

Treagust, D.F., Duit, R., & Fraser, B.J. (1996). Improving teachingand learning in science and mathematics. New York, NY:Teachers College Press.

Wittrock, M.C. (1985). Learning science by generating new con-ceptions from old ideas. In: L.H.T. West, & A.L. Pines (Eds.),Cognitive structure and conceptual change (pp. 259-266). NewYork, NY: Academic Press.

AM114-44101tAMA-MAO*W4RZHO

MAZfMRtff44-WWW-AN

N

21K-30K4*+*0011rOMMAO*VnlitN,AVAVAAAFiAM,W11-04t-faOtT.AIFInkistilOVR-614111WiTnrPT,VAft*A4V4in*w_Lning0-IttA0*42111M4MINEMM*MR8AltrOM5)-n41111RM-61-444*N_L-Mffi*nwfMA-,TAntArPRMNUAgi,t4-44,R&A.**fInttA MR.131MI*4#4t4VEM#M4fi*IYAKW,Itwft*MAKW,URtwO5Mn#111W,MnwitA**ERAHWAALE***XIMMW9MXtMg0fullIMAAZ_L0 MAAffAIIIIAX,UMkttWitMAIX*WnRO*nwt-gM0 Attalt*Vn42MtZ*,#114 AdVICAMMMONM,VAAVMAffriq.A* itit-4-OulT***ARatnt:

14


Effects of the Tansui River Education Program on HighSchool Students' Environmental Attitudes and

Knowledge

SHUN-MEI WANG

Institute of Environmental EducationNational Taiwan Normal University

Taipei, Taiwan, R.O.C.

(Received April 9, 1996; Accepted September 5, 1996)

ABSTRACT

Like many other places on the Earth, most rivers in Taiwan are polluted. Educational programsare needed to raise the public's concern about rivers and motivate them to take action. The Tansui RiverEducation Program (TREP) was designed for secondary students in the Tansui River watershed of Taiwan.The pilot TREP was conducted in two high schools to evaluate the program's effects on attitudes andknowledge. The experimental design included control and experimental groups, pre- and post-tests anda questionnaire. A paired t-test and a one-way ANOVA with a 0.05 level of significance were usedto analyze the data. Due to an inflexible regular curriculum, the two-day pilot program was conductedoutside of class during weekends and holidays. The control group did not receive any instruction andonly answered the questionnaire. Major findings of the research were: (1) TREP significantly increasedparticipants' awareness of the river; (2) TREP significantly increased their sense of empowerment forsaving the river; (3) TREP significantly increased their intention to take actions related to "advocacy"and "school environmental protection"; (4) TREP significantly increased their feeling of responsibilityfor the school environment in terms of planning and decision making; (5) TREP significantly increasedtheir knowledge of water quality. In light of the research results, the author suggests that the Ministryof Education in Taiwan should consider adding TREP to the national curriculum.

Key Words: water monitoring, river education program, evaluation, environmental attitudes, knowledgeof water quality.

I. Introduction

The environment is a type of social issue; thisissue is related to human beings' values, beliefs, andawareness about the environment and environmentalprotection (Stapp, 1983). About 30 years ago, peoplebegan to be aware of the environmental crisis andenvironmental protection movements began in west-ern countries (Swan, 1975) and Japan (Song, 1989).Gradually people began to understand the importanceof environmental education in addressing the rootcauses of environmental issues indifference towardthe environment and lack of knowledge about theenvironment.

The issue of river water quality is one of themost critical environmental issues around the world,particularly in Taiwan. The middle and lower reaches

7rg

of most rivers in Taiwan are moderately to seriouslypolluted (Environmental Protection Administration,1989). Although technical and political efforts to savethese rivers, such as building sewage systems and lawenforcement to prohibit illegal dumping have beenundertaken, river education programs are needed toincrease the general public's concern for the rivers andmotivate them to be involved in the tasks of improvingthe rivers' water quality. This research consisted ofthe design of the Tansui River Education Program(TREP) and its evaluation.

H. Tansui River Education Program(TREP)

TREP focused on the Tansui River watershed,which is the largest river system in the north of

10

Effects of the River Program on Students

Taiwan. Its three branches, the Keelung River, theSinten River, and the Dahan River drain a watershedthat includes 27 cities and a population of more than4.6 million (Fig. 1). Various land uses have seriouslypolluted the Tansui River and the water quality isusually very poor (BOD> 18 ppm) in its middle andlower reaches.

The overall goal of TREP was to motivate stu-dents to be concerned about this watershed and takeactions to improve river quality. In order fo-r this tooccur, the following four subgoals of TREP were set:

(1) To increase participants' awareness of currentstatus of the Tansui River

(2) To promote participants' environmental atti-tudes and empowerment

(3) To enhance participants' knowledge of riverwater quality

(4) To increase participants' skills in water moni-toring and problem solving

These subgoals match the aims, characteristics

Fig. 1. Program area: The 21 main rivers in Taiwan and theirpollution status, three main branches of the TansuiRiver and the location of Keelung High School andSonsan High School.

and hierarchical objectives of global environmentaleducation, including awareness, knowledge, attitudes,skills, and participation. (UNESCO, 1980) In addi-tion, Hines' meta-analysis (Hines, Hungerford, &Tomera, 1986) also pointed out that attitudes, empow-erment, knowledge and skills are important variablesin building an environmentally responsible citizen. Inthis research, the effect of TREP was examined byassessing the students' awareness, attitudes and knowl-edge.

TREP consists of five sections (Fig. 2): (1)Students shared their experiences and expectations ofthe Tansui River within groups involved in the pro-gram. (2) Students learned about the Tansui River andits problems through a slide presentation in order toget an overview of the river. (3) Students monitoredriver water with nine tests and investigated land usesalong the Tansui River near their school in order tounderstand the interaction between human activitiesand river water quality and build skills in using watermonitoring testing kits and analyzing water qualitydata. (4) Students read a story about citizens success-fully taking action to stop illegal fishing in a smallvillage of Taiwan in order to learn about citizens'environmental responsibilities and problem solvingstrategies. (5) Students undertook collective actionsin their school in order to increase public awarenessof the Tansui River's problems.

The program adopted the interdisciplinary andcommunity problem solving approaches used in theRouge Project (Bull et al., 1988). The program com-bined natural science investigation and social sciencediscussion to understand the big picture of water qualityand how it is related to social issues. The environ-mental issue focused on was river water quality near

Experience Sharing

Silde Show

Water Monitoring

Case Study

Action Taking

ProblemInvestigation

ProblemSolving

Fig. 2. The five sections of the Tansui River EducationProgram.

11

16

S.M. Wang

the schools. Students collected information on theriver, learned citizen action strategies and processesand gained direct experience solving real world issues.

TREP adopted two attitude change approacheschanging attitudes by giving information and letting

students learn by doing. Festinger' s cognitive disso-nance theory (Festinger, 1957) holds that a person' scognition, attitude and behavior must usually be con-sistent in order for the person' s psyche to be in bal-ance. He asserted that individuals cannot tolerate apsychological state of dissonance. The state of dis-sonance causes an uncomfortable tension and drivespeople to search for information, change their attitude,or change behavior to reduce or avoid an increase incognitive dissonance.

The slide show and case study sections of TREPprovided information concerning sense of responsibil-ity for the environment and the impact of some com-mon behavior on rivers. When students learn newmessages or truths concerning the target, at the sametime they learn values and attitudes toward the target(Koballa, 1988). However, the process of informing,assimilating, and confirming is conducted by thelearner's existing cognitive system (Phillips, 1981;Kaplan and Kaplan, 1982). Inculcation and valueclarification approaches were applied in the slide showand case study sections to build students' attitudes andknowledge.

Another effective learning approach is first-handexperience. Direct experience using more than oneof the senses can build more comprehensive conceptsand help the learner retain the concepts longer thanlearning by symbols only (Fazio and Zanna, 1981).Engleson (1985) views direct purposeful experiencesin the field as able to benefit students' awareness ofthe environment and their environmental responsibil-ity. TREP provided direct experience in the watermonitoring and action-taking sections, which createdanother opportunity for attitude change.

Another approach changing subjects' behaviorto lead to a change in their attitude was also appliedin TREP. According to Festinger' s cognitive disso-nance theory (Festinger, 1957), people usually changetheir attitude toward a target in public decision makingand compliant behaviors. Following a decision, thereis a process of actively seeking out information whichis consonant with the decision, strengthening confi-dence in the decision and increasing attitudes in favorof the alternative chosen (Festinger, 1957). Zimbardo,Ebbbesen, & Maslach (1977) explained that a persondoes not make questionable inferences about his pub-lic behavior. Therefore, the person usually maintainshis behavior and changes his attitudes toward the

behavior privately to match the behavior.By participating in TREP and going outside and

monitoring river water in the field, students werepersuaded to adopt more positive attitudes towardwater monitoring. In the action-taking section, stu-dents were asked to develop their action plan, makedecisions and take actions in front of other students

a compliant behavior. After this section, their at-titudes toward the actions they took and the workingprocess often changed.

Petty (1986) classified two types of persuasionprocesses, the central route and peripheral route. Thecentral route provides subjects with abundant infor-mation in order to help them carefully consider theinformation and change their attitudes. The other wayis to create social pressure or a reward to induce aperson to change his or her attitudes. TREP soughtto change a person' s attitude using these two ap-proaches at the same time.

III. Research Methods

The design of this research was based on the "pre-test/post-test nonequivalent control group design." Thisdesign tried to control for the effect of history andchecked for any differences between the experimentaland control group concerning the environmental at-titudes and knowledge. Eight classes (four fromKeelung High School and four from Sonsan HighSchool) were selected by school teachers to participatein this research. Keelung High School and SonsanHigh School are located along the Keelung River, onebranch of the Tansui River (Fig. 1). The two schoolsare very different. Keelung High School is in a sub-urban area, while Sonsan High School is a, typicalurban school. The water quality near Keelung HighSchool was good. However, the river near SonsanHigh School was polluted.

Four classes two classes each from KeelungHigh School and Sonsan High School participated inTREP as experimental groups and answered the pre-and post-test questionnaires (Appendix A). TREP wasconducted by the researcher and the teachers, duringtwo days of seven hours each over weekends or holidays,outside of classes, due to the highly structured classschedule during weekdays. Most participants werevoluntary. Three-quarters of the participants (105students) participated in whole program (Table 1).

There were also four control classes which cor-responded to the experimental classes in terms of theirmajor area of study and academic performance. Duringthe research, the control classes answered the pre- andpost-test questionnaires without attending any of the

12


Table 1. The Experimental Design of the TREP Program

Groups No. Pre-test Program Post-test

Experimental classesFull participantsPartial participantsNon-participants

Control classes

1054722

176

wholepartialnonenone

program.To evaluate the program, an evaluation instru-

ment was designed according to the program objec-tives of awareness, attitude, and knowledge. This wasa questionnaire consisting of five parts (Appendix A).The questionnaire was modified based on the com-ments of a panel of 11 educators who had experiencein environmental education or educatithial evaluation.These experts reviewed it according to the objectivesand content of the program.

The reliability of the "attitudes" and "confidenceabout water quality knowledge" sections of the ques-tionnaire was calculated using a Chronbach alpha withSPSS software to check internal reliability (Table 2).The reliability test for the "knowledge of water qual-ity" section was conducted using the split-half method.The reliability coefficients for most parts were higherthan 0.65 and were therefore accepted. Because theconstruct of "attitudes toward saving the KeelungRiver" is diverse, the internal consistence is low.However, the situation of low internal consistence wasovercome in analysis because each dimension of thispart had its own score, rather than being combinedto get the score of the part as a whole.

The data analysis included descriptive and infer-ential statistics. The assumption for normality wastested by the skewness and kurtosis processes. Mostresults indicated no serious departure from normaldistribution. The assumption for equal variance wasmet by a residual error term. Therefore, the pairedt-test was used to analyze the change between the pre-

test and post-test of most parts of the questionnaire,except the "awareness of the river" and "attitudestoward saving the ri ver" sections. The one-wayANOVA analysis was used to detect the differencebetween the four groups in the pre-test and the post-test. The level of significance for these statistics wasset at 0.05.

The "awareness of the river" and "attitudes to-ward saving the river" sections had to be analyzeddifferently in the pre- and post-test comparison, be-cause the post-test of the two parts were included instudent evaluation sheets which did not include stu-dents' names. Thus, the pre-test and post-test resultsin these two parts could not be matched individually.This caused the pre-test and post-test scores for theseparts to be dependent. The difference between pre-and post-test results was analyzed with one-wayANOVA and the significance level was set as 0.01 toovercome the "dependent variable" issue.

The research had three constraints to interpretingthe data. First, the results were not able to show thatTREP is more effective than a regular lecture approachor any other approach, because the control classes didnot receive any of instruction. Second, it cannot beassumed that the program effect would have beenexactly the same if the program were conducted duringregular class time. Third, using the same questionnairefor the pre- and post-test may have caused a postponedeffect.

Overall, the research dealt with program effectsin terms of attitude and knowledge and was carriedout using the following five null hypotheses.

(1) There is no significant difference in partici-pants' awareness of the studied river in thepre- and post-test for TREP.

(2) There is no significant difference in partici-pants' attitudes toward saving the river inthe pre- and post-test for TREP.There is no significant difference betweenthe control group and the experimental groupin terms of feelings of responsibility for

(3)(i)

Table 2. Reliability Coefficient for Each Section of the Questionnaire

Section Coefficient of Reliability

Awareness of the Keelung RiverAttitudes toward saving the Keelung RiverKnowledge of water qualityConfidence about water quality knowledgeAttitudes toward responsibility for environmental protectionIntentions of taking environmental actionsPerceptions of barriers to action taking

0.680.350.870.800.660.710.76

13

AS

S.M. Wang

Table 3. Pre- and Post-test Comparison of Awareness of the Keelung River

Name of schoolDimensions

Mean Differencebetweentwo tests

F valuePre-test Post-test

Keelung SchoolFamiliarity with river 3.98 5.67 1.69 63.06 0.000Evaluation of river's health 2.61 4.27 1.66 69.10 0.000Good feelings about river 2.91 4.02 1.11 28.63 0.000

Sonsan SchoolFamiliarity with river 3.62 5.49 1.87 49.96 0.000Evaluation of river's health 2.03 2.35 0.32 2.46 0.120Good feelings about river 2.36 2.56 0.21 0.93 0.336

Note: Score is from 1 to 7.

environmental protection after TREP.(ii) There is no significant difference between

the experimental group's pre-test and post-test in terms of feelings of responsibility forenvironmental protection.

(4)(i) There is no significant difference betweenthe control group and the experimental groupin terms of intention to take action afterTREP.There is no significant difference betweenthe experimental group's pre-test and post-test in terms of intention to take action.There is no significant difference betweenthe control group and the experimental groupin terms of the knowledge of river waterquality after TREP.

(ii) There is no significant difference betweenthe experimental group's pre-test and post-test in terms of knowledge of river waterquality.

(iii) There is no significant difference betweenthe control group and the experimental groupin terms of their confidence in their knowl-edge after TREP.

(iv) There is no significant difference betweenthe experimental group's pre-test and post-test in terms of their confidence in theirknowledge.

IV. Result and Discussion

1. Students' Awareness of the Keelung River

Null hypothesis 1 (familiarity with the river) isrejected according to an ANOVA test with 0.01 sig-nificance level. TREP significantly increases partici-pants' awareness of the Keelung River in terms offamiliarity of the river. The mean differences between

the pre- and post-tests of the experimental groups atKeelung High School and Sonsan High School were1.69 and 1.87, respectively.

Students' awareness of the river was measuredaccording to three dimensions: their familiarity withthe river, their evaluation of its cleanness, and theirfeelings about it. After the program, the participantsranked their level of familiarity with the KeelungRiver as higher than before the program (Table 3).Before the program, they were not familiar with theriver. However, students evaluated the Keelung Riveras very polluted and felt negative about it.

In addition, the program helped participantsclarify their perceptions of the river's health and theirfeelings about the river so that the correlation coef-ficient between these two variables in the post-testincreased (Table 4). The Keelung High School par-ticipants, who monitored a pristine area of the KeelungRiver during the monitoring section of the program,re-evaluated their original perceptions of the river'spollution and their feelings about the river becamemore positive. However the Sonsan High Schoolparticipants' evaluations and feelings about the riverdid not change after they had visited a polluted partof the Keelung River (Table 3).

Familiarity with a particular environment is aneffective predictor of preference for the environment(Medina, 1983) and an indicator of awareness of the

Table 4. Correlation Coefficient between Students' Evaluation ofthe River's Health and Their Feelings about the River

Evaluation of river's health

Pre-test Post-test

Feelings toward the river 0.69 0.85

14

11 9


environment. The baseline data showed that exam-inees were not familiar with the Keelung River beforethe program, although the two schools were locatedalong the Keelung River. These findings were similarto Hsiao' s results (Hsiao, 1991). His research showedthat the general public did not think of natural water,i.e. a river or lake, when they were asked about water.The author thinks that, in the case of high schoolstudents, the overwhelming tensions of academic studydo not often give them a chance to visit the Keelungor Tansui River. This implies that local people (in-cluding senior high school students) have lost a feelingof connection with rivers. These feelings are essentialfor stimulating them to take action.

Although they were not familiar with the KeelungRiver in the pre-test, they assessed it as polluted,dirty, and ugly. This phenomenon also occurredin surveys done by Hsiao (1991) and Hwang(1991). Their data show that the general public andvocational school students evaluated Taiwanese riversin general as "seriously polluted" in spite of differentlevels of pollution in different rivers. These prevailingideas that rivers are seriously polluted may causepeople to feel powerless and block them fromtaking action, according to Bardwell's study (Bardwell,1992).

These previous data proved that, by directly ex-periencing the river, learners have an opportunity touse their senses hearing, seeing, touching and smell-ing to build up salient and non-rejectable informa-tion on the studied river. The situation is supportedby Fazio and Zanna' s experiment (Fazio and Zanna,1981) and Bem' s theory (Bem, 1970). The benefit ofdirect, purposeful experiences, like the water moni-toring activity, for developing awareness and sensi-tivity towards the environment has been noted

(Engleson, 1985).

2. Students' Attitudes toward Saving the River

Null hypothesis 2 is rejected according to anANOVA test, with 0.01 significance level. TREPsignificantly increases participants' feelings ofempowerment toward saving the river. The meandifferences between the pre- and post-tests of the ex-perimental groups in Keelung High School and SonsanHigh School were 0.73 and 1.06, respectively.

After the program, participants from both schoolsfelt more empowered to save the Keelung River, asindicated by their scores on the statements, "the taskwill be successful," "I feel empowered to improve theriver," "I have ideas about what to do," and "I canhelp." However, only Sonsan participants' feelingsof responsibility for the task significantly increasedafter the program. The change in the Keelung par-ticipants' feelings of responsibility was on the statis-tical margin (Table 5). In addition, only the Keelungparticipants felt that saving the Keelung River was notas difficult as they had perceived before the program.This was perhaps because the river area they visitedseemed relatively clean based on its appearance andthe water quality data they obtained. This researchfound that the negative influence of perceived diffi-culty on empowerment was reduced by TREP (Table6). Possibly, this was because TREP provided theknowledge and skills necessary to overcome the nega-tive feelings of powerlessness.

With regard to students' attitudes toward savingthe Keelung River, the baseline data show that studentsin these senior high schools did not feel empoweredor responsible for undertaking the task, although most

Table 5. Pre-test and Post-test Comparison of the Two Schools in Terms of Attitudes Toward Saving the Keelung River

Mean DifferenceName of school between F value

Pre-test Post-testDimensions two tests

Keelung SchoolImportance 6.46 6.40 -0.05 0.14 0.764Difficulty 5.50 4.40 -1.10 19.26 0.000Empowerment 4.41 5.15 0.73 23.38 0.000Responsibility 5.05 5.32 0.28 4.82 0.03m

Sonsan SchoolImportance 6.58 6.65 0.07 0.20 0.659Difficulty 5.89 5.66 -0.23 1.20 0.277Empowerment 4.22 5.28 1.06 25.61 0.001Responsibility 5.12 5.56 0.43 6.42 0.010

Note: Score is from 1 to 7.M: Margin (p<0.05)

15

20

S.M. Wang

Table 6. Mean of Four Items in the Empowerment Dimension and the Correlation of Success and Empowerment with Difficulty in thePre-test and Post-test

Items ofEmpowermentDimension

Mean

Pre-test Post-test

Correlation coefficientwith difficulty dimension

Pre-test Post-test

The task will be successfulI feel empowered to do itI have ideas about how to help itI can help in achieving the task

5.203.543.575.0

5.95**4.56**473**5.57**

Perception of Difficulty 5.68 4.95***

-0.15 -0.12-0.34 -0.10

** the difference between pre- and post test meets a significance level of less than 0.001.

felt that saving the river was very important and urgent.Hwang' s survey (Hwang, 1991) also showed that 65%of 2,794 surveyed vocational school students agreedwith the statement, "I am a student, therefore what Ican do for the environment is limited." Therefore therewas a discrepancy between students' perceptions ofthe importance of saving the river and their feelingsabout doing it.

The case study on citizen action, as well as theirown action-taking experience in the program may haveincreased students' feelings of empowerment andresponsibility for saving the river. Bandura (1977)pointed out that people' s feelings of self-efficacy willincrease after studying a successful case and accom-plishing tasks.

However, the author raised a concern about watermonitoring activities in Taiwan. The research findingsshow that students' experience with the river stronglyinfluenced their feelings about the river and theirevaluation of the difficulty of saving the river. Gen-erally, a negative feeling towards the river oftenprevents people from accessing or discussing the river.

The findings lead to the question "What consti-tutes a good site for students to do water monitoring?"According to the community problem solving approach,the water monitoring sites ought to be close to par-ticipants' schools or homes (community-based) so thatteachers can link the river issues with students' livesand encourage them to follow up with action taking(problem solving) in their families, school or commu-nity. However, in Taiwan, the lower and middlereaches of most rivers, where most students live, areseriously polluted. If students visited the pollutedparts, their perceptions of river pollution would in-crease. Therefore, if there is more than one oppor-tunity to do water monitoring, it would be a good ideato visit an unpolluted part of the river in addition tothe rivers near their schools or homes. Students can

then build a sense of appreciation and respect for therivers.

3..Students' Sense of Responsibility forEnvironmental Protection

Null hypothesis 3(i) is rejected according to anANOVA test with 0.05 significance level. Null hypoth-esis 3(ii) is rejected by a paired t-test with 0.05significance level. TREP significantly increases stu-dents" feelings of responsibility for school environ-mental protection. The mean difference between thepre- and post-test of the experimental group was 0.2.The mean difference between the control group andthe experimental group after the program was 0.4.

Participants' attitudes toward who should be re-sponsible for the school environment were examinedaccording to four aspects of the process: providingideas to initiate a task, planning, decision making, andfunding. In terms of decision making and planning,participants significantly changed their attitudes afterthe program (Table 7). Before the program, they wereneutral about who should be responsible for thesetasks; after the program, they felt that students shouldhave greater responsibility for them than the admin-istratiop. With regard to providing ideas to initiateschool environmental protection, the participantstended to attribute responsibility to the students bothbefore and after the program. In the pre-test and post-test, the control and full participation group boththought.that the school administration should providemore funding for school environmental protection thanstudents.

In the pre-test, student attitudes tended to beneutral. They felt the administration and studentsshould equally share responsibility for the school en-vironment. Or, possibly, they had no particular feel-

16

21


Table 7. Examinees' Sense of Responsibility for School Environment in the Pre- and Post-test

Group

Mean Differencebetween the

two testsPaired t

Pre-test Post-test

Attitudes toward decision makingControl classes 2.92 2.76 -0.17 -1.75 0.082

Full participants 2.93 3.28 0.32 2.66 0.009

Attitudes toward planningControl classes 2.79 2.70 -0.08 -0.82 0.415


Attitudes toward responsibility for initiatingControl classes 3.99 3.73 -0.27 -3.31 0.001


Attitude toward fundingControl classes 2.15 2.22 0.04 0.55 0.580


Note: Score is from I to 5.

ings about it and their rating was influenced by the"middle-way philosophy" that is dominant in Chinesesociety.

The middle-way philosophy means maintainingbalance and not going beyond society' s limits. Itwould be good if students indeed strongly believedthat the administration and students should equallyshare the responsibility. However, in real-life situ-ations, people' s attitudes often do not exactly hold toa "middle-way" position and tend to be non-neutral.

A bottom-up and public participation approachfor environmental protection is usually promoted bysocial environmentalists. (Milbrath, 1989) This re-quires that people have very strong attitudes towardresponsibility for initiation, planning, decision mak-ing, and funding of local environmental protection.The author feels that a democratic atmosphere in school

allowing students to make their own plan and finishit may be a useful way for managing students'cleaning tasks in schools or communities and estab-lishing a sense of their environmental responsibility.However, generally, school administrators are domi-nant and decide on a working plan for environmentaltasks and mandate students to implement it. Studentsmust follow this mandate an opportunity for decision-making or taking responsibility on their own.

4. Students' Intention to Take EnvironmentalActions

Null hypothesis 4(i) is rejected by an ANOVA testwith 0.05 significance level. Null hypothesis 4( ii) isrejected by a paired t-test with 0.05 significance level.TREP significantly increases students' intentions totake actions. The mean difference between the pre-

and post-test of the experimental group was 0.5. Themean difference between the control group and theexperimental group after the program was 0.65.

Students were asked how likely it was that theywould take each of the 15 environmental actions.According to the mean level of intention, the 15 actionswere divided into three clusters by factor analysis:"personal behavior change," "advocacy and self-edu-cation," and "school environmental action". After theprogram, participants' intentions to take actions re-lated to "advocacy and self-education" and "schoolenvironmental actions" significantly increased (Table8). However, participants did not increase their in-tention to take "personal behavior change" approach,possibly because the pre-test scores were already quitehigh, 4.8 (The highest value is 6.0).

Chow' s survey (Chow, 1992) also indicated thatprimary and secondary school teachers had greaterintentions to change personal behavior than other op-tions (teaching students about the environment inclasses, persuading the whole school to protect theenvironment and participating in environmental ac-tivities out of school). In the U.S., Jordan' s study(Jordan, Hungerford, & Tomera, 1986) involvingAmerican students showed that their knowledge ofenvironmental action-taking was much more orientedtoward an eco-management approach than any otherapproach both before and after his program. This maybe because changing personal behavior (eco-manage-ment) relates more to common sense, is socially ac-ceptable, simpler, less controversial, and requires lesstime and energy than the other actions. The authorthink these attitudes are supported by beliefs commonin Chinese culture: that one's foremost concerns should

17

22

S.M. Wang

Table 8. Pre-test and Post-test Comparison of the Control Classes' and Full Participants' Intentions to Take Environmental Actions inthe Three Clusters

Mean

Clusters

Control (169) Full (104)

Pre-test Post-test Pre-test Post-test

Personal behavior Advocacy 4.73 4.80 4.70 4.89and learning 3.44 3.46 3.45 4.14**School env. protection 3.45 3.43 3.74 4.27**

Notes: Scale is from 1 to 6.Comparison between pre- and post-test used two-tailed, paired t-test with alpha=0.05.

** P<0.001

be about personal daily life and fulfilling one's duties.Beyond the individual and family levels, there are fewfeelings of responsibility for public affairs and littletendency to act for the public good.

Larson, Forrest, & Bostian (1981) argued thatenvironmental actions taken in the household are nota good indicator of people's commitment to solveenvironmental problems. Personal behaviors are lessexpensive, more expedient and/or habitual than politi-cal and legal actions. Therefore, it is meaningful thatTREP increased students' intention of taking actionsrelated to persuasion and school environmental tasks.Actually, at the end of this program the students wroteletters and made posters to increase their schoolmates'concern for the rivers.

5. Students' Knowledge of Water Quality

Null hypothesis 5(i) is rejected by an ANOVA testwith 0.05 significance level. Null hypothesis 5(ii) isrejected by a paired t-test with 0.05 significance level.TREP increases students' knowledge of water quality.The mean difference between the control and experi-mental group after the program was 0.06. The meandifference between the pre- and post-test of the ex-

perimental group was 0.06.Null hypothesis 5(iii) is rejected by an ANOVA

test with 0.05 significance level. Null hypothesis 5(iv)is rejected by a paired t test with 0.05 significancelevel. TREP increases students' feelings of confidenceabout this knowledge. The mean difference betweenthe control and experimental group after the programwas 0.31. The mean difference between the pre- andpost-test of the experimental group was 0.30.

Students assessed the impacts of the twelveactivities on river water quality as "positive," "nega-tive," or "zero" (Appendix A). After the program, theparticipants' accuracy in answering these questionsincreased and was higher than that of the controlclasses (Table 9). Participants' confidence in theirknowledge of water quality also increased and washigher than the control classes in the post-test (Table10). After the program, the students' ability to cor-rectly assess the impact of the twelve activities onoverall river water quality, as well as their confidencein their answers, had increased. But the mean differ-ence between the pre- and post-test was small becausethe students had been quite knowledgeable before theprogram (Table 9, 10 ). Talsma's evaluation (Talsma,

Table 9. The Mean Score of Water Quality Knowledge of Each Group in the Pre-test and Post-test and the Difference between Pre- andPost-test

Mean score' DifferenceGroup No. between pre Paired t

Pre-test Post-test and post-test

Control classes 170 0.82 0.82 0.00 0.05 0.957Experimental classes 172 0.80 0.85 0.06 4.81 0.000

Non-participants 22 0.80 0.82 0.02 0.88 0.389Partial participants 47 0.77 0.81 0.05 1.66 0.104Full participants 103 0.82 0.88 0.06 5.15 0.000

'Mean score is between 0 and I.

18

3


Table 10.The Mean Scores for Confidence in the Water Quality Knowledge of Each Group in Pre-test and Post-test and the Differencebetween Pre- and Post-test

Group No.Mean score'

Pre-test Post-test

Differencebetween

pre- and post-Paired t

Control Classes 169 3.18 3.20 0.01 0.42 0.675Experimental Classes 172 3.18 3.40 0.22 6.67 0.000

Non-participants 22 3.24 3.27 0.03 0.53 0.605

Partial participants 47 3.06 3.22 0.16 2.16 0.036

Full participants 103 3.22 3.51 0.30 7.00 0.000

'Four points scale covers from 1. very unsure, 2. unsure, 3. sure, 4. very sure.

1992) of the Rouge Project also found this to be thecase.

However, students' accuracy in assessing theimpacts of the three activities "fishing from theriver bank," "using soap instead of detergent," and"removing mud from the river bottom" did notimprove after the program (Table 11). Moreover,the mean scores for accuracy in assessing "fishing,""removing mud from river bottom," and "usingsoap instead of detergent" did not increase signifi-cantly and students' confidence for these threeitems was lower than their confidence for otheritems. This is not surprising, as the task of"removing mud from the river bottom" was notmentioned in the program. Also, the participantsmay have rated "fishing" as a negative impactbecause the third section of the program mentionedillegal fishing with poisons; students were also notvery sure of their answers. Nearly half of the full

participants assessed "using soap instead of detergent"incorrectly, possibly because students did not knowthe benefit of using soap, or because students answeredthe question without comparing detergent with soapand thought that soap was a pollutant. These resultsimply that TREP did not improve participants' waterquality knowledge in certain aspects, yet tended toincrease their confidence about their knowledge ofwater quality.

IV. ConclusionThe Tansui River Education Program is an

innovation within Taiwanese formal education.During the program, students go out of school tomonitor their rivers and take actions for the river(interdisciplinary characteristics and community prob-lem solving approach). The research found TREPsignificantly increased participants' environmental

Table 11. Comparison of Full Participants' Pre-test and Post-test Mean Scores in Terms of Accuracy and Confidence

No. Item

8 acid rain3 replacing lawns with cementI overuse of fertilizer4 discharge from septic tank9 garbage on the streets5 changing woodland into golf course7 fishing from river bank

10 hot water discharged from factory6 digging up sand along river bank

11 building roads over the river12 removing mud from river bottom

2 using soap instead of detergent

Accuracy Confidence

pre-test post-test pre-test post-test

0.98 0.97 3.52 3.67*

0.95 0.99 3.37 3.62***0.93 0.99 3.48 3.67**

0.90 0.98* 3.50 3.70**

0.89 0.99** 3.52 3.68*

0.88 0.96* 3.35 3.56**0.82 0.77 2.95 3.29***

0.80 0.90** 3.23 354***0.77 0.89** 3.10 3.47***0.75 0.89*** 3.06 3.42***

0.63 0.64 2.90 3.29***

0.51 0.58 2.68 3.22***

Note: A paired t-test with 0.05 significant level was used in comparing the pre- and post-test scores, which were independent.* p<0.05** p<0.01*** p<0.001

19 -

24

S.M. Wang

awareness, knowledge, attitudes, and empowerment.In addition, participating students and teachers werein favor of it. Therefore, the author suggests this typeof education program should be promoted withfurther research to compare it with other teachingapproaches.

Acknowledgments

I would like to thank the teachers and studentswho participated in the Tansui River Education Pro-gram and their schools. I am grateful for financialsupport from the National Science Council of theRepublic of China (NSC 81-0421-S-003-01-Z). Pro-fessors Kuang-Jen Yang and the Environmental Edu-cation Center at National Taiwan Normal Universityprovided administrative support which I deeply appre-ciate.

Appendix

Questionnaire

A. A couple of paired adjectives and short sentences are listed below. Please mark the level of each in "saving the Keelung River".

The Keelung River

I strongly agree 3 2with the leftdescription

I know it vaguelyI am unfamiliar

with the riverIt is dirtyIt is turbidIt is beautifulI dislike itI feel sad

about itI am worried

about it

I strongly agree 3 2with the leftdescription

It is necessaryIt is urgentIt is difficultSimple taskEfforts to save it

would beunsuccessful

It' s government' sresponsibility

I feel helplessI don't know what

I can doI can helpThe task is a

mission for meMy life will

change for it

1 1

Saving the Keelung River

1

20

1

25

2 3

2 3

I strongly agreewith the rightdescription

I know it wellI'm familiar with

the riverIt is cleanIt is transparentIt is uglyI like itI feel happy

about itI am content

with it

I strongly agreewith the rightdescription

It is unnecessaryIt is not urgentIt is easyComplex taskEfforts to save it

would no besuccessful

It's the citizens'responsibility

I feel empoweredI have ideas about

what I can do.I can not helpIt is a trouble for

meMy life will not

change for it


B.1. The following activities may affect the water quality of rivers in different ways by improving or degrading or not affecting thewater quality. Please circle the effect of each activity in column A."+". improves Water quality or reduces degradation of water quality"-" = degrades or worsens water quality"o"= does not affect water quality

B.2. At the same time, how confident do you feel about your knowledge when you answer each question? Please indicate the extent towhich you are confident in answering each question in column B.

Activitiescolumn A

impact on water qualityvery

unsure

column Bdegree of confidence

verysure

good none bad unsure sure

I. Overuse of fertilizer in the farmers' fields + 0 1 2 3 4

2. Using soap instead of detergent for laundry + 0 1 2 3 4

3. Changing lawns into cement along the river + 0 1 2 3 4

4. Discharge from septic tank + 0 1 2 3 4

5. Changing woodlands into golf courses + 0 1 2 3 4

6. Digging up sand along the river bank + 0 1 2 3 4

7. Fishing from the river bank + 0 1 2 3 4

8. Acid rain + 0 1 2 3 4

9. Garbage on the streets + 0 1 2 3 4

10. Hot water discharged from factories + 0 1 2 3 4

11. Building roads over the river + 0 1 2 3 4

12. Dredging the mud from river bottom + 0 1 2 3 4

C. The following are paired statements about who has responsibility for school environmental protection, the school administration orstudents. After you consider the ideal and actual situation (personal factors, academic work load, school administration and society),who do you think should take charge of school environmental protection? Please indicate your opinion about each pair of statementsby circling a number.

1=strongly agree left 2=agree left 3=neutral4=agree right 5=strongly agree right

I. Students should wait for directions 1 2 3 4 5

from the administration then take action.

2. The direction of school environmental 1 2 3 4 5

protection should be decided by theadministration(e.g. whether or not the schoolshould recycle)

3. School environmental protection tasks I 2 3 4 5

should beplanned by the administration(e.g. how to do recycling tasks)

4.. School environmental protection should be 1 2 3 4 5

funded by the administration.

Students should initiate school environmentalprotection (e.g. providing ideas on recycling).

The direction of school environmentalprotection should be decided by student

School environmental protection tasks shouldbe planned by students.

School environmental protection should befunded by students

D. I. The following are some environmental actions or activities. How likely do you think it is that you would take these actions orparticipate in these activities? Please circle the number that is closest to your opinion.1=very unlikely 2=unlikely 3=somewhat unlikely4=somewhat likely 5=likely 6=very likely

1. Telling your family to reduce use of detergents 2 3 4 5 6

2. Not throwing away bruised vegetables and fruits 2 3 4 5 6

3. Informing the school about dripping faucets 2 3 4 5 6

4. Watching TV programs about the environment 2 3 4 5 6

5. Advising neighbors not to dump illegally 2 3 4 5 6

6. Writing an article about the environment to persuade the public 2 3 4 5 6

21

26

S.M. Wang

7. Collecting articles about the environment 1 2 3 4 5 68. Writing a letter to the EPA for law enforcement 1 2 3 4 5 69. Calling the local EPA about polluting behavior 1 2 3 4 5 6

10. Making a poster to increase schoolmates' environmental awareness I 2 3 4 5 611. Turning off lights which are not needed 1 2 3 4 5 612. Cleaning drains in the school

1 2 3 4 5 613. Investigating where school waste water flows 1 2 3 4 5 614. Collecting and storing chemicals used in the school's lab 1 2 3 4 5 615. Being a member of an environmental group

1 2 3 4 5 6

References

Bandura, A. (1977). Self-efficacy: toward a unifying theory ofbehavioral change. Psychological Review, 84(2), 191-215.

Bardwell, L. (1992). Success stories: imagery by example. Journalof Environmental Education, 23(1), 5-10.

Bern, D. J. (1970). Belief, attitude, and human affairs. Berkley,CA: Brooks/cole Publishing Company.

Bull, J., Cromwell, M., Cwikiel, W., Chiro, G., Guarino, J., Rathje,R., Stapp, W., Wals, A., & Youngquist, M. (1988). Educationin action: A community problem-solving program for schools.Dexter, MI: Thomson-Shore Printers.

Chow, G. L. (1992). Study on primary and secondary schoolteachers' attitudes toward environmental issues (in Chinese).Unpublished master's thesis. National Taiwan Normal Univer-sity, Taipei, Taiwan.

Engleson, D. C. (1985). A guide to curriculum planning in envi-ronmental education. Madison, WI: Wisconsin Department ofPublic Instruction.

Environmental Protection Administration (EPA). (1989). Status ofenvironmental protection in Taiwan, R.O.C. (in Chinese).Taipei, Taiwan: EPA.

Fazio, R. H., & Zanna, M. P. (1981). Direct experience and attitude-behavior consistency. Advances in Experimental Social Psy-chology, 14, 162-202.

Festinger, L. (1957). A theory of cognitive dissonance. IL and NY:Row, Peterson Company.

Hines, J. M., Hungerford, H., & Tomera, A. (1986). Analysis andsynthesis of researches on responsible environmental behavior:a meta-analysis. Journal of Environmental Education, 18(2),1-8.

Iisiao, S. H. (1991). The general public's awareness, attitudes, andbehavior related to water. In: Water concern (in Chinese) (pp.2-13). Taipei, Taiwan: China Times Publishers.

Hwang, C. C. (1991). Study on environmental awareness of juniorcollege students (in Chinese). The Proceeding of the First ROCSymposium on Research on Environmental Education (pp. 61-90). Taipei, Taiwan.

Jordan, J. R., Hungerford, H., & Tomera, A. (1986). Effects oftwo residential environmental workshops on high school stu-

dents. Journal of Environmental Education, 18(1), 15-22.Kaplan, S., & Kaplan, R. (1982). Cognition and environment:

functioning in an uncertain world. New York, NY: Praeger.Koballa, T. R. (1988). Attitude and related concepts in science

education. Science Education, 72(2), 115-126.Larson, M., Forrest, M., & Bostian, L. (1981). Participation in

pro-environmental behavior. Journal of Environmental Educa-tion, 12(3), 21-24.

Medina, A. Q. (1983). A visual assessment of children's and en-vironmental educators' urban residential preference patterns.Unpublished doctoral dissertation, University of Michigan, AnnArbor, MI.

Milbrath, L. W. (1989). Envisioning a sustainable society learningour way out. New York, NY: State University of New YorkPress.

Petty, R. E. (1986). Communication and perivasion: Central andperipheral routes to attitude change. New York, NY: Springer-Verlay.

Phillips, J. L. (1981). Piaget's theory: A primer. Seatle, WA:Freeman and Company Publishers.

Song, M. S. (1989). The connection of nature, society, and culture- social psychology of the human living environment (inChinese). Environmental Education Magazine, 4, 2-7.

Stapp, W. B. (1983). Status and analysis of unesco's effort to furtherenvironmental education internationally - toward a nationalstrategy for environmental education. In: M. Cowan & W. B.Stapp (Eds.), Environmental education in action (pp. 7-9).Columbus, OH: ERIC.

Swan, M. (1975). Forerunners of environmental education. In: N.McInnis & D. Albrecht (Eds.), What makes education environ-mental? Louisville, KY: Data Courier, Inc. and WashingtonD.C.: Environmental Education, Inc.

UNESCO. (1980). Environmental education in the light of thetbilisi conference. Paris: UNESCO Press.

Talsma, V. L. (1992). The interactive rouge river water qualitymonitoring program: An evaluation of an environmental edu-cation program. Unpublished master's thesis. University ofMichigan, Ann Arbor, MI.

Zimbardo, P. G., Ebbesen, E. B., & Maslach, C. (1977). Influencingattitudes and changing behavior. New York, NY: Addison-Wesley Publishing Company.

22

2 7


-it 11 T=It 5V-it

-t;Ff

_I_ PIA

IALA- %!Vofi-ii.k alAliRf-Vir

tf4N

2PifMt5On?ralf-A2MA*q3*trIM*ir-JRMMIRWHM.kfilMftWHMtItH@M:MROW*M.Mtg.06,XPJA-EM.0*WKM.F5t°(5)MItYMM .fit."VAKReff,XICE0Te4047*RAql-t4LICP-7-*VAMOMIMNHO'iLkftWAM'WMUffi.ZMWOMINffnVVVIWWWOMMWARMI'V4HtAt401±RM.A8MM&KM''.4.R53-PlIgn-MMYQMI'W7141tO*RMARWHMISOJEMI'Mft0AVOMV.-AllAiittRWVUEVMMOMPA'ft$A11,1i-MgMVtiffPAIlltMXfaRnMoONt*MOgilA.aRRJWAP*.(3)At*MftflWW11AfFlOgilMIWAiffnlIMSA.WEtiffMqM0-04VIVIliffilWAMMtniffe.Eavtammtuffm*wmnficog4nmft.xvimm*,*AllmftwowgtamiPJniTumvomsivAlmrP.

23

4


Taiwanese Elementary Students' Perceptions ofDinosaurs

CHI-CHIN CHIN

Department of Science EducationNational Museum of Natural Science

Taichung, Taiwan, R.O.C.

(Received July 8, 1996; Accepted October 11, 1996)

ABSTRACT

In a child's mind, dinosaurs are creatures with mystical charm. Every major natural history museumin the world has a rich dinosaur collection and fascinating dinosaur galleries which are very popularamong children. Although little information on dinosaurs exists in Taiwanese k-12 textbooks, childrenin Taiwan obtain related information through various informal channels including science museums.

This study conducted by curators at the National Museum of Natural Science(NMNS) mainlyinvestigated the perceptions of Taiwanese children of dinosaurs. A questionnaire composed of multiplechoiceand open-ended questions was developed as an instrument for data collection. A total of 651 childrenbelonging to three age groups (grades 2, 4 and 6) were randomly selected from 5 elementary schoolsin the Taichung Area and asked to answer the questionnaire. Qualitative data was collected by interviewingboth children who answered the questionnaire (N=24) and who visited the dinosaur gallery (N=12). Theresults showed that there was a significant difference in preference towards dinosaurs between boys andgirls. This study also reports on the causes of children's perceptions, children's informational sources,and their understanding of the reasons for the extinction of dinosaurs. The findings are not only providedas a reference for science educators, but will also help NMNS understand what young visitors thinkand give it an empirical basis for improving the dinosaur gallery.

Key Words: elementary students, dinosaurs, perceptions, Taiwan

I. Introduction

In a child's mind, dinosaurs are creatures withmystical charm. Every major natural history museumin the world has a rich dinosaur collection and fas-cinating dinosaur galleries which are very popularamong children. Thus the dinosaur gallery at theNational Museum of Natural Science (NMNS) inTaiwan has been a big attraction for school childrensince it opened to the public in 1988.

The dinosaurs constitute a biological populationwith great diversity in size, shape and feeding habits.Their evolution and controversial processes of extinc-tion can provide a suitable framework for children tounderstand the scale, time sequence and chronologicalrelation of life on Earth. This can also provide aknowledge base helping children to understand earth'sbiological diversity, the impact of their environmenton living organisms, the causes of extinction and themechanisms of competition. If approached correctly,the study of dinosaurs could become one of the best

tz,

topics for introducing children to the fasinating worldof natural history.

Since 1841, the year the word "dinosaur" wasfirst coined by British paleontologist Richard Owen,numerous publications introducing dinosaur issues withboth scholar and popular orientations have been pub-lished. Besides publications relating scientists' inves-tigations of dinosaurs (British Natural History Mu-seum, 1979; Lambert & the Diagram Group, 1990;Norman, 1985, 1991; Benton, 1992), much materialabout dinosaurs in the form of writings, films, andgames etc. have been created incorporating the author'sown imagination. The most notable instance of thisin the recent years is the film Jurassic Park, which wasdirected by Steven Spielberg based on a fictitious storywritten by the American author Michael Crichton.Such films have not only attracted big audiences inthe theater, but have also spurred the prevailing in-terest in dinosaur issues.

The reasons for such public interest in a groupof species extinct for 65 million years has always been

24

2 9

Elementary Students' Perceptions of Dinosaurs

an interesting question for biologists, educators, psy-chologists, and sociologists. As museum curators,Gardom & Milner (1993) tried to provide explanationsfor this phenomenon from three perspectives:

(1) First, dinosaurs were such special animals thatpeople somehow look at them as if they weremonsters described in legends and myths. Sucha belief fills people with feelings such asexcitment, astonishment and enchantment.

(2) Second, since dinosaurs are all extinct, peopledo not need to be concerned about what peoplemight have done to hurt these species. Thislessens people's guilt while dealing with dino-saur issues.

(3) Finally, because knowledge about dinosaurs isstill limited, novel findings are emerging all thetime. Therefore, the conceptual framework aboutdinosaurs created by the scientific communityoften changes when new materials are added.Such a situation provides amateurs a broad scopefor satisfying their desire to investigate.

Science educators have investigated children'smisconceptions concerning various aspects of naturalphenomena in both Taiwan and other countries (Gil& Carrascosa, 1990; Rowell, Dawson, & Lyndon, 1990;Wu, 1993). In the life sciences, Taiwanese researchershave investigated children's conceptions concerningecology, plants, life and so on (Yu, 1995; Hsiung,Chang, & Hsiung, 1996; Dai, 1995; Huang, 1996).However, none has done research on conceptions ofpaleontology. In the United States, Chi & Koeske(1983) selected dinosaurs as a vehicle to study howchild's memory performance was affected by theirknowledge. In Barba's study (Barba, 1995), she foundthat children had developed their concepts of dino-saurs at a tacit level by experiencing models, pictures,and films. Even though young children had difficultyverbalizing their conceptual understandings of ancientfauna, they were able to classify representations offauna as being Mesozoic or non-Mesozoic species withhigh degree of accuracy. As children matured and/or had more encounters with dinosaur-related con-cepts, they were able to verbalize more geological-time-related explanations of ancient life. Barba con-cluded that her findings supported Polyani's Theoryof Tacit Knowledge in that children' s conceptualunderstanding was built first at a tacit level and laterdeveloped at an explicit level.

As a part of the informal learning system, sciencemuseums have put much effort into studying theirvisitors (Finson & Enochs, 1987; Dierking & Falk,1994; Hood & Roberts, 1994; Diamond, 1994; Boisvert& Slez, 1994; Lewenstein, 1993; Screven, 1993), the

design of exhibits (Bernfeld, 1993), and the effect ofpublic services (Silverman, 1993). Being a focal exhibitarea in the museum, dinosaurs can serve as a "hot"issue for curators of NMNS to learn about the inter-actions between visitors and exhibits, and about waysto facilitate such interactions.

Although dinosaur fossils have never been dis-covered in Taiwan, and there is no topic introducingdinosaurs in Taiwanese k-12 textbooks, through thechannels of informal learning such as mass media,scientific publications, dinosaur exhibits and peercommunication, Taiwanese children have been able toconstruct a knowledge framework about dinosaurs.According to Solomon (1987), social influences playimportant roles in children's understanding of scien-tific issues. Based on this premise, it is especiallyinteresting for science educators to investigate whatTaiwanese children's conceptual frameworks of dino-saurs are like and how they blend information froma variety of sources. The results of this study couldbe used to help NMNS to understand what youngvisitors think about such of creatures and provideempirical evidence for improving dinosaur exhibits.

II. Research Questions

This study mainly focuses on answering thefollowing research questions:

(1) What is the degree of preference of Taiwaneseelementary students for dinosaurs?

(2) What are the most popular and unpopular di-nosaurs according to Taiwanese elementary stu-dents?

(3) What is(are) the main source(s) of informationabout dinosaurs for elementary students in Tai-wan?

(4) What are Taiwanese elementary students' per-ceptions of the existence and the features ofdinosaurs?

(5) What reasons for the extinction of the dinosaursare known to elementary students in Taiwan?

(6) What factors play roles in forming Taiwanesechildren's perceptions of dinosaurs?

III. Design and ProcedureIn order to achieve the objectives of this study,

a questionnaire composed of multiple-choice and open-ended questions and including topics such as theexistence of dinosaurs, their size, reasons for theirextinction, and the relationship between dinosaurs andman was developed. The main topics in the question-naire were based on research questions proposed,

25

30

C.C. Chin

discussed and agreed on by the 7 members of the studygroup. A. prototype questionnaire was distributed toa small group of students in a pilot study. The com-ments from both the children and their teachers wereused as feedback for refining the way the questionswere written and the criteria for choosing appropriatequestions.

Five elementary schools in the Taichung areawere selected as the target schools for this study. Inthese schools, 2nd, 4th, and 6th grade classes wererandomly selected to answer the questionnaire. Thequestionnaires were distributed to the sampled classeswith assistance from their teachers. Before studentsgave their answers, the teachers described the purposeof the questionnaire and how to fill it out. For the2nd-grade students, their teachers read and explainedthe meaning of each question to make sure the studentsunderstood what the questions asked. The time allot-ted for completing the questionnaire was 30 to 40minutes. A total of 651 children from 15 classesbelonging to three age groups answered the question-naire (Table 1). Among them, 85% respondentsmentioned they had visited NMNS during the last twoyears.

To obtain more in-depth information, childrenwho gave relatively correct responses for either themultiple-choice or open-ended questions wereselected for a further interview. A total of 24(14 male,10 female) students including 10, 8, and 6 studentsfrom the 6th, 4th, and 2nd grade respectively partici-pated in the interview process. In addition, someinformal interviews with respondents were alsoheld with assistance from the elementary schoolteachers. Furthermore, in order to obtain direct in-formation for the purpose of exhibit improvement, asmall number of elementary students (N=12) wereinterviewed while they were visiting the dinosaurgallery of NMNS.

Quantitative data from the questionnaires wasanalyzed by statistical methods such as frequencyanalysis and the chi-square test. Data from the inter-

Table 1. Demographic Data of the Respondents

Variables N (%) Variables

SchoolA 142 (21.8)

118 (18.1)121 (18.6)150 (23.0)120 (18.4)

SexMaleFemale

Grade Level2nd4th6th

N (%)

337 (51.8)314 (48.2)

221 (33.9)216 (33.2)214 (32.9)

Note: N=Number of respondents

views was analyzed by the analytic-induction method.Qualitative data from various interviewees was firstreviewed and then classified into primitive categoriesby individual researchers. The primitive categoriesthen served as a basis for the researchers' furtherdiscussion for the purpose of achieving consensus onthe categorization. Based on these results, assertivestatements were then generated. In presenting thefindings, selected statements from interviewees weredirectly quoted as supporting evidence.

IV. Limitations of the StudyThe quantitative data shown in this study was

mainly used as the basis for selecting students forinterviews. The statistical results generated in thisstudy were not expected to totally represent otherregions in Taiwan or Taiwan as a whole. Comparedwith the results from the random sampling procedure,the significance of the qualitative information in thisstudy should be examined carefully, although a sam-pling process and statistical methods were also uti-lized.

V. Findings

1. The Majority of Students Expressed a HighLevel of Preference for Dinosaurs

Nearly half of sampled elementary students(49.3%) expressed preference for dinosaurs (Table 2).Less than 10% of students dislike or strongly dislikedinosaurs. Significant differences also exist betweendifferent age groups (X2=36.4, df=10, P<0.001). More4th graders are fond of dinosaurs than 2nd and 6thgraders. If gender is considered, boys show greaterpreference for dinosaurs than girls (X2=56.9, df=5,P<0.001). Boys are likely to express their attitudesabout dinosaurs agressively; however, girls are rela-tively more conservative. In the interview, a girl whorepresented the typical feminine view said:

The dinosaurs are boys' stuff.

It seems that dinosaurs are included in the same categoryas things like monsters, machine guns, and otherscience-based objects, which boys are much moreinterested in.

2. The Most Popular and Unpopular DinosaursHave a High Degree of Overlap

26

The most popular and unpopular species of di-

31


Table 2. The Extent of Preference for Dinosaurs of Sampled Elementary Students Based on the Different Variables

Variables1

N(%)2 3 4 5 0

Grade2 73(33.0) 34(15.4) 11(5.0) 12(5.4) 74(33.5) 17(7.7)4 85(39.4) 35(16.2) 8(3.7) 2(0.9) 76(35.2) 10(4.6)

6 47(22.0) 47(22.0) 13(6.1) 9(4.3) 82(38.3) 16(7.5)

X2=36.4, df=10, ***P<0.001

GenderMale 146(43.3) 58(17.2) 8(2.4) 12(3.6) 94(27.9) 19(5.6)Female 59(18.8) 58(18.5) 24(7.7) 11(3.5) 137(43.8) 24(7.6)

X2=56.9, df=5, ***P<0.001

Total 205(31.5) 116(17.8) 32(4.9) 23(3.5) 232(35.6) 43(6.6)

Notes: 1=Strongly Like, 2=Like, 3=Dislike, 4=Strongly Dislike, 5=Like Some; Dislike Others, 0=No ResponseN=Number of respondents

Table 3. The Most Popular Species .of Dinosaurs Named by El-ementary Students

Rank Species Frequency

Apatosaurus 143

2 Triceratops 963 Tyrannosaurus 734 Stegosaurus 43

5 Brachiosaurus 366 Pterosaur* 247 Tanystropheus 16

8 Flying Dragons 8

9 Deinonychosaurs 7

10 Plesiosaurus* 6

11 Bubble Dinosaurs 5

12 Deinonychosaurus 5

13 Baby Dinosaurs 3

14 Plant-eating Dinosaurs 3

15 Allosaurus 3

16 Ichthyosaurs* 3

* indicates species easily mis-recognized as a dinosaur.

nosaurs named by children are also reported in thisstudy (Tables 3 and 4). Apatosaurus, Triceratops, andTyrannosaurus are regarded as the top three mostpopular dinosaurs by sampled elementary students inthis study. About 22% of students (N=143) whoanswered the questionnaire expressed a preference forApatosaurus. The reasons they gave such an answerincluded: its genial character (N=61), herbivorous habit(N=36), huge body size (N=26), wouldn't eat humans(N=26) and cute appearence (N=25), etc. The cuteappearance of Triceratops, especially the horns on itshead, is the major reason it attracts students. Insummary, the reasons mentioned by the respondents

Table 4. The Most Unpopular Species of Dinosaurs Named byElementary Students

Rank Species

1

2

3

45

67

8

9

Tyrannosaurus rexTriceratopsStegosaurusOviraptorPterosaur*DeinonychosaurusAllosaurusApatosaurusMeat-eating Dinosaurs

Frequency

21814

9

8

8

5

5

5

3

* indicates species easily mis-recognized as a dinosaur.

can be categorized as (1) the features, and (2) thenature of the dinosaurs. Some reasons also reflectedrespondents' own personality. For instance, an over-weight male 4th grader said:

I like Apatosaurus because I also have a fat body.

A 4th-grade girl who had strongly expressed herpreference for dinosaurs in general mentioned thatApatosaurus was her favorite species because of itsgenial nature:

Although Apatosaurus looks huge and strong,however it is genial and will not attack other species. I

feel its character is just like mine. I like herbivorousspecies of dinosaurs because they maintain a peacefulatmosphere.

Both students highlighted the features of the dino-

27

32

C.C. Chin

saurs, from either a morphological or psychologicalperspective, to justify their feelings. In the case ofTyrannosaurus, although it ranked third place on thelist of popular species, the reasons were quite differentfrom those for plant-eating dinosaurs. Students indi-cated that its strong body and fierce nature attractedthem. A male respondent in 6th grade expressed thatTyrannosaurus was his idol. He said:

I like strong figures. Tyrannosaurus represents sucha figure in my mind. Its mighty power attracts me. I alsolike its dominating nature. I look at Tyrannosaurus as a

model for me to emulate.

Interestingly, Tyrannosaurus was also named themost unpopular species of dinosaur by 218 studentsbecause its potentially human-eating and carnivoroushabits. The following response from a female 6thgrader represents the typical reasoning behind such anopinion:

I don't like Tyrannosaurus. It was fierce and likelyto mistreat other animals. Tyrannosaurus even used itspowerful teeth and claws to catch and beat the peacefulspecies. Lots of animals were killed by it. I don't likethe violence. That's why I don't like Tyrannosaurus,

either.

The number of students (N=218) who namedTyrannosaurus as the most unpopular is greater thanfor other species, for instance Triceratops (N=14),Stegosaurus (N=9), and Oviraptor (N=8), by a ex-tremely large margin.

From the species listed by the students, we alsofind some kinds of animals which do not belong amongdinosaurs. These species such as Ichthyosaurs,Plesiosaurus, and Pterosaur are usually misrecognizedas dinosaurs because of their similar appearance. Howchildren judge whether an animal is a dinosaur or notwill be discussed in the following section.

Some dinosaurs created by designers such asbubble dinosaurs and baby dinosaurs were also listedby respondents. This fact tells us children are notconfined by the scientific system in expressing theiropinions about such kinds of questions.

From the species names of the Most popular andunpopular dinosaurs listed by children, we may inferthat some species are more familial' to Taiwanesechildren. These "dinosaur stars" include Apatosaurus,Triceratops, Tyrannosaurus, Stegosaurus,Brachiosaurus, etc. Because of children's existingknowledge about dinosaurs, the species shown in Tables3 and 4 have a high degree of overlap.

3. One of the Reasons Children Are Fond of theDinosaurs Is Their Diversity in Featuresand Habits

As stated above, most students determined theirfavorite dinosaurs according to the nature and the habitof the dinosaurs. While interviewed, some childrenexpressed they were curious about the dinosaursbecause of their various appearances and habits. A6th grader said:

I was in touch with this subject since I was in kin-dergarten. At that time, I heard others talking about di-nosaurs. I felt that dinosaurs must be interesting andattractive. Thus I began to read books introducing dino-saurs. The more information I got, the more I liked dino-saurs.

When asked why he liked the dinosaurs so much, hereplied:

Because the dinosaurs are a group of animals. Thisgroup is composed of different species of animals. I foundthis very interesting.

When the interviewer asked about dinosaurs' hugebody size and tall figure, the student replied:

When I first studied dinosaurs, I didn't pay muchattention to their sizes. This is not the major reason I likethem. I think the fact that dinosaurs were animals existingin an ancient period aroused my curiosity. I love novelthings which I have never seen before, such as monstersin strange worlds.

According to his words, it was the life history andbiodiversity of the dinosaurs, not the image of theirstrong and tall bodies, that attracted his interest.Moreover, another 6th-grade student who had; ex-pressed in-depth thought about the dinosaurs in herinterview also described the diversity of dinosaurs interms of skin color, food sources, habitats, etc. toillustrate the fascination of the dinosaur group. Inparticular, she mentioned the value of dinosaurs inhelping her learn the concept of biodiversity throughcomparison:

28

By examining the species in the dinosaur group, Ifound they not only had similar characteristics to each other,

but also the specific features individual species. From such

examinations, I begin to realize that the dinosaurs were aninteresting group that could help me to understand thesimilarity and diversity of biological species.


A 4th grader also emphasized a similar point:

Some dinosaurs were tall; others were tiny. Theywere diverse in body size and height. I always thought that

the larger means stronger before, however, after I learnedsome information about the dinosaurs, I began to realizethat some small ones could defeat the bigger ones. Afterthat I felt that dinosaurs not only gave me a lot of newinformation, but also let me observe nature from the otherside.

This child studied dinosaurs in terms of theirinteractions and the diversity between species. Fur-thermore, she also began to understand the concept ofcompetition by means of special examples. Althoughdinosaurs are usually described as huge animals,according to the interviews, some students showedthey understand dinosaurs are a diverse group. Inconclusion, interesting features and mysterious leg-ends are the major motivatives for childern to learnabout dinosaurs.

4. Most Students Judged Dinosaurs Superfi-cially by Their Appearance, Not in a Sys-tematic and Scientific Manner

Although children could describe the specificfeatures of dinosaur that were familiar to them, theyusually could not illustrate common features that definethe whole dinosaur population. Some similar creaturessuch as Ichthyosaurs, Plesiosaurus, Pterosaur, andDimetrodon were usually mistakenly regarded asdinosaurs. When asked what are dinosaurs' commonfeatures, most of students answered "I don't know"without any hesitation. They usually emphasizeddinosaurs' diverse appearance and habits, but none ofthem could provide answers based on scientific defi-nitions. Some children even misinterpreted dinosaursas follows:

Most dinosaurs lived on the land; some inhabited inthe water; some could fly across the sky.

According to this description, some' flying andmarine reptiles existing in the age of the 'dinosaurswere apparently included in the "dinosaur family."However, according to children's answers,; only cer-tain flying and marine reptiles were categorized asdinosaurs. After the species most frequently proposedby children were checked, we found that species oftenintroduced to children along with the dinosaurs-suchas Ichthyosaurs, Plesiosaurus, Pterosaur, andDimetrodon-were most commonly the subject of such

misunderstandings. In addition, the morphology ofdinosaurs has an effect on children's judgements, too.Children usually judge if a animal is a species ofdinosaur roughly according to its appearance. Fewstudents paid much attention to accurate scientificways of defining dinosaurs.

5. What Factors Play Roles in CausingMisconceptions about Dinosaurs?

A. Translated Terms Make Children Mistakenly Con-sider Some Species as Dinosaurs.

In Chinese, dinosaurs are translated as OEfrom the word's meaning in English. However, intraditional Chinese legends there is also an artificialanimal species called the dragon [E]. Every speciesin the group of dinosaurs has a common name withthe Chinese Character MEdragon. However, fewchildren are confused by the relationship betweendinosaurs and the traditional Chinese dragon. Whenasked if a Chinese dragon was a dinosaur while beingshown illustrations, only 10.4% of the respondentsthough it was a kind of dinosaur and nearly 79% ofthe respondents stated it was not. However, 10.4%of the students had no idea about this.

Table 5 shows both the English and Chinesenames for some prominent species. No matter if theyare dinosaurs or other reptiles, all of the last charactersfor the species shown in this table are [it]. Not onlythe character [2], but also their similarity in theappearance make students include them in the dinosaurgroup.

B. Although Descriptions in the Texts of Labels in theDinosaur Gallery Provide Students Abundant Infor-mation about Dinosaurs, Some MisconceptionsAppeared in Some Students' Minds during theProcess of Informational Processing

Table 6 shows children's views about the size ofdinosaurs. 50% of the elementary students regarded

Table 5. English and Chinese Names of Some Typical Dinosaursand Dinosaur-like Reptiles

English Name Chinese Name Dinosaur or not

ApatosaurusTriceratopsTyrannosaurusStegosaurusPterosaurPlesiosaurusIchthyosaursDimetrodon

YesYesYesYesNoNoNoNo

29

C.C. Chin

Table 6. How Large Were the Largest and Smallest Dinosaurs?

Item Frequency Percentage (%)

How largeElephant 36 5.5A Classroom 10 1.52- to 3-Story Building 100 15.4Lower than a 10-story Building 114 17.5Taller than a 10-story Building 331 50.8Don't Know 60 9.2

How smallRhinoceros 225 34.6Pig 70 10.8Chicken 112 17.2Rat 148 22.7Smaller than a Rat 42 6.5Don't Know 54 8.3

the largest dinosaur as "taller than a ten-story build-ing". Over 30% of respondents thought the largestdinosaur was between 2- to 10-story building. Fewstudents give the answers such as "classroom-" or"elephant-size". On the other hand, the results forthe size of the smallest dinosaurs are more diversethan for the size of the largest dinosaurs. Many stu-dents chose rhinoceros-size (34.6%) and rat-size(22.7%). Only 112(17.2%) students selected the correctanswer chicken-size. From students' perceptionsabout their sizes, we might conclude that dinosaursare regarded as relatively large-sized animals.

When asked the reason for choosing "rat-size"as the answer in the interview, a male 6th grader whohas visited the dinosaur gallery of NMNS several timesstated:

I have read the text on the labels in the dinosaurgallery. I am sure the smallest dinosaur is calledCompsognathus. The text mentioned that Compsognathus

is also known as Mussaurus which means rat-like dinosaur.

That's why I am sure the size of the smallest dinosaur isjust like a rat.

According to his statement, although he answeredcorrectly about the smallest dinosaur, he was misledby the Chinese translation of Mussaurus-ME-whichmeans rat-like dinosaur.

On the other hand, in another case although astudent chose the right answer' for the size of thesmallest dinosaur, he did not name the right species.He said:

The smallest dinosaurs were like chickens. Theirname was Gallimimus (chicken-like dinosaur).

In Chinese, Gallimimus is directly translated as [41),(ft] (chicken-like dinosaur). Such a translation shown

in the exhibit text was supposed to transmit the messagethat Gallimimus looked like a chicken. However, somechildren were led to believe that the size of Gallimimuswas similar to that of a chicken: AlthoughCompsognathus and Gallimimus are nearby exhibitsin the dinosaur gallery of the NMNS, the texts on theirlabels easily cause visitors to become confused andcombine the information together.

In conclusion, those students who acquired moreinformation in the dinosaur gallery of the sciencemuseum may be assumed to be students who study hardand are active learners. Students had to be familiarwith the basic terminology used in the exhibit hallbefore they could make such mistakes. In case of theinterviewees mentioned above, their teachers com-mented that these students had strong motivation tolearn new things. However, it seems that for suchstudents the components which form ,the differentconcepts were reorganized and disconnected beforethe complete information was firmly fixed into theirlong-term memory.

Ill

C. Some of Children's Misconceptions about Dino-saurs Originate from the Blending Information fromthe Mass Media with Their Own Ima,Onations

Most students agreed that dinosaurs appearedbefore human beings were on earth (N=313, 78.8%).However, 5.8% of elementary students thought manappeared earlier; nearly 10% of respondents said di-nosaurs and man appeared during the same period(Table 7). Many students knew that dinosaurs wereanimals living on this planet long before human beingsappeared. For instance, a 4th-grade student who usedrelatively more scientific terms in his interview thanother students stated as follows:

The dinosaurs were extinct in the Cretaceous. Afterthat, man began to occupy the earth. So they did not overlap

Table 7. Who Appeared on Earth First? Dinosaurs or Man?

Item Grade levelTotal: N(%) 2 4 6

ManDinosaursContemporaneousDon't Know

38 ( 5.8) 6 ( 2.7) 12 ( 5.6) 20 ( 9.3)513 (78.8) 181 (81.9) 166 (76.9) 166 (77.6)

59 ( 9.1) 16 ( 7.2) 24 (11.1) 19 ( 8.9)41 ( 6.3) 18 ( 8.1) 14 ( 6.5) 9 ( 4.2)

X2=13.5, df=6, *P<0.05


30

35


with each other.

When asked how long was this gap, the student said:

About ten thousand years.

Although he knew the Cretaceous was an ancient periodof time, he did not realize how long ago this periodwas. In his point of view, ten thousand years repre-sented a period long enough to be the interval betweenthe first emergence of humans and the end of thedinosaur era. Judging from his description, this stu-dent might understand the interval was not short, butwas not certain about its exact length. It seemed thathis answer was merely based on a surmise from hisown limited information.

It is interesting that 46% of elementary studentsmentioned that ancient man had seen dinosaurs (Table8). Although most children put the appearance of manand the dinosaurs in order correctly, it seems that somethought that the dinosaurs and man existed contem-poraneously on earth for a period of time.

A male 2nd grader narrated a story that he hadread from a book:

Some ancient people were not afraid of dinosaur attacks.

They bravely fought with dinosaurs. They used tools madeby themselves to shoot the dinosaurs. However, the mencould not withstand the continuous assaults from the di-nosaurs. They were eventually beaten. Since then, men

began to fear and ran away from dinosaurs.

Other students also mentioned that they had seenman and dinosaurs appearing contemporaneously inthe same cartoon or picture. The story generallydepicted them fighting with each other. For instance,a first-grade student told the interviewer that he hadwatched Tyrannosaurus chasing people in the filmJurrasic Park while he was viewing the exhibit show-ing the attempt of Tyrannosaurus to catch Triceratops.Even from simulation games on computers, somestudents saw people and dinosaurs on the same screen.Men and dinosaurs are also found contemporaneouslyin books and drawings. All of these are usually fabulous

Table 8. Did Ancient Man See the Dinosaurs?

Item (%)

Yes 301 46.5No 236 36.5Don't Know 110 17.0


stories. However, some students, especially lowergraders, regarded the narratives in films, cartoons andpictures as real stories and adopted these as an elementof their perceptions about dinosaurs.

6. Major Sources of Information about Dino-saurs Changed Gradually from Lower toHigher Graders

According to the survey, grade is a factor influ-encing the source of elementary students' knowledgeabout dinosaurs (X2=35.3, df=8, p<0.001). Over 60%of respondents mentioned that the science museumwas a major source for acquiring information aboutdinosaurs (Table 9). Other sources include movies andtelevision (16.2%, N=106), newspapers/books/maga-zines (14.8%, N=96), and discussions with other people(3.6%, N=23). More 4th and 6th graders learned aboutdinosaurs from newspapers/books/magazines (17.6%and 18.5%) than 2nd graders (8.1%). On the otherhand, more 2nd graders (70.6%) regarded the sciencemuseum as a major informational source about dino-saurs than students in the other two grades. It seemsthat students in the higher grades obtain informationfrom more diverse sources than lower graders, andchannels other than the science museum become moreimportant for their informational acquisition. Whilegender does not have a significant effect on the sourceof knowledge about dinosaurs (X2=9.1, df=4, p>0.05),more girls learned such information from the sciencemuseum than boys. Among male students, the olderthe students the more they mentioned books and

Table 9. Major Sources of Knowledge about the Dinosaurs forElementary Students

Variable 1 2 3 4 0

N(%)

Grade2 18( 8.1) 32(14.5) 156(70.6) 9(4.1) 6(2.7)4 38(17.6) 31(14.5) 141(65.3) 5(2.3) 1(0.3)6 40(18.5) 43(19.9) 119(55.1) 9(4.2) 3(1.3)

X2=35.3, df=8, ***P<0.001

GenderMale 63(18.7) 56(16.6) 202(59.9) 12(3.6) 4(1.2)Female 33(10.6) 50(16.0) 213(68.3) 11(3.5) 5(1.6)

X2=9.1, df=4, P>0.05

Total 96(14.8) 106(16.2) 416(63.8) 23(3.6) 10(1.6)

Notes:1=Newspapers/books/magazines, 2=Television or movies,3=Science museum, 4=Communication with other persons,0=No response; N=Number of respondents

31

3 6

C.C. Chin

Table 10. Major Sources of Knowledge about the Dinosaurs forMale Elementary Students

Source Grade 2 4N (%) N (%)

6 TotalN (%)

Books, etc. 12(10.4) 23(20.3) 28(25.7) 63Films 18(15.7) 16(14.2) 22(20.2) 56Science Museum 79( 6.9) 70(61.9) 53(48.6) 202Other Persons 4( 3.3) 4( 3.6) 4( 3.7) 12No Response 2( 1.7) 0( 0.0) 2( 1.8) 4

Total 115(100.0) 113(100.0) 109(100.0) 337

X2=23.4, df=8, "P<0.01


magazines as a major source; on the other hand, thescience museum was a less important source of dino-saur-related information for them (X2=23.4, df=8,p<0.01) (Table 10).

Some children noted that they began to learnabout dinosaurs during an early stage of their elemen-tary school years when they were attracted by theseinteresting animals. Moreover, parents usually playedthe role facilitators when their children were learningabout dinosaurs. A 6th grader recalled the ways herparents helped her acquire information about dino-saurs when she was a lower grader:

After my parents discovered my interest in dinosaurs,

they encouraged me by bringing me to the dinosaur galleryin the science museum, selecting and purchasing dinosaur-related reading materials such as books, and using maga-zines or films to introduce me to the dinosaurs.

The probable reason parents assisted their chil-dren so actively is that they thought of dinosaurs aspart of scientific knowledge; they felt their childrencould nurture a good basis and positive attitude to-wards science learning by learning about dinosaurs.However, according to the findings of this study,children's interest in studying dinosaurs declined afterentering higher grades. One of the major reasons isthat older elementary students have become moreinterested in other novel things. A male elementaryschool science teacher who helped interview the students tried to explain such changes based on hisobservations and his experience with children:

Their work load also influenced children's willing-ness to keep their focus on dinosaurs. Their teachers andparents begin paying more attention to the students' aca-demic achievements. Therefore they no longer have timeto study subjects, such as dinosaurs, that are unrelated to

their classes. ...Moreover, parents usually arrange lots oflearning activities such as Chinese composition, Englishconversation, and music lessons etc. after class for theirkids. The fact that many skills must be learned directsstudents' attentions to a relatively broader field.

In this case, parents are no longer facilitators fortheir children to learn about dinosaurs. On the con-trary, their role becomes that of an "inhibitor" as theirchildren grow up.

7. Perceptions of Elementary Students Towardsthe Existence of Dinosaurs

Most of the children believed that dinosaurs donot exist it the present world: All kinds of dinosaursare extinct (Table 11). However, 7.5% of respondentsthought that dinosaurs are still alive on earth. About13.5% of students possess other naive perceptionsabout the existence of the dinosaurs. Such perceptionsinclude: (1) Dinosaurs are mythical animals, (2) di-nosaurs are man-made creatures in from films, and (3)dinosaurs are artificial models or toys for enjoyment.Among these ideas, the dinosaurs as models or toyswas chosen by the fewest students.

Students at different grade levels also had sig-nificantly different perceptions about these proposi-tions (X2=50.9, df=10, p<0.001). The fact that dino-saurs are extinct was recognized by 83.8% of the 4thgraders. In comparison, about 70% of the 2nd and 6thgraders hold such a belief. Moreover, more 6th graders

Table 11. Elementary Students' Perceptions about the CurrentExistence of Dinosaurs

Variables1 2 3 4 5 0

N (%)

Grade2 161(72.9) 7( 3.2) 12(5.4) 28(12.7) 7(3.2) 6(2.6)4 181(83.8) 12( 5.6) 10(4.6) 9( 4.2) 2(0.9) 2(0.9)6 158(73.8) 30(14.0) 10(4.7) 8( 3.7) 2(0.9) 6(2.9)

X2=50.9, df=10, ***P<0.001

GenderMale 254(75.4) 36(10.7) 14(4.2) 16( 4.7) 7(2.0) 10(3.0)Female 245(78.3) 13( 4.2) 18(5.8) 29( 9.3) 4(1.2) 4(1.2)

X2=19.1, df=5, "P<0.01

Total 500(76.8) 49( 7.5) 32(4.9) 45( 6.9)11(1.7) 14(2.1)

Notes:1=Extinct animals, 2=Still alive on earth, 3=Mythical ani-mals, 4=Fabulous animals in films, 5=Artificial models ortoys, 0=No response; N=Number of respondents

32


Table 12. Perceptions of Elementary Students about the CurrentExistence of Dinosaurs

Variables2 3 4 5 0

MaleGrade

246

88(76.5) 5( 4.3) 5(4.3) 12(10.4) 3(2.6) 2(1.7)93(82.3) 9( 8.0) 5(4.4) 2( 1.8) 2(1.8) 2(1.8)73(67.8) 22(20.2) 4(3.7) 2( 1.8) 2(1.8) 6(5.5)

X2=40.2, df=10, ***P<0.001

FemaleGrade

2

4

6

73(68.9) 2( 1.9) 7(6.6) 16(15.1) 4(3.8) 4(3.8)87(85.3) 3( 2.9) 5(4.9) 7( 6.9) 0(0.0) 0(0.0)85(81.0) 8( 7.6) 6(5.7) 6( 5.7) 0(0.0) 0(0.0)

X2=29.7, df=10, "P<0.01

Notes:1=Extinct animals, 2=Still alive on earth, 3=Mythical ani-mals, 4=Fabulous animals in films, 5=Artificial models ortoys, 0=No Response; N=Number of respondents

(14%) considered dinosaurs to be living animals thanstudents in the other two sampled grades. Over 12%of the 2nd graders held naive ideas such as dinosaursbeing merely creatures existing in films. If thethree kinds of naive ideas are combined, they areheld by a relatively higher proportion of 2ndgraders (21.3%) than 4th (9.7%) and 6th (8.3%) grad-ers.

If the two genders are considered separatelythrough the three age groups, significant differencesexist not only in the male (X2=40.2, df=10, p<0.001)but also the female group (X2=29.7, df=10, p<0.01)(Table 12). In both male and female groups, more4th grade students regarded the dinosaurs as extinctcreatures than those in the other two grades. Inaddition, more 6th graders think dinosaurs are still

Table 13. The Perceptions of 6th-Grade Students about the Cur-rent Existence of Dinosaurs

Items Male FemaleN (%) N (%)

Extinct animalsStill alive on earthMythical animalsFabulous animals in filmsArtificial models or toysNo Response

73(67.0) 85(81.0)22(20.2) 8( 7.6)

4( 3.7) 6( 5.7)2( 1.8) 6( 5.7)2( 1.8) 0( 0.0)6( 5.5) 0( 0.0)

X2=21.2, df=5, ***P<0.001

alive on earth for both girls and boys. The proportionof male 6th graders with such a belief exceeds 20%.Compared with the other two grades, there are moremale and female 2nd graders with the belief thatdinosaurs are fictitious animals existing only in films.For the 2nd graders, the proportion of girls with naiveideas was 25%, but that of the male group was only17%.

If we consider the effect of gender for eachsampled grade, a significant difference exists only inthe 6th grade for belief in the survival of dinosaurs(X2=21.2, df=5, p<0.001) (Table 13). Over 80% offemale 6th-grade students believe that dinosaurs areextinct. Only 67% of the boys in the same grade holdsuch a belief. But for another item, relatively moreboys in the 6th grade believe dinosaurs are still aliveon earth. There is a 12.6% discrepancy between thetwo sexes for this question.

According to the interviews, some students in thelower grades regarded some existing reptiles such ascrocodiles, turtles, etc. as the descendants of the ancientdinosaurs. Based on biologists' research, this is amisconception about the classification of reptiles.However, those 6th graders who indicated dinosaurswere still alive on earth held relatively more thought-ful arguments. Some of them stated that dinosaurs arenot extinct because they evolved into the birds. Othersmentioned that certain monsters reported in the massmedia might be the descendents of the dinosaurs. Amale 6th grader's explanation represents the firstposition:

I think dinosaurs theoretically still exist on earth.From science books, I learned birds evolved from dinosaurs.

Based on such evidence, dinosaurs exist on earth in the other

forms.

Another view held by children about the exist-ence of the dinosaurs is that they believe some reportedmonsters might be the living dinosaurs. A child toldthe interviewer:

The monster described by the visitors and residentsnear Loch Ness is one kind of dinosaur. A few people saidthey have seen it. Its picture was taken, too. From thepicture, it has a long neck; however, the large portion ofits body is in the water. Judging from the scale, I estimateit is very large.

Another student said:

I believe there are dinosaurs still alive. Although

Note: N=Number of respondents no direct evidence has been provided, a lot of reports tell

33

38

C.C. Chin

us about the existence of the dinosaurs. Since human being

cannot explore every corner of the earth, we cannot to be

sure obscure creatures are not dinosaurs.

Both students are 6th graders. The possession of moreinformation related to but not confined to the subjectof dinosaurs apparently led them to construct their ownlogical explanations about the existence of dinosaurs.Such an effect existed especially for students in thehigher grades. Their knowledge about fabulous mon-sters or obscure creatures makes them think somedinosaurs still exist on Earth.

8. Most of Reasons for the Extinction of theDinosaurs Given by Elementary StudentsWere Connected with EnvironmentalChanges

Reasons for the extinction of the dinosaurs givenby elementary students include:

(1) a hit on Earth by a meteorite,(2) the eruption of volcanoes,(3) earthquakes,(4) lack of food,(5) environmental changes: cold weather, ...,(6) poisoning from eating harmful plants,(7) failure to compete with the mammals,(8) continental drift,(9) emigration to other planets,

Children usually gave more than one reason toexplain why the dinosaurs were extinct. Students inthe higher grades or who possessed more informationabout dinosaurs were more able to explain the reasonsfor the extinction of the dinosaurs by organizing someof the above factors in a more logical way. On theother hand, younger students revealed their relativelynaive ideas when they explained why the dinosaursdid not survive on earth. The logical connectionsbetween the factors mentioned in their explanationswere less well-developed. For instance, when askedwhy there were no dinosaurs in the current age, a first-grade student who had visited the dinosaur gallery ofthe science museum accompanied with his parentsexplained:

All of the dinosaurs were dead so there are no di-nosaurs left on earth. At first volcanoes erupted. A lotof hot rocks were spilled. After the eruptions, it began to

snow. The dinosaurs could not touch the snow because it

was extremely cold. Lots of dinosaurs were frozen to death.

Only if a typhoon came, the snow would be blown far away

from the area. Therefore, snow would not cause the ex-

tinction of the dinosaurs. Besides, the heat from the erup-

tion of the volcanoes also made the snow melt. Thus, there

was no disaster for the dinosaurs.

The fact that this student connected the eruptionof volcanoes with snowing is no consistent with thecause-effect relationship. In his point of view, weatherchange induced a difficult environment for the dino-saurs. However, the cause of cold weather was quiteirrelevant to the real phenomena. This student usedsnow caused by the volcanic eruption to explain whythe cold weather occured. The various kinds of naiveexplanations were more easily discovered in the in-terviews. For instance, a second-grade student indi-cated that the dinosaurs became extinct because theydirectly touched hot materials that had erupted fromthe volcanoes. Although both students mentionedvolcanoes, one thought cold and another thought hotwas the immediate factor that killed the dinosaurs.

Compared with the younger students, the expla-nations provided by higher graders were more science-based and conformed better with logical relationships.For example, a fifth-grade student interviewed in thedinosaur gallery also mentioned volcanic eruptions asthe major reason, but his reasoning was more logical:

The dust from the eruption of the volcano covered

the atmosphere. The clouds were so thick that sunlight was

not able to penetrate. Consequently, the plants growing on

the surface of the planet could not absorb enough lightenergy to produce nutrition. Therefore these plants failed

to survive. The herbivous dinosaurs were affected and died

first. Other dinosaurs with meat-eating habits died because

there were no living animals as food. All of the dinosaurs

died in the long run.

This student adopted the concept of the food chain asthe basis for his explanation. Although also mention-ing clouds in the atmosphere, another student indi-cated a different sourcethe impact on the earth bya meteorite from outer space. But he shortened thetime period needed for the extinction to occur. Hesaid:

The earth was suddenly hit by a comet. And then

the dust rose from the surface of the earth. The planet was

shaded by the thick cloud layer formed from the dust. Such

a cloud layer lasted for several months. All of the dinosaurs

finally became extinct because they could not endure theconsequent humid and dark weather.

The following responses can be also provided to helpus understand children's tendency to simplify both theelements and the time scales of the story:

34

3 9


(1)The dinosaurs were killed by falling meteorites.(2)The dinosaurs were killed by falling materials caused

by earthquakes.(3)The dinosaurs were scorched to death by boiling sub-

stances from volcanoes.(4) The dinosaurs could only survive in warm areas. If they

moved to the south pole, they would die. Actually thedinosaurs were carried to the arctic zone in the ice agebecause of continental drift. Finally, all of the dinosaurs

were frozen to death.

Judging from the reasons provided by students,it is the life history of the individual, not the wholepopulation, that serves as the basic knowledge for theirexplanation of the extinction of the dinosaurs. Ex-amining the way they answered this unsolved scien-tific problem, we may conclude that most elementarystudents do not have explicit prior knowledge aboutthe difference between the life history of individualorganisms and of a whole population of a species. Thestudents are inclined to shorten the time needed forthe whole event, simplify or delete some of the steps,and/or connect the possible reasons in an illogical way.Besides their level of mental development, the limitedinformation possessed by the kids is also believed toplay role in this.

VI. Discussion and Recommenda-tions

The task of studying children's perceptions and/or misconceptions about a specific scientific conceptinvolves two perspectives. First, in a relatively short-term perspective, the findings could be regarded as abody of descriptive findings elicited and categorizedby science educators. Such findings provide informa-tion about children' s understandings of scientificconcepts. In the second perspective, the findings mayserve as a foundation for further prescriptions. Forclassroom teachers, the development and adoption ofappropriate teaching strategies should be based onwhat is known about the students. However, as edu-cators in the science museum, the possession of knowl-edge about the students' responses toward certain topicswould allow the curators to consider what role theymay play in the collaborative relationship with theschool system. But if the scientific concepts studiedare directly based on the exhibits themselves, theresults could serve as a basis for the work of designersand educators in the museum. Therefore, not only isan evaluation indispensible to the process of devel-oping exhibitsthe so-called formative evaluationbuta summative evaluation conducted by the curators is

required for refining permanent exhibits. For scienceeducators in the broad sense, the results of the evalu-ation provide knowledge to help understand students'science learning. However, on the other hand, if thedata is analyzed from the standpoint of the museumeducators, the prescriptive function for the educationalfunction of the museum may be emphasized.

The dinosaur gallery is one of the most focalpoints for children visiting NMNS. Most of the stu-dents participating in the study indicated that thedinosaur gallery in NMNS was their main source ofinformation concerning dinosaurs. However, unlikeother countries, no dinosaur fossils have been foundin Taiwan. Therefore, the permanent exhibits in thedinosaur gallery are not real fossils. But when weexamined the level of the students' interest in dino-saurs, the response of youngsters in Taiwan was foundto be at quite a high level, even if dinosaurs are notintroduced well in their elementary school textbooks.It seems that the lack of direct experience of realfossils is somehow replaced by other experiences suchas exhibits in the form of miniatures, models, films,simulation games, and demonstration boards and alsoinformation provided by the mass media and personalcontact. The results of this study have supported

-Solomon's arguments (Solomon, 1987) concerningsocial influences on children's science learning.

Based on the results of this study, the followingrecommendations may help both school and museumeducators consider ways for applying the topic ofdinosaurs to science teaching in the school classroomand museum exhibit halls.

(1) Due to the interest expressed by the majorityof elementary students, the topic of dinosaurscould be used as a vehicle and motivationaldevice for children to understand the biologicaldiversity of the Earth, the impact of the envi-ronment on living organisms, the reasons forextinction and the mechanism of competition.If approached correctly, the study of dinosaurscould become one of the best topics for intro-ducing children to the fascinating world ofnature.

(2) Children acquired their knowledge of dinosaursthrough a variety of the channels outside theclassroom. Most of them, constructed knowl-edge of dinosaurs that is incompatible with whatscientists have reported. It is suggested thatteachers mention this topic while teaching re-lated topics in the classroom and introduce itin a systematic way to help students correct theirown concepts.

(3) Children's perceptions about dinosaurs come

35

40

C.C. Chin

from several channels. Certain channels prima-rily serve entertainment purposes rather thanachieving educational goals. Moreover, somepublished books introducing dinosaurs are foundto contain mistakes. In this situation, the imagesof dinosaurs formed by children are more or lessdivergent from the scientific view. For in-stance, Plesiosaurus, which is not dinosaur, isfound include as one of the typical examplesdinosaurs in some publications for children.There is no doubt children will gain incorrectinformation if they read such books. Becausethe editors of children's books don't necessarilyhave a science background, it is suggested thatbooks being published should be examined byexperts on dinosaurs.

(4) More students in the lower grades mentionedthe science museum was their main source forobtaining information about dinosaurs. Com-pared with younger students, more older stu-dents learned about dinosaurs by reading ma-terials such as books and magazines, etc. It issuggested that awareness and interest should beincreased in appropriate ways for younger stu-dents. In contrast, the provision of more infor-mation about dinosaurs should be emphasizedin the higher grades.

(5)The content of exhibit labels in the dinosaurgallery is one of the sources of informationabout dinosaurs for elementary students. How-ever, some Chinese terms translated from thescientific names are likely to mislead students'understanding of the dinosaur. Based on sucha finding, we suggest that the Chinese namesof some dinosaurs need to be re-considered toavoid causing misunderstandings.

(6) In order to help students construct knowledgeof dinosaurs based on scientific findings, bothteachers and parents should be encouraged tovisit the dinosaur gallery of the science museumtogether with their children. It is also suggestedthat exhibits, models, simulation games, films,multimedia, and the services of interpretersshould be utilized in a systematic way.

According to Lucas (1983), most of people learnoutside the school system most of their lives. Alongwith the concept of life-long learning, informal edu-cational resources will play a more important role inlearning in the future. Especially for schoolchildren,the opportunity to explore the science museum willhelp them to nurture interest in and motivation forlearning and utilizing social educational resources froman early stage of life. Tamir's study (Tamir, 1990)

also indicated that field trips to informal institutionssuch as museums, zoos, and outdoor areas have apositive effect on students' learning. As a focal pointfor visits, the dinosaur gallery could be used to re-inforce children's recognition of certain biologicalconcepts, and also provide informal educational re-sources for long-term utilization over a lifetime. Sincedinosaur concepts are not introduced well in text-books, children's perceptions of dinosaurs are notnarrowed by the scope of formal education. Under-standing of the freedom of children' thoughts on thistopic may give us an opportunity to re-consider thegoals of science teaching and the ways we achievethem.

Acknowledgments

This study was financially supported by the National ScienceCouncil (NSC 84-2511-S-178-002). I would like to express myappreciation to Dr. Hsiao-lin Tuan, Dr. Francis Yu-Teh Chang, Ms.Mann-hsi Tso, Ms. Shu-Yun Chung, Ms. Chia-Yu Hsi, Mr. Min-Ren Wang, and Mr. Yi-Long Chen. Without their support, this studycould not be completed within such a limited period.

References

Barba, R. H. (1995). Children's tacit and explicit understandingsof dinosaurs. Paper presented in 1995 Annual Meeting andConference for National Association of Research in ScienceTeaching, San Francisco, CA.

Benton, M. (1992). Dinosaur and other prehistoric animal factfinder,New York, NY: Grisewood & Dempsey Ltd.

Bernfeld, D. (1993). The participative museum. Museum Inter-national, 179, 50-52.

Boisvert, D. L., & Slez, B. J. (1994). The relationship betweenvisitor characteristics and learning-associated behaviors in ascience museum discovery space. Science Education, 78(2),137-148.

British Museum of Natural History (1979). Dinosaurs and theirrelatives. Cambridge: Cambridge University Press.

Chi, M. T. H., & Koeske, R. D. (1983). Network representationof a child's dinosaur knowledge. Development Psychology, 19,29-39.

Dai, M. F. Wang (1995). Young Chinese children's conceptionsof life. Paper presented in 1995 Annual Meeting and Confer-ence for National Association of Research in Science Teaching,San Francisco, CA.

Diamond, J. (1994). Sex difference in science museum: A review.Curator, 37(1), 17-24.

Dierking, L. D., & Falk, J. H. (1994). Family behavior and learningin informal science settings: A review of the research. ScienceEducation, 78(1), 17-24.

Finson, K. D., & Enochs, L. G. (1987). Student attitudes towardscience-technology-society resulting from visitation to science-technology museum. Journal of Research in Science Teaching,24(7), 593-609.

Gardom, T., & Milner, A. (1993). The natural history museum bookof dinosaurs. London: The Natural History Museum.

Gil, D., & Carrascosa, J. (1990). What to do about science "mis-

36

41


conceptions". Science Education, 74(5), 531-540.Hsiung, C. T., Chang, J. Y., & Hsiung, T. H. (1996). A study of

Taiwan elementary students' understanding of ecological sta-bility. Paper presented in 1996 Annual Meeting and Conferencefor the National Association of Research in Science Teaching,St. Louis, MS.

Hood, M. G., & Roberts, L. C. (1994). Neither too young nor tooold: A comparison of visitor characteristics. Curator, 37(1),36-45.

Huang, D. S. (1996). A study of children's conceptions of life,animals, and plants as well as their alternative conceptions.Proc. Natl. Sci. Counc. ROC(D). 6(1), 39-46.

Lambert, D., & the Diagram Group (1990). The dinosaur data book,facts and fictions about the world's largest creatures. NewYork, NY: Avon Books.

Lewenstein, B. V. (1993). Why the "public understanding of thescience" field is beginning to listen to the audience. Journalof Museum Education, 18(3), 3-6. .

Lucas, A. (1983). Scientific literacy and informal learning. Studiesin Science Education, 10, 1-36.

Norman, D. (1985). The illustrated encyclopedia of dinosaurs.Salamander, CA: Salamander Books Ltd.

Norman, D. (1991). Dinosaurs! Londong: Boxtree Ltd.Rowell, J. A., Dawson, C. J., & Lyndon, H. (1990). Changing

misconceptions: A challenge to science educators. EuropeanJournal of Science Education, 12, 167-175.

Screven, C. G. (1993). Visitor studies: An introduction. MuseumInternational, 178, 4-5.

Silverman, L. (1993). Making meaning together. Lessons fromthe field of American history. Journal of Museum Education,18(3), 7-11.

Solomon, J. (1987). Social influences on the construction of pupils'understanding of science. Studies in Science Education, 14,63-82.

Tamir, P. (1990). Factors associated with the relationship betweenformal, informal and nonformal science learning. Journal ofEnvironmental Education, 22(2), 34-42.

Wu, P. (1993). The rationality model and students' misconceptions.Unpublished doctoral dissertation. University of Texas at Austin,Austin, TX.

Yu, S. M. (1995). Elementary students' conceptions of ecology.Paper presented in 1995 Annual Meeting and Conference forNational Association of Research in Science Teaching, SanFrancisco, CA.

/1\ **II Fel_4111011 iRict 4:0112...VM

*ROM

-f-tt Aft

ril N

itiA-mmtlfanat ; HI'A.stir4Itt4q=, 0-VIATAMILVOUN77.-gi5FIYErtglAS ILUAbit4IINZ*I&AftlElgrf-TZA=1 40*-Vtititt-F-3-1-AIR tfitniwcwmattu

RiVVJASI WrinIPX4gMfrItA2Sffitl , gt141\f-IPIZEMMStfitaffrtANVII-ArOlititt*-tfl--)A101%* MA-M4nZ10{1 -

*WE ANY-MAI /Jvf-tttbiEZtH EOM ° MR111 ol=i-gtazmegt= AVA*-tAtURAMONXWit**4 Fr-taiNE 4\N A-I fUtbilM,I-04A-NA-XA- ;

-1-#Mfin4iNtittt. AzPiJfAillAssE11-4-3-1YIJ*T4AVIIIMM 1)-(SoLf_4-#X01,441401X ° AltatriumamtvAnzatuRNA.E4c-.-o4Acx 1444WittififflOrA3OPAVit #tZfiR4Eil4J1-8- 1).(ScAtOtnwiti4mH

37

42

Proc. Natl. Sci. Counc. ROC (D)V ol. 7, No. 1, 1997. pp. 38-45

The Relationship between Biology Cognitive Preferencesand Science Process Skills

YEONG-JING CHENG*, JAMES A. SHYMANSKY**, CHIOU-CHWEN HUANG***, AND BIH-Ju LIAW*

*Department of BiologyNational Taiwan Normal University

Taipei, Taiwan, R.O.C.**Science Education Center

The University of IowaIowa City, IA, U.S.A.

***Taiwan Provincial Chung-Ho Senior High SchoolTaipei County, Taiwan, R.O.C.

(Received July 16, 1996; Accepted September 24, 1996)

ABSTRACT

This study employed the Test of Biology Cognitive Preference (TBCP) to determine the biologycognitive preference styles of 1,339 seventh grade students in the Taipei area of Taiwan. The relationshipbetween biology cognitive preferences and achievement in science process skills was then explored.

The results showed that the subjects exhibited a cognitive preference style of Questioning (Q)> Principles (P), Applications (A) > Recall (R). This indicated that the subjects displayed the highestpreference for questioning and the lowest preference for memorizing biological information presentedto them.

Performance in science process skills was measured by the Test of Science Process Skills (TSPS).Results from TBCP and TSPS were analyzed using a correlation technique. TSPS scores correlatedpositively with Q preference scores, but negatively with R preference scores in biology (p<0.01). However,no significant correlations were found between performance in science process skills and scores for theP and A preference modes.

When the subjects were categorized into three different achievement levels, according to the TSPSscores, namely as high (H), medium (M), and low (L) achievers, oneway ANOVA revealed significantmain effects for biology cognitive preferences across the three groups of subjects.

A strong preference for questioning and a weak preference for recall of biological informationwere,associated with high achievement in science process skills. On the other hand, strong preferencesfor application and principles and weak preferences for recall and questioning were associated with lowachievement in science process skills. Generally, it was evident that the biology cognitive preferencestyles exhibited by high achievers were significantly different from those exhibited by low achievers.

Key Words: biology cognitive preference, science process skills

I. Introduction

The construct of cognitive preferences was firstintroduced by Heath (1964) as an alternative way toevaluate the effectiveness of science curriculum. Theidea was based upon the hypothesis that, whenlearning science, students not only acquire scientificknowledge, but also develop particular modes ofpreference in dealing with scientific information.Consequently, the science cognitive preferences ex-hibited by students in inquiry-oriented and conven-

38

tional science curricula are expected to be different.Students using a more inquiry-oriented science cur-riculum should exhibit a higher preference for prin-ciples and questioning compared to those who use amore traditional science curriculum (Heath, 1964;

Tamir, 1977).

Heath (1964) proposed four different modes ofcognitive preference in science, namely, Recall (R),Principles (P), Questioning (Q), and Application (A).Based upon Heath's definitions, van den Berg (1978)described the four cognitive preference modes as

4 3

follows:Factual Information or Recall (R)A person responding in this mode likes to storeor memorize scientific information in the sameform it is presented. The "storing" is withoutany processing such as trying to relate infor-mation to other information, or abstracting ageneral rule or principle from specific data, orquestioning the validity of the information. Apreference for recall indicates interest in learn-ing terms, definitions, facts, observations, orconcepts.

(2) Principles (P)A peron responding in this mode likes to extractgeneral principles or rules from the informa-tion presented, or looks for interrelationshipsbetween scientific information. A preferencefor principles indicates interest in identifyingrelationships between variables, in learningrules that can be applied to a class of objects,organisms, phenomena or variables, or in ex-plaining phenomena in terms of general prin-ciples or theories.

(3) Questioning (Q)A person responding in this mode likes to evalu-ate scientific information critically, looking forovergeneralizations and limitations. A prefer-ence for questioning indicates interest in criti-cally palyzing and commenting on the validityof scientific information, and/or in generatingsuggestions and hypotheses for further research.

(4) Applications (A)A person responding in this mode tends toevaluate scientific information in terms of itsapplicability. A preference for applicationsindicates interest in using scientific informa-tion to solve problems in commerce, industry,farming, daily life, or science.

Cognitive preference tests based upon Heath' snotion typically consist of items including an intro-ductory statement followed by four alternative ex-tended statements, one for each of the four preferencemodes. The examinees are told that all the four al-ternatives are correct and are asked to rank or rate thefour alternatives according to their degree of prefer-ence. Since Heath proposed his idea of cognitivepreferences, quite a few instruments purported to mea-sure science cognitive preferences have been devel-oped. It was evident that cognitive preferences inscience could provide useful information about stu-dents' thinking that might better inform biology teach-ers' planning and teaching.

The importance of cognitive preference assess-

(1)

Cognitive Preference & Science Process Skill

ment lies, as suggested by Heath (1964), not in whetherthe student can identify correct information, but ratherin what (s)he is likely to do with information intel-lectually. Therefore, a cognitive preference test doesnot attempt to measure students' achievements, butrather attempts to adequately reveal students' particu-lar modes of thinking or preferences in dealing withscientific information presented to them. Suchepistemic information can be useful in structuringlearning experiences and addressing specific abstractor practical topics.

1. Importance of the Study

Since the 1970s, major science education reformsin Taiwan have transformed science learning frommerely memorizing scientific facts and concepts tofostering understanding of fundamental principles anddeveloping intellectual skills related to the acquisitionand evaluation of scientific knowledge. Hence, thebiology curriculum currently implemented in juniorhigh schools places more emphasis on the learning ofscientific principles and developing inquiry skills thandid the former biology curriculum. When this cur-riculum attains its intended goals, then the studentswill have developed basic science process skills inaddition to the learning of biological concepts, andexhibit a preference for analyzing scientific informa-tion critically rather than merely memorizing scien-tific knowledge. Unfortunately this assumption hasnot been formally tested. Exploration of the modesof preference in cognition that the students exhibit inprocessing scientific information and of the relation-ships between students' science cognitive preferencesand achievements in science are reasonable first stepsin testing the assumption.

Within the last two decades, research has indi-cated that students' cognitive preferences may varyfrom subject to subject, and even from topic to topicwithin a subject (Cheng, Huang, Tsai, & Liaw, 1993a;Rost, 1983; Tamir, 1975; Tamir & Jungwirth, 1984;Tamir & Lunetta, 1978). A recent study done inTaiwan (Cheng, Huang, Tsai, & Liaw, 1993b) hasshown that seventh grade students, after studying the"new" biology curriculum, exhibited a strongpreference for the P mode and a weak preference forthe R mode. The study also indicated that there weresignificant differences in cognitive preferences be-tween males and females, and among students indifferent types of schools. Some results of the studywere different from what has been found in studiesdone in other countries (Tamir, 1985). What thismeans and how cultural and educational backgrounds

39

4 4

Y.J. Cheng

contribute to the development of cognitive preferencesin students are questions that need to be exploredfurther.

The relationship between cognitive preferencesand achievement in science has also been a major focusof investigation in previous researches. Althoughmany studies have shown that there was a close re-lationship between cognitive preferences and achieve-ment in science (Tamir, 1976, 1977, 1988; Tamir &Kempa, 1978), their results were relatively diverse.Several studies found that science achievement cor-related positively with the Q mode, and negativelywith the R mode (Cheng & Yang, 1995; Kempa &Dube, 1973; Tamir, 1988; Tamir & Yamamoto, 1977).Others showed that while biology achievement cor-related positively with the A mode, no significantcorrelation was found between biology achievementand the Q mode (Barnett, 1974).

Although studies have attempted to investigatethe relationship between cognitive preferences andseveral aspects of achievements in science, very fewhave focused on science process skills achievement.In a study utilizing ninth grade students as subjects,Atwood & Stevens (1978) found that students with astrong preference for the A mode showed greaterachievement gains in science process skills than stu-dents with a weak preference for the A mode. How-ever, the conclusions of their study were neither com-prehensive nor definitive. Since science process skills,such as interpreting data and formulating hypotheses,constitute an essential part of inquiry-based logicalthinking and reasoning, it is anticipated that studentsshould develop these skills in their learning of science.Hence, whether or not substantial associations existbetween students' cognitive preferences and the learn-ing of science process skills is an issue that needs tobe further investigated.

2. Purpose of Study

The purpose of this study was to determine thebiology cognitive preference styles of seventh-gradestudents in the Taipei area of Taiwan, and to explorehow the students' cognitive preferences related to theirlearning of science process skills.

II. Methods

1. The Subjects

The subjects consisted of 1,339 seventh-gradestudents selected from the Taipei area (Taipei City andTaipei County) employing stratified random sampling

and cluster sampling methods. Firstly, the junior highschools in Taipei City and Taipei County were dividedinto three types according to their size, i.e., totalnumber of classes. The three types of schools weretermed large (above 76 classes), medium (31-75classes), and small (below 30 classes) schools. Then,an adequate number of schools and classes weredrawn from the three types of schools based on theratio of the total number of classes in each type ofschool. All of the students in the classes selected werethe subjects of this study. A total of 1,339 studentswere selected from three large, two medium and twosmall schools.

2. Assessment of Biology Cognitive Preference

The Test of Biology Cognitive Preference (TBCP),developed by Cheng et al. (1993b), was employed toassess the biology cognitive preference styles of thesubjects. The TBCP consisted of 32 items. Each itemwas composed of a statement and four optional ex-tended responses, each of which represented one ofthe four cognitive preference modes. In order toensure that the subjects were familiar with the infor-mation contained in the TBCP, the construction of theitems was based exclusively upon the current junior-high biology textbook.

The reliabilities of the TBCP, as reported in thevalidation study, were satisfactory for the purpose ofthis study. The internal-consistency reliability coef-ficients were 0.86, 0.79, 0.90, and 0.67 for the R(Recall), P (Principles), Q (Questioning), and A (Ap-plication) modes, respectively. It was also evident thatthe TBCP possessed fairly satisfactory content andconstruct validity (Cheng et al., 1993b).

The TBCP was administered to the subjectsipsatively at the end of the second semester. Thesubjects were informed that all of the four options werecorrect and were asked to rank the four options ac-cording to their relative preference. The answer sheetswere scored using a data transformation program whichallotted "4" to the most preferred response, "3" and"2" to the next preferred, respectively, and "1" to theleast preferred response.

3. Assessment of Science Process Skills

The science process skills of the subjects weremeasured by the Test of Science Process Skills (TSPS).The TSPS consists of two types of items: 17 multiple-choice items and 13 short constructed-response items.Among them, 4 items were developed by the author,and 26 items were selected, adapted and intensively

40


modified from constructed-response type items origi-nally developed by Mao (1988, 1989) for measuringinference, prediction, and data interpretation skills.

Since the TSPS was recomposed from parts oftests developed previously, it was imperative for thisstudy to examine the reliability of the test. In thisstudy, the internal-consistency reliability (Cronbachalpha) was 0.80 and the test-retest reliability was 0.84(Table 1). This reliability was considered satisfactoryfor the purposes of this study. The content validityof the TSPS, which was established by an expert panel,was satisfactory.

Ill. Results and Discussion1. Reliability and Interrelationship among

Cognitive Modes

There are two different ways reported in theliterature for a person to respond to a science cognitivepreference test. They are ipsative and normativemethods (Tamir & Lunetta, 1977). Although theipsative method has been questioned regarding thevalidity of the data obtained, Tamir & Lunetta (1977)provided strong support for the use of the ipsativemethod in science cognitive preference research. Theyfurther indicated that the ipsative method discrimi-nated more clearly between different cognitive pat-terns are that the danger of distorted relationships isnot as severe as might be expected.

In the present study, the TBCP was administeredto the subjects ipsatively (ranking). Data analysisindicated that the Cronbach alpha reliability coeffi-cients of the four preference modes were satisfactory(Table 2), and were consistent with what has beenfound in previous studies of 7th graders (Cheng et al.,1993b; Cheng & Yang, 1995).

The intercorrelations of scores of the four cog-nitive preference modes are also presented in Table2. Since the data were ipsative in nature, an individual' sscores on the four preference modes were interdepen-dent. Students who displayed a high preference fora particular preference mode would naturally displaya low preference for other preference modes and viceversa. It was thus anticipated that some of the cor-

Table 1. Mean, Standard Deviation (SD), and Reliability Coeffi-cients of the TSPS Scores

Mean SD Coefficient Alpha Test-Retest Correlation

19.07 5.04 0.80 0.84(N=1339) (N=91)

relation coefficients among the four cognitive prefer-ence scores would be negative. This interdependencyshould be taken into account when interpretations ofthe correlation coefficients are made.

An examination of the intercorrelations amongscores of the four cognitive preference modes shownin Table 2 indicates that a positive but somewhat lowcorrelation exists between scores of R and P prefer-ence modes. However, strong negative correlationsare found between scores of R and P modes and thoseof Q and A preference modes. In all respects, theresults shown in Table 2, except the correlation be-tween the Q and A modes, are quite similar to thosepredicted by the cognitive preferences model (Heath,1964) and to those reported in previous studies inwhich students of the same grade level were inves-tigated (Cheng et al., 1993b; Cheng & Yang, 1995).However, no statistically significant positive correla-tion was found between scores of Q and A preferencemodes. This phenomenon was somewhat differentfrom theoretical expectations and the findings of thetwo studies mentioned above. A recent meta-analyticstudy of cognitive preferences showed that the meancorrelation between Q and A modes was -0.15 witha standard deviation of 0.29 (Tamir, 1985). This im-plies that relatively weak, moderately fluctuatingcorrelations between scores of Q and A preferencemodes might be anticipated.

The pattern of intercorrelations shown in Table2 confirmed the existence of two opposite scales, theQ-R and A-P scales, within the four biology cognitivepreference modes. They were referred to as "scientificcuriosity" and "scientific utility" scales, respectively,in previous studies (Kempa & Dube, 1973; Tamir,1975).

2. Biology Cognitive Preferences and ScienceProcess Skills

The mean and standard deviations of scoresobtained for the four preference modes are shown in

Table 2. Coefficient Alpha and Intercorrelations of Scores of theFour Cognitive Preference Modes (N=1339)

Preference Mode A

Coefficient Alpha 0.84 0.75 0.88 0.67

R (Recall) 1.00P (Principles) 0.25* 1.00Q (Questioning) -0.81* -0.52* 1.00

A (Application) -0.35* 0.58* 0.07 1.00

* Significant at the 0.01 level.

41

\

4 6

Y.J. Cheng

Table 3. The ranking order of the mean scores forthe four cognitive preference modes from the highestpreference to the lowest preference was Q>P,A>R.The ranking order of the P, Q and A preference modeswas different from what had been found in a previousstudy (Cheng & Yang, 1995) in which the rankingorder of mean scores from highest preference to lowestpreference was P>A>Q>R. Oneway ANOVA and sub-sequent multiple range tests showed that scores for theQ, P. and A modes were significantly higher than thoseof the R mode. However, no significant differenceswere found among scores of Q, P, and A modes. Thissuggested that the differences between the findings ofthe two studies were not drastic. The results indicatedthat the subjects of the present study prefer to questionbiology information presented to them more than tojust memorize it. They also indicated that the currentjunior high biology curriculum has largely attained itsintended goals.

The mean value of Q-R (Q minus R) as shownin Table 3 was 0.25, a value higher than that obtainedpreviously (Cheng & Yang, 1995). The high andpositive Q-R value (representing the scientific curi-osity scale) may be, as stated by Kempa & Dube(1973), "an indication of students' willingness anddesire to acquire scientific knowledge beyond thatalready possessed" (p. 287). But Jungwirth (1980)gave an alternative interpretation for this bipolar Q-R phenomenon by contending that "high Q scoresindicate willingness to go beyond the familiar, low Rscores indicate a distaste for the already familiar-anintrinsic dissatisfaction with knowledge already gainedand hence memorized" (p. 92). Since biology achieve-ment was correlated positively with Q scores andnegatively with R scores, it seems that Jungwirth'sstatement is more appropriate for interpreting the dataobtained in this study. The mean value of P-A (P minusA) was around 0. This indicated that the "scientificutility scale" was not as distinct as "scientific curiosityscale". This phenomenon might also indicate that, ingeneral, the subjects as a whole did not express apreference for conceptual knowledge or application of

Table 3. Means and Standard Deviations (SD) of Scores of the FourCognitive Preference Modes (N=1339)

Preference R P Q A Q-R P-A MRT AmongMode Four Modes*

Mean 2.32 2.55 2.58 2.55 0.25 0.00 R:P, R:Q,SD 0.46 0.36 0.54 0.33 0.94 0.61 R:A

* In the Multiple Range Test, only pairs of comparisons significantlydifferent at the 0.05 level are reported.

knowledge. However, a standard deviation of 0.61(Table 3) showed that the preferences for conceptualknowledge (R) and for application of knowledge (A)as exhibited by individual students were somewhatdifferent.

Since the biology curriculum received by thesubjects was inquiry-oriented, and one of the goals ofthe curriculum is to foster inquiry skills relating to theprocesses of science, the biology cognitive prefer-ences of students taking this curriculum should reflectthis trend. Therefore, it is important for this studyto examine the relationships between students' biol-ogy cognitive preferences and achievement in scienceprocess skills. Data presented in Table 4 shows thatachievement in science process skills as measured bythe TSPS correlated positively with scores for the Qpreference mode and Q-R scale, and negatively withscores for the R preference mode (p<0.01). No sig-nificant correlations were found between achievementin science process skills and scores for the P and Apreference modes together with the P-A scale. How-ever, a low but significant correlation was found be-tween scores on the multiple-choice subtest of theTSPS and scores for the P preference mode.

Since the effects of the "new" biology curriculumwill likely be different for more successful studentsand less successful students, the sample was catego-rized by TSPS achievement. If the subjects werepurposefully categorized into high (H, above +0.5 SD),medium (M, between +0.5 SD and -0.5 SD), and low(L, below -0.5 SD) performance groups according toTSPS scores, comparisons of biology cognitive pref-erences among the three groups of subjects could thusbe made. Oneway ANOVA revealed that significantdifferences in scores for R and Q preference modesand the Q-R scale were found across the three groupsof subjects. No significant differences in scores forP and A preference modes and the P-A scale werefound across the three groups of subjects (Table 5).A subsequent multiple range test revealed that the highgroup exhibited the strongest preference for the Q

Table 4. Correlations betWeen Scores of TBCP and TSPS (N=1339)

Preference Mode R P Q A Q-A P-A

TSPS# -0.30* 0.05 0.25* -0.03 0.29* 0.05TSPSCR# -0.24* 0.01 0.21* -0.02 0.24* 0.02TSPSMC# -0.29* 0.07* 0.22* -0.04 0.26* 0.06

* Significant at the 0.01 level.# TSPS : Test of Science Process Skills

TSPSCR : TSPS Constructed Response SubtestTSPSMC : TSPS Multiple-Choice Subtest

42


mode and weakest preference for the R mode. Con-versely, the low group exhibited the strongest pref-erence for the R mode and the weakest preference forthe Q mode.

Data presented in Table 5 also showed that thebiology cognitive preference orientation of the sub-jects, which was represented by the rank order of thefour preference modes from highest to lowest, wasQ>P,A>R, P,Q,A>R, and A,P>R,Q, for the high (H),medium (M), and low (L) groups, respectively. Theseresults seemed to show that, for 7th grade students,a strongest preference for critical questioning (Q mode)and a weakest preference for recall (R mode) of bio-logical information were associated with superiorperformance in science process skills. Strong pref-erences for principles (P mode), questioning and ap-plication (A mode), and a weakest preference for recallof biological information were associated with me-dium performance in science process skills. On theother hand, strong preferences for application andprinciples, and weak preferences for recall and ques-tioning of biological information were associated withinferior performance in science process skills.

These findings were different from, if not con-tradictory to, those reported by Atwood & Stevens(1978). In a study with 9th-grade students as subjects,they reported that a strong preference for the appli-cation of information (A mode) indicated science

process skill achievement, and that a preference forrecall or questioning scientific information was nei-ther an advantage nor a disadvantage to achievementin science process skills.

Science process skills are essentially intellectualskills which students use to process scientific infor-mation. Science cognitive preferences represent par-ticular preference modes students display intellectu-ally in dealing with scientific information presentedto them. Therefore, to know, in essence, what rela-tionship exists between these two cognitive processesshould be an indication of students' willingness to usescience processes to interpret information. This shouldalso be of great interest to science educators. Sincevery little study has been done in this area so far,obviously, more studies need to be done before we candraw any conclusions concerning how science cogni-tive preference styles are related to the learning ofscience process skills, and what a science teacher cando to promote superior performance in science processskills.

IV. Conclusions

From the above findings, it was concluded that7th grade students in Taiwan exhibited a strong pref-erence for questioning (Q mode) and a weak prefer-ence for memorizing (R mode) biological information

Table 5. Differences in Biology Cognitive Preferences among Three Groups of Subjects Categorized by TSPS Scores

PreferenceMode

TSPS Scores Category#F Prob.(2,1338)

MultipleRangeTest*(N=446) (N=513) (N=380)

Mean 2.16 2.35 2.48 0.000 H : M; H : LSD 0.48 0.44 0.38 M : L

Mean 2.56 2.56 2.53 0.320SD 0.38 0.35 0.33

Mean 2.76 2.55 2.43 0.000 H:M;H:LSD 0.57 0.52 0.44 M :

A Mean 2.53 2.55 2.56 0.328 -SD 0.36 0.34 0.27

Q-R Mean 0.60 0.20 -0.05 H : M; H: LSD 1.00 0.90 0.77 M : L

P-A Mean 0.03 0.02 -0.04 0.269SD 0.67 0.61 0.53

MRT AmongFour Modes*

R : P, R : QR : A, P : QQ : A

R : P, R : QR : A

R : A, P : QQ : A

TSPS Mean 24.40 19.12 12.74SD 1.90 1.43 3.01

# H, L, and M represent above +0.5 SD, below -0.5 SD, and in between, respectively.* In the Multiple Range Test, only pairs of groups significantly different at the 0.05 level are reported.

43

4 8

Y.J. Cheng

presented to them. This provided strong support forthe claim that the biology curriculum currently usedin junior high schools has partially attained its in-tended goals.

It is evident that, for 7th grade students, achieve-ment in science process skills correlates positivelywith questioning (Q preference mode), but negativelywith recall (R preference mode) of biological infor-mation. No statistically significant correlations werefound between achievement in science process skillsand the P and A preference modes.

A strong preference for critical questioning (Qmode) and a weak preference for memorizing (R mode)biological information were associated with superiorachievement in science process skills. However, strongpreferences for application (A mode) and principles(P mode) and weak preferences for recall (R mode)and questioning (Q mode) were associated with infe-rior achievement in science process skills. It is evidentthat the biology cognitive preferences exhibited bystudents with high science process skills achievementwere significantly different from those of low achiev-ers.

Acknowledgments

This study was financially supported by the National ScienceCouncil, Taiwan, under Grant No. NSC 80-0111-S-003-34. Theauthors are indebted to Ms. Lih-Mei Chen for providing commentsand assistance during the process of the study, and to ProfessorSong-Ling Mao for permission to adapt items from his scienceprocess skills tests for this study. We are grateful to Dr. Larry D.Yore, Professor, University of Victoria, Victoria, British Columbia,Canada for invaluable suggestions and critical comments.

References

Atwood, R. K., & Stevens, J. T. (1978). Do cognitive preferencesof ninth-grade students influence science process achievement?Journal of Research in Science Teaching, 15(4), 277-180.

Barnett, H. C. (1974). An investigation of relationships amongbiology achievement, perception of teacher style, and cognitivepreferences. Journal of Research in Science Teaching, 11(2),141-147.

Cheng, Y. J., Huang, C. C., Tsai, T. S, & Liaw, B. J. (1993a).Cognitive preferences in science of the seventh grade students.Chinese Journal of Science Education, 1(1), 51-76.

Cheng, Y. J., Tsai, T. S., Huang, C. C., & Liaw, B. J. (1993b). Astudy on biology cognitive preferences of junior high students.Bulletin of National Taiwan Normal University, 38, 223-249.

Cheng, Y. J., & Yang, K. Y. (1995). The relationships betweenbiology cognitive preference and academic achievement. ChineseJournal of Science Education, 3(1), 1-21.

Heath, R. W. (1964). Curriculum, cognition, and educational mea-surement. Educational and Psychological Measurement, 24(2),239-253.

Jung wirth, E. (1980). Alternative interpretations of findings incognitive preference research in science education. ScienceEducation, 64(1), 85-94.

Kempa, R. F., & Dube, G. E. (1973). Cognitive preference ori-entations in students of chemistry. British Journal of Educa-tional Psychology, 43, 279-288.

Mao, S. L. (1988). An analysis of ability levels of grade 7 and9 students in data processing and interpreting. Proceeding ofthe Third Symposium on Science Education (pp. 241-278). Taipei,Taiwan: National Science Council, R.O.C.

Mao, S. L. (1989). A growth study on the ability of plotting andinterpretating information in 7-9 grade students. Proceedingof the Fourth Symposium on Science Education (pp. 373-406).Taipei, Taiwan: National Science Council, R.O.C.

Rost, J. (1983). Cognitive preferences as components of studentinterest. Studies in Educational Evaluation, 9, 285-302.

Tamir, P. (1975). The relationship among cognitive preference,school environment, teacher' curricular bias, curriculum, andsubject matter. American Educational Research Journal, 12(3),235-264.

Tamir, P. (1976). The relationship between achievement in biologyand cognitive preferences styles in high school students. BritishJournal Educational Psychology, 46, 57-67.

Tamir, P. (1977). A note on cognitive preferences in science. Stu-dies in Science Education, 4, 111-121.

Tamir, P. (1985). Meat-analysis cognitive preferences and learning.Journal of Research in Science Teaching, 22(1), 1-17.

Tamir, P. (1988). The relationship between cognitive preferences,student background and achievement in science. Journal ofResearch in Science Teaching, 25(3), 201-216.

Tamir, P., & Kempa, R. F. (1978). Cognitive preference stylesacross three science disciplines. Science Education, 62(2), 143-152.

Tamir, P., & Jung wirth, E. (1984). Test scores and associationsas measures of cognitive preferences. Studies in EducationalEvaluation, 10, 149-158.

Tamir, P., & Lunetta, V. N. (1977). A comparison of ipsative andnormative procedures in the study of cognitive preferences. TheJournal of Educational Research, 71, 86-92.

Tamir, P, & Lunetta, V. N. (1978). Cognitive preferences in biologyof a group of talented high school students. Journal of Researchin Science Teaching, 15(1), 59-64.

Tamir, P., & Yamamoto, K. (1977). The effects of the junior high"FAST" program on student achievement and preferences inhigh school biology. Studies in Educational Evaluation, 3(1),7-17.

Van den Berg, E. (1978). Cognitive preferences: A validationstudy. Unpublished doctoral dissertation, Iowa City, IA: TheUniversity of Iowa.

44 -

49


-trii fAfl*bJ Miffagg* Jame A. Shymansky** Mt)(.t *** jggtn *

* MOINEt4-W4-tcg**Science Education Center, The University of Iowa, Iowa City, IA, U.S.A.

***

N

21iff9--06111 rt.tMiligOR1*(TBCP) NIlLit,E, (fllt$ ) Att4t 1339 fv.lal ***-treJtttinlAff AtNil-ItiMA§TA-444-38V45MinitZraireJINHA

ii3F9MAYVrtal5f214J1tfitiMIA§IfflA Eti A 'EZJIERIt.-34 (Q)) N-310j11.1(P) Iigffl (A)>Eta (R) ' olLA, Nip **4-itte:AffAfihW 42.14'0%11g ZifftitYMMINTaT4#1-rt2,t2

62K243MittifIRREAPA r f31-*AUROLNIJO (TsPs) 1143,143-1-*1-M Q ICT0413)- zrAimENER1403)-ItZrAIDECUfigi (p<0.01) ; ffigq-P RI A ig03-1-tZral.I.RgA*4f1H14A °

agtf2M TSPS fiT143-1-131-AA (H) (m) (L) hittitERVI4 VZI-TVAZOisth (ANOVA) n MAPAW,. ErM2MttfiiMACI-TAAIMAVTI

AMIAZ (Q) MSMZ rEt2 (R)J 20-130-7-04J*t 101-*3.811RfigatitAA ; R-AgiNEAZ 'OM (A) AIEWW.141.1 (P) IMAVIEIZ 'ES (R) (Q) AO ri4J*t. 'ZU °

f=1-*311#140gAgitAreif4-1tZttfifitMMITA-anUgatctEtTIA

45

50


Investigation of the Structure and Dimensionaliiy ofILPS Tests Items

MIAO-HSIANG LIN , YEONG-JING CHENG**

, SONG-LING MAO**

, HONG-MING GUO**, TAI-SHAN FANG**,

JEN-HONG LIN**

, AND JIN-TUN LIN

*Institute of Statistical ScienceAcademia Sinica

Taipei, Taiwan, R.O.C.**College of Science

National Taiwan Normal UniversityTaipei, Taiwan, R.O.C.

(Received August 2, 1996; Accepted October 16, 1996)

ABSTRACT

Employing statistical methods such as the Quasi-Simplex model, principal component analysis,cluster analysis, and Stout' s nonparametric unidimensionality test, we investigated the complexitystructure and dimensionality of ILPS test items seeking to measure student performance in six sciencesubject areas.

It was found that the four biology subtests for Grades 6 and 9 (all of which contained items designedto measure the cognitive processes involved in knowledge, understanding, application, and integration)fit the Quasi-Simplex model; however, the observed hierarchical order of complexity for the four Grade6 subtests differe'd from the original hypothesis. The results imply: (1) both biology tests are capableof measuring more than simple factual knowledge, and (2) it is more difficult for test writers to createitems which accurately measure application and integration abilities than items which measure simpleknowledge and Understanding.

Use of the single linkage method failed to find consistent patterns of distinct clusters in dendrogramsassociated with the ten ILPS tests, meaning that items within each ILPS test set maintained a homogenousstructure. Results from Stout's procedure revealed that the Grade 9 biology and physics tests failedto support the null hypothesis of essential unidimensionality. The deletion of all items with a first factorloading of less than 0;40 from these two tests resulted in the 10 ILPS tests becoming essentiallyunidimensional.

Key Words:. cluster analysis, dendrogram, indicators of science education, item response theory, non-parametric, Unidimensional test, principal component analysis, quasi-simplex model

I. Introduction

Under the auspices of the ROC's National Sci-ence Council, Shin & Cheng (1994) designed aproject to develop pre-college indicators for monitor-ing and upgrading Taiwan' s secondary science edu-cation. As suggested by Murnane & Raizen (1988),Smith (1988), Shavelson, McDonnel, Oakes, Carey,& Picus (1987), Shavelson, Carey, & Webb (1990) andOdden (1990), this monitoring system included input,process, and outcome indicators. However, during theproject' s first phase, Shin and Cheng focused on de-

veloping only a single type of outcome indicator:Indicators of Learning Progress in Science Education(ILPS). The completion of five sub-projects resultedin the creation of ILPS test sets for biology, physics,chemistry, earth science, and science process skills(Cheng & Lin, 1994; Fang, 1994; Guo, 1994; Lin,1995). The stated goals for these ILPS tests are to:(1) evaluate individual science education achievement;(2) compare science education achievement acrossschool districts or geographical regions; and (3) tracklongitudinal trends locally and nationally over time.

The ILPS test sets were constructed according to

46

51

Structure & Dimensionality of ILPS Test Item

two criteria. First, test items must correspond to majorcurricular objectives that is, they must allow for theassessment of conceptual knowledge, processing, andhigher-order thinking skills in addition to factualknowledge. Second, in order to allow for fair com-parisons based on test scores, test item parameters(e.g., difficulty and discrimination) must be invariantto compensate for the effects of those parameters 'interms of individual test items. To meet the firstrequirement, subject-matter experts were asked toconstruct test items to measure the four cognitiveprocesses of knowledge, understanding, application,and integration for each of the five subprojects. Tomeet the second requirement, panels of experts in-tended to use Item Response Theory (IRT) models(Lord, 1980) to calibrate item statistics and assessstudent performance; the IRT model was chosenbecause it provides a means of securing item param-eters with desired invariance properties.

Assessing items designed to measure cognitiveprocesses involves subjective judgments, and thusrequires the use of statistical tools to investigate hi-erarchical order of learning complexity among itemsassociated with the four cognitive categories. Inaddition, the legitimacy of using preferred item re-sponse models is predicated on an assumption ofunidimensionality. This in turn calls for statisticaltesting for unidimensionality before applying IRTmodels to calibrate ILPS tests items.

The goal of the present study was to investigatethe complexity structure and dimensionality of ILPStest items. It was hoped that the results of such aninvestigation would: (1) show how quantitative meth-ods can be incorporated to verify the objectivity ofa panel study; and (2) provide guidelines for the ap-propriate use of the IRT model for the calibration ofILPS tests items.

II. Methodology

1. Complexity Structure

Guttman's Simplex Theory (Guttman, 1969) ad-dresses the question of whether or not a set of similartests form a simple hierarchical order in terms ofcomplexity. This order-factor theory includes twomodel types: Perfect Simplex and Quasi-Simplex.

Given a set of n tests t1, tn possesses a PerfectSimplex model that is, the n tests can be arrangedin order from least to most complex and correlationsbetween any two tests can be expressed as : rik=riglrkg where j<k in terms of. degree of complexity andg is a hypothetical total complexity factor. In addition,

the matrix of inter-correlations among the n testscontains the following features: (1) all of the highestcorrelations are found next to the main diagonal; (2)as correlations move further away from the maindiagonal, they decrease to either the left or right; and(3) the highest column or row of sums associated withtests of intermediate complexity are found at or nearthe middle of the matrix, while the lowest sums (rep-resenting the least and most complex tests) are foundat the extremes.

Guttman (1969) also showed that any such setof n tests will possess the following regression prop-erties: (1) when the complexity order is given as ti<tk<tk,the correlation between ti and tk when th is partialledout is zero. In other words, the partial correlationr3k.n=0, since the partialling effect results in a completelack of common qualities remaining between tj and tic;(2) when regressing ti in the remaining tests, forexample, f =a+biti+...+bj_iti_i+bj+iti.,i+...+bt, the setof multiple regression coefficients (b's) retains a patternby which those b's associated with tests closest to ti(e.g., 121_1 and b11) are larger than those which arefurthest from tj; and (3) the multiple correlations ofR1 J... o_0(i+1)... which are associated with tests of in-termediate complexity are higher than those associatedwith the most and least complex tests.

Under the Quasi-Simplex model, any test score(Ti) is assumed as being composed of an underlying.Perfect-Simplex score lb and an error score (ei) whoserelationship can be expressed as:

Ti=111+ep j=1, 2, n, (1)

where the variables rti form a Perfect Simplex structurewhen (1) obeys three laws of deviation of the ei (e.g.,Pe7k=0, jk; Pk=0, j, k=l, 2, n; Pejek=0,In terms of the orthogonal common-factors, rpj is thelinear combination of the first j common-factors (C),for example, ii1=C1+C2+...+C1 (j=1, 2, ..., n). Thus,(1) could be expanded as:

T1=C1

T2=C1+C2

T3=C1+C2+C3

Tn-1=C14-C2-1--1-Cn-1

+el

+e2

+e3

Tn=C1+C2+ +C, +en. (2)

Moreover, (2) implies rivenh=0, a partial correlation

47

u 5 2

M.H. Lin et al.

which occurs because the effects of the same common-factors are removed from both rij and rik. In addition,the set of variables rt, also produces the same regres-sion properties and matrix pattern as those yielded bya set of Perfect-Simplex model tests.

As previously noted, items for each ILPS test setwere divided into four subtests, each measuring aseparate cognitive process associated with subjectknowledge, understanding, application or integration;it was also assumed that the four subtests possesseda simple order of learning complexity from least tomost complex. Since the four subtests within eachsubject area are of the same kind, the principles of theSimplex Theory may be used to determine if the foursubtests fit a Perfect or Quasi-Perfect Simplex model.Such an examination is also needed to assess the extentto which consistency can be established between theresults of a panel study and statistical methodology.

2. Analyses for Homogeneity and Unidimen-sionality

As stated earlier, test item sets must satisfythe criteria of unidimensionality prior to calibrationvia an IRT model. Statistically speaking, test itemhomogeneity is a required characteristic of any uni-dimensional test, and therefore ILPS tests must bescreened in order to eliminate heterogeneous items;all remaining items must still be subjected tofurther analyses for homogeneity and unidimension-ality.

A. Principal Component Analysis (Harman, 1976)

The primary concept of this analysis type is that,given a set of n items, the performance of individualitems can be expressed as a linear composite of nuncorrelated common-factors. This type of analysiscan be expressed as: Zi=Ei= latiF; =ai 1F1+a,2F2

where au is the factor loading of item i onfactor j. Principal Component Analysis has as itsunderlying technique the computation of the eigen-value (A) of the matrix (R) of item intercorrelationsvia the characteristic equation IR-2d1=0. In addition,the computation process must meet the criterion thatthe eigenvalue associated with the first factor(A1=Ei. ) should be larger than those associatedwith the remaining n-1 factors (A1=E; .14 , j=2, 3, ...,n). In other words, this criterion allows the first factorF1 to extract the highest percentage variance, whichis defined as Aq/number of items. Since the first factorholds this property of maximum variance (includinglarge loadings) from most, if not all, of the n itemsbeing analyzed, a practical approach is to measure item

homogeneity in terms of the magnitude of the firstfactor loading. For example, items whose an<0.30 areconsidered heterogeneous relative to those with anai10.30. For the present study, we adopted ail?_0.30as the critical factor loading when selecting homoge-neous items for cluster analysis.

B. Cluster Analysis

Cluster analysis evolved from methodology usedto explore the classification of fauna as part of theLinnean Taxonomic System. Essentially, this analysisis a scheme for grouping n unclassified objects intoa small number of distinct clusters. This classificationprocess uses predetermined criteria such that objectswithin any cluster hold greater similarity.

Despite the great proliferation of clustering tech-niques (see Everitt, 1980, for a full review), for ourpurposes we will limit our summary of the underlyingprinciples of hierarchical clustering to the following:given a set of n objects, each measured by p variables(which constitute a frame of reference for establishingclusters), one must choose between forming a nxnmatrix of either similarity or dissimilarity; in otherwords, the matrix elements may consist of associationmeasures or distance measures of the p variables.Regardless of the similarity/dissimilarity choice, thenxn matrix serves as input data for the creation of adendrogram a two-dimensional diagram which dis-plays the groups that are formed, as well as the associ-ated distance coefficients at each fused stage. Forexample, if the underlying structure of n objects con-sists of two distinct clusters, such a structure can bevisualized as a dendrogram similar to that shown inPanel A of Fig. 1. The dendrogram in Panel B, however,indicates that the data lacks cluster structure.

To investigate whether or not an ILPS test itemset is homogeneous, for this study we chose a single-linkage method subsumed under hierarchicalagglomerative techniques to analyze the input data ofthe inter-item correlation matrix. This decision wasmade in accordance with the argument presented byJardin & Sibson (1971) that this method is the onlyone capable of satisfying certain mathematical con-ditions which they believe are required when trans-forming a measure of (dis)similarity into a hierarchicaldendrogram. In addition, in accordance withCormack' s argument (Cormack, 1971), we chose touse cluster analysis for dissecting groups ofhomogeneous objects in other words, to test thehypothesis that a dendrogram associated with any ILPStest item set will fail to display a group structure,which of itself is considered an indication of homo-geneity.

48

53

distancecoefficient

225.36

201.77

178.16

154.55

130.94

107.32

83.71

60.10

36.49

12.68

-10.73


Panel ATwo-Group Structure

distancecoefficient

2.28

2.05

1.81

1.57

1.33

1.09

0.85

0.61

0.37

0.13

-0.11

Panel BNo-Group Structure

Hr

Fig. 1. Dendrograms representing a two-group structure and a no-group structure. (From Everitt, 1980: 87 and 90.)

C. Unidimensionality Analysis

According to item response theory (Lord &Novick, 1968; Lord, 1980), models are established viathe use of an item response function to describe to whatdegree an examinee's response is dependent on his orher ability. The ogive model in (3) represents one formof item response function, one which states that theprobability of a correct response to a dichotomouslyscored item g, P8(0) or P g(U 8=110) increases as abilityo increases, where 0 is a one-dimensional latent vari-able

Pg(0)= Prob(Ug= 119)

fag(e- bg)

J_c 2dt. [ag(O bg)] , (3)

where

bs -- P (bs) 0 50 .-.. item difficulty, ,s

as NI-rP (bs)=- item discrimination..s

Formula (3) also serves as an item-ability, regressionfunction (the regression of Ug on 0) since Pg(9)=E(U 51 0). A regression function including its inflec-tion point as well as the slope at that point remainsunchanged regardless of the frequency distribution of

nn

ability in the group being tested. Since the itemparameters bg and ag are respectively related to theinflection point and the slope of the ogive, they serveas invariant item parameters regardless of which groupis under analysis. It appears that, relative to its clas-sical counterpart, IRT serves as a better response tothe problem of securing item parameters with invari-ance properties. However, its legitimacy is dependenton the model successfully meeting an assumption ofunidimensionality.

The meaning of an unidimensionality test maybe clarified in terms of two mathematically equivalentstatements shown as (4) and (5). Let u=(u1, u2,un) represent an examinee' s binary responses to a setof n test items; (4) which represents Lazarsfeld' sassumption of local independence (Lazarsfeld & Henry,1968) states that for a fixed value of 0, an examinee'sprobability of success on all n items equals the productof the separate probabilities of success on individualitems.

g(u10)= g(u1, u2 , . . . , un10)

where

49

5 4

fg(ug10)= gliliProb(Ug= ug10)

.:1' [Ilowg[1_11(0)]1-.8, (4)

fg(ttgle) =I g

if ug = 1 ,

if tt = 0 .

M.H. Lin et al.

Equivalently, the viewpoint of unidimensionalitytaken. by (5) is that the marginal distribution g(u) ofU=(U1, U2, ..., Un) for an examinee chosen at randomcan be expressed as the conditional distribution of (4)and a prior distribution f(0) for the random variableof latent ability 0.

g(u) = f 1g(u10)f(0)cle for all u

= [Pg(0)].g[i _pg(ov-ug}f(049. (5)

Lord & Novick (1968) and Lord (1980) has pointedout that if U is normally distributed in the group tested

that is, f(0)= co(0) in (5) then it is possible tointerpret unidimensionality from the single common-factor model. In this case, 0 acts as the latent traitwhich underlies performance among test items, whichis, by definition, unidimensionality. However, theabsence of a statistical significance test for unidimen-sionality as defined in (4) or (5) forces a search forsubstitute procedures. Two such procedures havebeen proposed for practical purposes, both of thememploying a comparison of latent roots extracted fromdata matrices of item intercorrelations. For example,one can treat data as arising from one-dimensionallatent space if: (1) the first latent root is at least fivetimes as large as the second, and the second is notsubstantially larger than any of the others (Lord, 1980);or (2) if the percentage variance of the first factor

(defined as ) is over 20% (Reckase,number of items1979). Generally speaking, these two unrefined pro-cedures are only rules of thumb, that is, they are onlycapable of indicating the 'degree' to which a set oftest items departs from unidimensionality.

In view of this great need for a significance testfor unidimensionality, Stout (1987, 1990) derived thefollowing equation for testing the hypothesis of uni-dimensionality of the latent space.

P(u) = . . P(ule)dRe) . (6)

(6) serves as a nonparametric and multidimensionallatent trait model, since P g(0) is capable of taking allfunctional forms other than a normal ogive, and avector of latent space is assumed for abilities such as

0=( , 02, Ok). After taking into consideration thattest items are inherently multiply determined, Stoutset out to determine the potential for a set of test itemsto hold to an assumption of 'essential unidimension-ality' the existence of one (and only one) 'dominant'factor. In other words, Stout based the statistical indexused to assess the departure from unidimensionalityon a less restrictive assumption, namely 'essentiallocal independence,' defined as:

n(n1 1) I Cov(Ui , U'jet 0 , (7)

where n is test length, that is, the conditional cova-riances on the average approach zero. This principleimplies that examinees with approximately equal testscores on a reasonably long test should have equalability when that test has a single dominant dimension(d=1).

Stout' s formula tests 1/0: d=1 v.s. Hi: d>1 , fol-lowing the empirical variant of (7) as follows:

Stu = Covali , 0 , (8)

where

M : length of an assessment subtestY proportion right on a long partitioningP

subtest with (n M) items .

The procedure underlying (8) is the division of a testcomposed of n items into one assessment subtest withM highly homogeneous items, after which a secondsubtest is partitioned using the remaining (nM) items.When test dimensionality d=1, both subtests are saidto measure the same dimension or construct. On theother hand, when d>1 , the partitioning subtest is saidto contain items that measure dimensions which werenot measured by the assessment subtest.

Responses to the items in these subtests are usedto assess the test' s overall unidimensionality and todivide examinees into K groups. As shown by Stout(1987: 394), for examinees in any subgroup k, theirresponses to the M items of the assessment subtest areused to compute the usual score variance estimate aswell as the unidimensional variance estimate. Whenc1=1 the difference between the two estimates is small;when d>1 , the difference is much larger. Stout'sstatistic for unidimensionality is formed by summingthe normalized differences across all K subgroups. Inaddition, he also proposed a correcting statistic forshorter tests in order to control two sources of biasresulting from variability in (1) the ability of exam-

50

55


inees, and (2) item difficulty.Regarding the unidimensionality analysis, for the

present study we employed both the traditional rulesof thumb and Stout' s testing statistics to examine theunidimensionality of the seventeen ILPS tests exam-ined (see Table 2). In addition, we examined the extentto which similar results were attained by both proce-dures.

Both subtests so named are due to the fact thatthe responses on them are respectively used to assessthe test' s unidimensionality and to partition examineesinto K groups. As shown by Stout (1987: 394), forexaminees in any subgroup k, their responses on theM items of the assessment subtest are used to computethe usual score variance estimate and the unidimen-sional variance estimate. When d=1, the differencevalue between both estimates is small; when d>l, thevalue is large. Stout's statistic for unidimensionalityis formed by summing over the normalized differencesacross all K subgroups. In addition, Stout also pro-posed a corrected statistic to reduce two sources ofbias resulting from the variabilities in examinee' sability and in item difficulty when the length of a testis not long.

With respect to the unidimensionality analysis,this study will employ both the traditional rules ofthumb and Stout's testing statistic to examine whetherunidimensionality holds for all the 17 tests of ILPS(see Table 2). Moreover, this study also examines theextent to which similar results are attained by bothprocedures.

III. Data Resources

1. Testing Materials

The original NSC-commissioned research projectfocused on science achievement assessment for stu-dents in grades 6, 9, and 12. The five sub-projectsresulted in ILPS test sets being created for each of thefive subject areas across the three grade levels, withtwo test sets each for biology, earth science, and physicsand single test sets for chemistry and science processskills (SPS). In addition, to conform to the nationalcurricula established for grades 6 and 9, items ran-domly chosen from the item pools of the five sciencesubjects were used to create two "natural science"tests. In all, Shin and Cheng' s work resulted in thecreation of 26 tests. Table 1 lists the content areaswhich the ILPS sets measure. Form A of the chemistrytest was specifically designed to measure a singlecontent area (Acids and Bases); in all other cases, thecombined subject tests measure two or more contentareas.

2. Samples

For the present study we adopted a two-stagesampling frame similar to that described in the IAEPSampling Manual (1990). Each target population(corresponding to grades 6, 9, and 12) constituted auniversal list of schools in a district which includesTaipei city and county. First-stage sampling unitsconsisted of the school numbers 11, 9, and 12 forgrades 6, 9 and 12, respectively; second-stage unitsconsisted of individual classes within the three schoolsbeing sampled. (It should be noted that all senior highschool classes consisting of students majoring in thesocial sciences were excluded). We then randomlychose class samples by means of balancing such fac-tors as school quality and student ability. The number

Table 1. Content Areas Measured by ILPS Test Forms, Their Subject Matter and Grade Levels

Subject

Content Area

GradeForm A Form B

Biology Structure and Function of Plants Heredity 6, 9, 12

Earth Science Meteorology Astronomy 6, 9, 12

Physics Mechanics/Optics Electromagnetism 6

Mechanics/Motion Electromagnetism 9

Mechanics Electromagnetism 12

Chemistry Acids and Bases 6, 9, 12

Science Process Skills Communication/Interpretation 6, 9, 12

Natural Science Structure and Function of PlantsMeteorology/Astronomy 6, 9

Mechanics/ElectromagnetismAcids and Bases

51

50

M.H. Lin et al.

of students sampled was over 7,000 for grade 6 andover 9,000 each for grades 9 and 12.

The assignment procedure also entailed both in-dependent and dependent measures of the subject areasunder investigation. For example, identical groups ofgrade 12 students took chemistry and SPS tests, butindependent samples were assigned for assessingachievement in biology, earth science, and physics.For grades 6 and 9, independent samples were takenfor assessing biology, earth science, and physicsachievement while identical samples took the chem-istry and natural science tests. Finally, all sampledstudents for grades 9 and 12 were assigned for theassessment of science process skills.

Table 2 presents sample size lists for the sixsubject areas across grades 6, 9, and 12; the table alsopresents the test lengths associated with each of the26 test forms. As shown in columns 2, 7 and 12, samplesizes for five subject areas (minus science processskills) have very small ranges: 1787 to 1852, 2243 to2327, and 2170 to 2394 for grades 6, 9 and 12, re-spectively. Columns 6, 11 and 16 list revised testlengths following the deletion of items having anyformat other than multiple choice. The ranges of thenumbers of items across the five subject tests weremuch more broad: 26 to 80, 32 to 76, and 29 to 56for grades 6, 9 and 12, respectively. The data matrixfor each subject in each grade is composed of binaryresponses, and as described earlier, each matrix (17in all) was subject to cluster and unidimensionalityanalyses. In contrast, since the requirement for theminimum number of cognitive levels needed for usingthe simplex model was met only by the biology testitems, our complexity structure analysis was limitedto the use of biology test data only.

IV. Data Analysis

1. Complexity Structure

As judged by a panel of experts, the biology itemson the "A" forms for grades 6 and 9 were broken downinto four subtests according to the distribution of thecognitive processes being tested. Table 3 shows thedistributions for both forms, with T1, T2, T3 and T4respectively representing the subtests correspondingto knowledge, understanding, application and integra-tion.

Since the degree of learning complexity of thefour subtests followed the hypothesized order ofT1<T2<T3<T4, according to (2) the subtest scores canbe described as:

Tr= CI

T2= C1 + C2

T3=C1+C2+C3

+e2

+e3

T4= C1+ C2+ C3 C4 +6'4.

In addition, the Simplex part variances (m) of the foursubtest scores can be described as:

2 a2a =Th

a2 a2 2112 C +

aC2

a2 a2 a2 a2q3 C2 C3

2 a2 2 a2 a2a 114 el

ae2 e3 + e4

Table 2. Sample Sizes and Test Lengths Across Six Subject Areas and Three Grade Levels

Grade 12 ' Grade 9 Grade 62 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Test Form Original Revised Test Form Original Revised Test Form Original RevisedSample Length Test Test Sample Length Test Test Sample

Subject Size A B Length Length Size A B Length Length SizeLength

ATest Test

Length LengthBiology 1826 40 40 80 80 2286 38 38 76 76 2213 28 28 56 56-Earth Science 1796 30 30 60 60 2327 30 30 60 60 2394 25 24 49 49Physics 1787 18 18 36 26 2302 38 29 67 36 2170 35 21 56 41Chemistry 18721 35 35 35 22711 43 43 43 2391' 33 33 33SPS2 18521 35 35 31 9171 37 37 32 9132 32 32 29Natural Science 2243' 36 36 36 23901 32 32 32

' indicates identical groups of students2indicates Science Process Skill

52


Table 3. Distribution of Cognitive Processes Measured by Grades 6 and 9 Biology Test Items

Cognitive Process

Grade 9 (Form A) Grade 6 (Form A)

ni Item Number ni Item Number

T1: Knowledge 2 1 25 4 1 2 6 14T2: Understanding 13 4 7 14-19 22 30-32 35 14 4 5 7-10 12 19-21 24-27

T3: Application 12 2 5 6 8 9 12 13 20 23 24 36 37 3 5 13 28T4: Integration 11 3 10 11 21 26-29 33 34 38 7 11 15-18 22 23

40 28

Clearly, m variances increase with subtest rankorder, that is, ot < 0742 < 0.43 < 0-44, and the correlationbetween m and ilk is the ratio of an, to ark, .j1(.Guttman (1969) used this ratio to define the PerfectSimplex matrix of intercorrelation according to theQuasi-Simplex model (Joreskog, 1970). LISREL-8software (1993) has made it possible to estimate nintercorrelations by imposing certain identificationconditions on the model' s parameters. For the presentstudy we used the LISREL-8 program to measure thegoodness of fit of the Quasi-Simplex model to the lendata that is, the correlation matrix of the subtestscores on 7'1 to T4. It follows that if len fits the model,the n correlation matrix (Rm) could be used to bothcompute the complexity coefficients and investigatethe following features all of which are characteristicof a Perfect Simplex structure:

(1) As mentioned in Section II of this paper, R(n)maintains the general northeast-southwest char-acteristic;

(2) The partial correlations r14.3, r13.2, r24.3 ShOUldapproach zero;

(3) When predicting a single subtest in terms of itsthree associated subtests, the four multiplecorrelations should display a border effectand the regression weights should show aneighboring effect. The border effect refers tothe idea that the values of R1,234 and R4,123 arelower than those of R2,134 and R3,124. Theneighboring effect refers to the following com-parisons:

(i) b2 is higher than b3 and b4 in the equa-tion /71 =b2172+b3773+b4n,4;

(ii) b3 is higher than b2 and b1 in the equa-tion T4=b1711-1-b2/72+b3713;

(iii) b4 is lower than b1 and b3 in the equationf2=b1rh+b3n3b4774; and

(iv) b 1 is lower than b2 and b4 in the equation1/3=b1711+b2712+b4714.

2. Principal Component Analysis for Homoge-neity and Unidimensionality

The input data required for the principal com-ponent analysis consists of the matrix of itemintercorrelations. The measure of correlation betweenindividual items is based on the tetrachoric correlationcoefficient (see Lord & Novick, 1968: 345; Lin, 1992)

a choice supported by the assumptions underlyingthe normal ogive of (3), which is equivalent to (9):

pg(e)E E(Ugle) Prob(Ug= 110)

= ( u° (9)g le

where Y g is a hypothetical trait determining whetheror not examinees answer item g correctly and whereyg is a constant which characterizes that item. Formula(9) also indicates that, for any given item g, theunobservable continuous quantity Y is dichotomizedat yg. When Y ?yg, its observed binary quantity U g=1;when Y g<yg, then Ug=0. This hypothetical relationshipbetween U and 0 in P g(0) is illustrated in Lord &Novick (1968: 371).

In other words, given a set of n items, the IRTassumes that the multivariate distribution of (U1, U

Un) may be the result of a multivariate normaldistribution ( Y1, Y Yn) after dichotomizing eachY-variable at some point y Subsequently, the under-lying statistical problem in terms of any pair of itemsi and j is how to infer the product moment correlationp(Y Y.) from the tetrachoric correlation (Ui, Ui), sincethe latter measure is used to correlate two artificiallydichotomized variables that have a bivariate normaldistribution. For this reason, we obtained the matrixof tetrachoric item intercorrelation (Rtet) for each ofthe seventeen data sets listed in Table 2. Each 12,was subjected to the principal component analysis,with ai1F1.0.30 used as the criterion for eliminatingheterogeneous items from each data set.

For those remaining homogeneous items, we alsocomputed the corresponding Rte, matrices, with eachnew 12, once again being subjected to a principalcomponent analysis for extracting eigenvalues by

53

58

M.H. Lin et al.

solving IR,X.II=0. The two rules of thumb proposedby Lord (1980) and Reckase (1979) were then usedto compare the magnitudes of each eigenvalue as partof a unidimensionality analysis.

3. Homogeneous Structure via Clustering

Using the single linkage method, dendrogramswere obtained for each data set following principalcomponent analysis. According to this method, inputdata consisted of the matrix of similarity measuresbetween items namely, S. Here we used R, as S(that is, S i= (ski=rtel(i, j)}, j 1=1, 2, n) for com-paring results based on the same measurement units.

The single linkage procedure consists of a seriesof (n-1) fusion stages. During the procedure's firststage, the two items (i, j) with the largest (ij) inthe S1 matrix are fused to form a group because theyare the most similar item pair. To illustrate, let items3 and 6 serve as one group. At stage two, a newsimilarity matrix S2 with the order (n-1) by (n-1) isformed by treating items (3, 6) as a group; similaritymeasures between group (3, 6) and the remaining (n-2) items are obtained under the condition of S(3, 6)k=MaX{S3k, S6k} (k3, 6). In other words, as shown inthe diagram below, S2 presents group-individual aswell as interindividual item similarities.

1 2 31

2

S1= 31J

(3 6)1

) S2 = 2

nxn rz

Next, the largest S2 entry is searched for the purposeof forming a second group, with the process describedabove being followed once again until Sn_1 is formedat the final stage. During that final stage, a conclusivefusion of two groups takes place, with the final resultbeing a complete dendrogram showing all fusions madein each successive stage. The dendrograms are thenexamined in order to check for the emergence of dis-tinct clusters. For this study we used a dendrogramsimilar to that shown in Panel B of Fig. 1 to indicatea homogeneous structure of test items.

4. Stout's Unidimensionality Test

Stout' s statistic for unidimensionality which

underlies (8) is represented by the formula T=TLTB

whose asymptotic distribution is N(0, 1). TB compen-sates for the biases of item difficulty and examineevariability which occur in TL for shorter tests. Asshown in the following diagram, computation proce-dures for TL and TB differ in how different assessmentsubtest items are used.

TL

n items in a test

assessmentsubtest

partitioningsubtest

M items (n M) items

assessmentsubtest

partitioningsubtest

M2 items (n M M2) items

Both assessment subtests have identicalnumbers of items (M=M2) and similar item difficultydistributions. However, to enhance the power of thestatistical test when d>1, M items are selected toserve a nonhomogeneous role in comparison withthe (nM) partitioning items, while M2 items serve a

(3 6) 1 2

su from S1

< so = Max ts3k ,

- 1) X (n- 1)

homogeneous role to the (nMM2) partitioningitems. With the exception of this single difference,the computation process is the same for TL, and TB. Thefollowing serves as a summary of the main steps takenfor the computation of TL for a given binary data [0,1 ]Jxn with J2000 examinees and n<80 items. (Thoseinterested in reading a detailed description of thesteps involved can find them in Stout (1987: 592-595))

(1) J examinees are randomly assigned to groupsA and B in a ratio of 1:3 for the purpose ofdividing the raw data matrix [0, 1]JX, intomatrices of [0, 1bAxn and [0, 1 ]en.

(2) The data matrix [0, 1 ]JAxn is used to obtain Mitems with the largest loadings in the bi-polarfactor; these constitute the most homogeneous

54

59


set of assessment items.(3) The data matrix [0, 11 m) is used to separateJJB>qii

J B examinees into k groups based on the (nM)subtest scores such that examinees within anygroup k will have the same test score.

(4) For a fixed group of J k examinees, associatedresponses of J k to M items (as shown below)are used to calculate the score variance estimate82 in (10), the unidimensionality variance es-timate 82u,k in (11), and the normalizing con-stant ,q in (12);

(5) Statistic (10),each k group

K

T E

(11) and (12) are computed forto obtain

2 ..2Akl.

=L K1/2 k 1

(6) T B is computedthe data matrixviously mentioned,constructedof item difficultygeneous to theitems.

(7) Calculate T=(TT> Z a, wherelevel.

For the presentjected to both traditionaland Stout's Unidimensionality(n) varied accordingproportion to n.approximately 2,000tioned 1:3 ratio wasJ A and J B groups.intercorrelations basedduring Step 2 for thewere compared withdimensionality analysis"rules of thumb" in

Sk

according to steps 2-5 based on[0, 11./Bx(n-M). However, as pre-

M2 items are specificallyto be similar to M items in terms

distribution and to be homo-(nMM2) partitioning subtest

LT B)1,11 and reject 110: d=1 ifZN(0, 1) and a is the significance

research, test data sets were sub-unidimensionality analysis

test. Since test lengthto data set, we established M in

As sample size (J) consisted ofsubjects, the previously men-

used to separate J examinees intoIn addition, tetrachoric item

on data [0, l]JAxn were usedselection of M items. Results

those yielded by traditional uni-to check the validity of the

examining test unidimensionality.

ItemExaminee 2 . Score Statistic

1 U11 U by

2 U21 LI 2M

U11 UJM

y(k)y(k)

v(k)

, k)]

Proportion p1(k)

Statistic [ab , k (84 , k M4)]

( Y(k) - (k))2

E (10)=1= 1

where

M Y(k)y(k) E J y ( k E Jm J=lJk

i-e)(1 fik)

i = 1 Al2

where

=

where

4 u.EJ=1

( 11;54 k (17/4 k at) S4 k

)1124 , k k -'- m4

4 (y.00 _ rky/14, k =

1=1

k = le)(1 13k) (1 21))2

(12)

V. Results and Discussion1. Complexity Structure

Table 4 presents the results of our complexitystructure analysis on the four Form A biology subtestsfor grades 6 and 9. A goodness of fit test performedon the grade 9 data set R(T) showed x2=0.22 (P=0.64),indicating that T1T4 complied with the Quasi-Sim-plex model. The resulting R(") pattern maintains anapproximate northeast-southwest character in tworespects: (1) correlations decrease as they move to theleft or right away from the main diagonal; and (2) thetwo lowest column sums (3.00 and 3.61) are associ-ated with the least and most complex subtests rh andrh, whereas the two highest totals (3.65 and 3.64) areassociated with the two intermediate complexitysubtests, 772 and 773. We found that the values of thethree partial correlations approached zero that is,r14.3=0.004, r13.2.0.003, and r24.3.0.003. The resultsalso revealed that all four Form A biology subtests forgrade 9 displayed the border effect, since R1,234(=0 .67)and R4,123(=0.97) were lower than the R2,134 and R3,124

55

6 0

M.H. Lin et al.

Table 4. Results of Complexity Structure Analyses on Four Form A Biology Subtests for Grades 6 and 9

CognitiveLevel

Correlations amongFour Subtests

Multiple-Regression Weights

(b's)

Multiple-Correlations

(R' s)

ComplexityLoading

T1 T2 T3 T4 T1 T2 T3 7.4

Knowledge T 1 1.00 0.47 0.44 0.42 0.24 0.20 0.16 0.51 0.42

R(')Understanding T2 0.47 1.00 0.66 0.64 0.15 0.39 0.35 0.74 0.91

Application T3 0.44 0.66 1.00 0.58 0.14 0.43 0.25 0.70 0.95Integration T4 0.42 0.64 0.58 1.00 0.11 0.41 0.26 0.68 1.00

Form A sum 2.33 2.76 2.68 2.64

(Grade 9) 7/1 712 113 7/4 711 772 713 714

Knowledge rh 1.00 0.67 0.67 0.66 0.84 -0.17 0.00 0.67 0.66

R(n)Understanding 112 0.67 1.00 1.00 0.98 -0.03 1.03 0.01 1.00 0.98

Application 113 0.67 1.00 1.00 0.97 0.01 1.49 -0.50 1.00 0.98Integration 714 0.66 0.98 0.97 1.00 0.00 -0.02 0.99 0.97 1.00

SU M 3.00 3.65 3.64 3.61

T 1 T2 7.4 T3 T1 T2 T4 T3

Knowledge T1 1.00 0.26 0.20 0.10 0.21 0.11 0.04 0.29 0.10

RmUnderstanding T2 0.26 1.00 0.38 0.22 0.19 0.32 0.15 0.45 0.40

Integration Ta 0.20 0.38 1.00 0.17 0.10 0.33 0.09 0.40 0.53Application T3 0.10 0.22 0.17 1.00 0.04 0.17 0.10 0.24 1.00

Form A sum 1.56 1.86 1.75 1.50

(Grade 6) 711 712 714 713 11 I 772 714 7/3

Knowledge ni 1.00 0.51 0.49 0.29 0.59 -0.09 0.01 0.51 0.29

R(/)Understanding 172 0.51 1.00 0.97 0.56 0.05 0.95 0.00 0.97 0.57

Integration 714 0.49 0.97 1.00 0.58 -0.01 0.94 0.05 0.97 0.59Application 113 0.29 0.56 0.58 1.00 0.01 -0.05 0.62 0.58 1.00

sum 2.29 3.04 3.04 2.43

figures associated with the two subtests of interme-diate complexity. The neighboring effect was presentas well: the regression weight values b2 and b3 werefound to be the highest and b4 and b1 the lowest relativeto their respective b' s in corresponding equations. AsTable 4 shows, the complexity loadings calculated for111-414 were 0.66, 0.98, 0.98 and 1.00, respectively,which reflects a trend towards increasing complexity.Differences in loading associated with subtests112-414 were found to be negligible; the complexityloading of in was substantially lower than those of112-114.

A goodness of fit test on the grade 6 data set withthe order of T3 and T4 reversed resulted in x2=0.21(P=0.64). In other words, in the order T1-T2-T4- 7'3the four subtests fit the Quasi-Simplex model. Resultsin support of this conclusion include: (1) the R(n)pattern maintains a northeast-southwest character; (2)the sums of the four columns are consistent with thefeatures of the Perfect Simplex models in that thelowest totals (2.29 and 2.43) are associated with niand 13, while the highest totals (3.04 and 3.04) are

associated with 712 and 774; (3) the values of the threepartial correlations approach zero (r14.2=-0.022,r13.4=0.008, and r13.2=0.006); (4) the four resultantmultiple correlation coefficients comply with ourexpected pattern that 1S, R1,234 (=0.51) and R3,124(=0.58) were found to be lower than R2,134 (.0.97)and R4,123 (.0.97); and (5) all b' s displayed aneighboring effect in corresponding regression equa-tions.

The complexity loadings for ni-712--T74-773 werecalculated as 0.29, 0.57, 0.59 and 1.00, respectively(Table 4). The substantial difference between 0.59 and1.00 indicates that the complexity order of subtests interms of application and integration as judged by thepanel of experts was incompatible with the orderestablished by means of statistical methods. On theother hand, the substantial differences between 0.29and 0.57 and between 0.57 and 1.00 suggest that thesubjective judgment of the panel was the same as theconclusion arrived at statistically with respect to thecomplexity order of the knowing, understanding andintegration subtests.

56


Table 5. First Common-Factor Loadings for Items From 10 ILPS Test Data Sets

NewLength

ScienceBiology Chemistry Process Skills

G-9 G-6(n=76) (n=56)

G-9 G-6 0-9(n=43) (n=33) (n=32)

(n'=70) (n'=42)A B CA B C

(n'.41) (n'.27) (n'.29)A B CA B CA B C

0.250.450.450.490.460.230.110.450.450.410.130.410.220.470.500.570.260.560.300.300.530.130.220.120.250.260.490.240.23

I

235678121314151617181920212224252627282930313233343536373940414243444546474849505152535455565758596061626364656667686970717273747576

Mean

Notes: A: Item numbers; B: Loadings; C: Squares of Loadings

0.48 0.23 1 0.36 0.13 1 0.48 0.23 2 0.57 0.32 1 0.500.57 0.33 3 0.30 0.09 2 0.64 0.41 3 0.56 0.32 2 0.670.45 0.20 4 0.35 0.12 3 0.71 0.50 4 0.57 0.32 3 0.670.60 0.36 5 0.30 0.09 4 0.55 0.30 5 0.55 0.30 4 0.700.44 0.19 6 0.41 0.17 5 0.59 0.35 6 0.54 0.29 5 0.680.53 0.28 7 0.38 0.14 6 0.45 0.20 7 0.65 0.42 6 0.480.52 0.27 9 0.42 0.18 7 0.44 0.19 8 0.44 0.19 7 0.330.52 0.27 12 0.30 0.09 8 0.71 0.50 9 0.68 0.46 8 0.670.39 0.15 14 0.32 0.10 10 0.52 0.27 10 0.47 0.22 9 0.670.47 0.22 16 0.55 0.30 11 0.70 0.49 11 0.61 0.37 10 0.640.38 0.14 17 0.39 0.15 12 0.62 0.38 12 0.57 0.32 12 0.370.65 0.42 18 0.30 0.09 13 0.69 0.48 14 0.33 0.11 13 0.640.61 0.37 19 0.46 0.21 14 0.77 0.59 15 0.66 0.44 14 0.470.55 0.30 20 0.41 0.17 16 0.73 0.54 16 0.69 0.47 15 0.680.56 0.32 21 0.33 0.11 17 0.70 0.50 17 0.41 0.17 16 0.710.34 0.12 23 0.41 0.17 18 0.52 0.27 18 0.43 0.18 17 0.760.37 0.14 25 0.30 0.09 19 0.61 0.37 20 0.63 0.39 18 0.510.58 0.34 26 0.41 0.17 20 0.58 0.34 21 0.62 0.38 19 0.750.41 0.17 29 0.37 0.14 21 0.65 0.42 22 0.64 0.40 21 0.550.63 0.40 31 0.39 0.15 22 0.56 0.31 23 0.70 0.49 22 0.550.56 0.32 32 0.36 0.13 23 0.60 0.36 24 0.50 0.25 23 0.730.70 0.49 33 0.42 0.18 24 0.53 0.28 26 0.61 0.38 24 0.360.62 0.38 35 0.32 0.10 25 0.57 0.32 27 0.66 0.44 26 0.470.55 0.31 36 0.46 0.21 26 0.59 0.35 28 0.70 0.48 27 0.350.57 0.33 38 0.52 0.27 27 0.58 0.33 29 0.70 0.50 28 0.500.44 0.19 39 0.50 0.25 28 0.65 0.42 30 0.71 0.51 29 0.510.41 0.17 40 0.41 0.16 29 0.62 0.38 32 0.38 0.14 30 0.700.45 0.20 41 0.41 0.17 30 0.49 0.24 31 0.490.46 0.21 43 0.50 0.25 31 0.53 0.28 32 0.470.43 0.18 44 0.31 0.09 32 0.39 0.150.43 0.18 45 0.34 0.12 33 0.37 0.130.37 0.14 46 0.71 0.50 34 0.72 0.520.39 0.15 47 0.59 0.35 35 0.72 0.520.68 0.47 48 0.58 0.34 36 0.49 0.240.51 0.26 49 0.60 0.36 37 0.77 0.590.38 0.15 50 0.35 0.12 38 0.53 0.280.49 0.24 51 0.35 0.12 39 0.63 0.400.45 0.20 52 0.57 0.33 40 0.64 0.410.63 0.40 53 0.50 0.25 41 0.46 0.210.56 0.31 54 0.65 0.42 42 0.63 0.390.39 0.15 55 0.67 0.44 43 0.61 0.370.40 0.16 56 0.57 0.320.67 0.440.34 0.120.67 0.450.52 0.270.45 0.200.53 0.280.54 0.300.63 0.400.69 0.470.62 0.380.60 0.370.67 0.450.60 0.360.62 0.390.61 0.370.65 0.430.60 0.360.53 0.280.67 0.450.32 0.100.58 0.330.58 0.340.63 0.390.69 0.480.41 0.170.70 0.490.64 0.410.57 0.32

0.53 0.43 0.59 0.58 0.57

A

1

34567910121314151617202122232425272829

Physics Natural Science Eearth Science

G-I2(n=26)

G-9(n=36) (n=36)

0-9 G-6(n=32) (n=60)

0-9

(n'=23)B CA

(n'.29)B C

(n'=33)AB C (n'=28)AB C (n'.53)A B C

0.51 0.26 1 0.61 0.37 1 0.37 0.14 3 0.30 0.09 1 0.55 0.310.49 0.24 2 0.70 0.48 2 0.46 0.22 4 0.41 0.16 2 0.69 0.470.50 0.25 3 0.46 0.21 3 0.61 0.38 5 0.32 0.10 3 0.61 0.370.44 0.20 7 0.58 0.33 4 0.53 0.29 6 0.38 0.15 4 0.60 0.360.42 0.17 9 0.80 0.63 5 0.59 0.35 7 0.57 0.32 5 0.53 0.280.57 0.33 12 0.83 0.68 6 0.44 0.19 8 0.66 0.43 6 0.38 0.140.64 0.41 19 0.33 0.11 7 0.62 0.39 9 0.40 0.16 7 0.58 0.340.50 0.25 20 0.78 0.62 9 0.41 0.17 10 0.47 0.22 8 0.60 0.360.46 0.21 21 0.72 0.53 10 0.53 0.28 13 0.58 0.33 10 0.58 0.330.72 0.52 23 0.38 0.14 11 0.57 0.33 14 0.45 0.20 11 0.61 0.370.46 0.21 26 0.46 0.21 12 0.58 0.33 15 0.38 0.14 12 0.60 0.360.36 0.13 30 0.53 0.28 14 0.42 0.18 16 0.48 0.23 13 0.63 0.400.55 0.31 31 0.61 0.37 15 0.35 0.12 17 0.73 0.53 14 0.42 0.170.57 0.33 36 0.45 0.20 16 0.61 0.38 18 0.73 0.53 15 0.50 0.250.58 0.34 39 0.61 0.37 17 0.40 0.16 19 0.74 0.55 16 0.49 0.240.45 0.20 40 0.71 0.51 18 0.41 0.17 20 0.61 0.37 17 0.70 0.490.62 0.39 41 0.57 0.32 19 0.71 0.51 21 0.40 0.16 18 0.73 0.530.55 0.30 42 0.47 0.22 20 0.70 0.49 22 0.73 0.53 19 0.59 0.340.51 0.26 43 0.31 0.09 21 0.68 0.46 23 0.69 0.48 20 0.60 0.360.51 0.26 44 0.71 0.51 22 0.60 0.36 24 0.34 0.12 21 0.67 0.440.53 0.28 45 0.48 0.23 23 0.57 0.33 25 0.40 0.16 22 0.66 0.430.38 0.14 52 0.70 0.49 24 0.61 0.37 26 0.37 0.14 23 0.73 0.540.44 0.19 53 0.64 0.41 25 0.56 0.32 27 0.51 0.26 24 0.67 0.44

54 0.64 0.40 26 0.67 0.44 28 0.49 0.24 25 0.64 0.4158 0.68 0.46 27 0.60 0.37 29 0.36 0.13 26 0.69 0.4860 0.35 0.12 28 0.69 0.47 30 0.33 0.11 27 0.68 0.4665 0.74 0.55 29 0.68 0.46 31 0.42 0.17 28 0.46 0.2166 0.67 0.45 30 0.57 0.32 32 0.50 0.25 29 0.47 0.2267 0.55 0.30 31 0.51 0.26 30 0.59 0.35

32 0.40 0.16 31 0.63 0.3933 0.62 0.38 32 0.37 0.1434 0.53 0.28 33 0.50 0.2536 0.68 0.47 34 0.43 0.19

35 0.67 0.4536 0.48 0.2338 0.64 0.4139 0.45 0.2040 0.51 0.2641 0.45 0.2042 0.33 0.1143 0.59 0.3545 0.43 0.1846 0.41 0.1747 0.30 0.0948 0.43 0.1849 0.41 0.1653 0.39 0.1554 0.53 0.2855 0.44 0.1956 0.45 0.2057 0.57 0.3358 0.36 0.1359 0.34 0.12

0.51 0.59 0.55 0.49 0.53

57

t, 6 2

M.H. Lin et al.

Table 6. Extracted Variances (Eigenvalues) of First Two Common Factors for 10 ILPS Test Data Sets

Subject Grade

Sample

Size(J)

Test

Length(n')

1

X22

RatioA.1/2.2

Eigen-Value

Percentvariance

Eigen-value

Percentvariance

Biology 9 (2286) 70 20.6 (29%) 3.34 (5%) 6.176 (2213) 42 8.36 (20%) 1.79 (4%) 4.67

Chemistry 9 (2271) 41 14.85 (36%) 1.66 (4%) 8.976 (2391) 27 9.27 (34%) 1.24 (5%) 7.51

Natural Science 9 (2243) 33 10.51 (32%) 1.7 (5%) 6.186 (2390) 28 7.3 (26%) 1.49 (5%) 4.91

Physics 12 (1787) 23 6.18 (27%) 1.27 (6%) 4.899 (2302) 29 10.61 (37%) 2.41 (8%) 4.41

Earth Science 9 (2327) 53 15.85 (30%) 3.29 (6%) 4.82Science Process Skills 9 (9171) 29 9.95 (34%) 1.68 (6%) 5.93

First Factor Extracted Variance2 Second Factor Extracted Variance

Eigartha

9.0 T

8.0

7.0

6.0

5.0

3.0

Biology

(Grade 6)

0.0 294567891110 1 2

Eigarralue

1.0

7.0

6.0

Natural Science

(Grade 6)

Eiproolue Biology

12.0 (Grade 9)

20.0

11.0

16.0

14.0

12.0

10.0

1.0

6.0

4.0

0 0 Naar1 2 1 4 5 6 7 9 1 1 1 1

0 1 7 3

eigarvalue

12.0

10 0

1.0

4.0 0 6.0

2 0

1 0

0 0 1

2

1 1

1 4 5 6

0.0

Natural Science

(Grade 9)

9.0

.0

7.0

6.0

5.0

4.0

3.0

2.0

1.0

0.0

Eiger:m.1u.

10.0 .

Science Process Skill'

(Grade 9)

ChemisUy

(Grade 6)

7

Eigarrolue

16.0

Earth Science

(Grade 9)

14.0

17.0

10.0

0.0

1 1 3 4 5 6 7 11 9 1 11 kaor

0- 1

F.igasvolue Chemistry

16.0 (Grade 9)

14.0

12.0

10.0

11

6.0

4.0

2.0

0.0

k000r

Eiger, Ow

11.0

10.0

LO

6 0

4.0

2.0

0.0

Physics

(Grade 9)

Eiseavalta

6.0

5.0

4 0

3.0

2.0

1.0

0.0 I 1 1 I 4 II

2

Physics

(Grade 12)

2 1 0 6

Fig. 2. Largest eigenvalues (<1.00) in order of size for the tetrachoric correlation matrices of 10 ILPS test data sets.

2. Principal Component Analysis for Homoge-neity and Unidimensionality

Each of the 17 ILPS tests of length n listed inColumns 6, 11 and 16 of Table 2 was subjected to aprincipal component analysis on the corresponding nby n matrix of 11.,. Seven ILPS tests showing non-negative definite lite, results were suspended for fur-ther analysis. Table 5 summarizes the first common-factor loadings for items in the 10 remaining ILPS

tests.As shown in Table 5, according to a +0.30 cri-

terion for first common-factor loading, the number ofitems deleted from the remaining 10 tests varied withsubject and grade; the number of items deleted rangedfrom 2 to 14. The lengths of these 10 ILPS tests werethus reduced to 70, 42, 41, 27, 29, 23, 29, 33, 28, and53 items, with their loading means for n' items re-corded at 0.53, 0.43, 0.59, 0.58, 0.57, 0.51, 0.59, 0.55,0.49, and 0.53, respectively. A high mean value (from

- 58

63

4

t

t

2

1 1

. i I

tttiiAnWTARA2=RSM5r,,,,72qa2R;:,Z..q.'RE...Mm^44.,-.274


Biology-Grade 6 (a' = 42) Biology-Grade 9 (n 70)

I

1

-1

1

11

ScienceProce9sSkilCh-ade9(n'=29)

-11-111

111111111

- --,---men2q..51-empm..mmrsm2roarl-xN

Chenilstry-Grade6(n'=27)

I

Physics-Grade 12 (n' = 23

mi

1

1

1 1

22.4.-.^2NAR^.MR-.r12=4.,.42

ill1-11 I

It:- t- [

t1.11 1

1111 1 III,I 1

i 1-11111,.;!:;; i;

1 I,11;it

S3g34,,..SW=S?:-.-P4G5IrrS2=nGN2M4VMP1324.1A4A2.^:,,,,,,-13402P(P=m^WP=ggas,2".2

Earth Science-Grade 9 (n' = 53);

71

I

ar,119=2aRXM'a°=ASAR7A2=2171,...,m2-m3511419$47SAMS4WM4

(a)

Chendstry-Grade 9 (n' = 41) Physics-Grade 9 (n' = 29)

IT

74-.42A,M=7244A"WZa="132,2Wq77-4a.g,a2R^=1 :2a.',-,1"242DINSMM,,,^^a51-.11.128M

Natural Science-Grade 6 (n' = 28)

.1

i

1 11 1

11 1

1 11 1

11 1

1

1

1-.-

1 1

1

is)

2..

Natural Science-Grade 9 (1,' = 33)

_

12 I 1

1 1 1

Z22P14.1ir=m14244411A224...,m,,^g FI,T,24£119418W-PM".211RMAR=.2.°A

(b)

1

Fig. 3. (a) Dendrograms associated with biology, SPS, and earth science ILPS test items. (b) Dendrograms associated with chemistry,

natural science, and physics ILPS test items.

0.43 to 0.59) indicates item homogeneity within a test Using Kaiser's criterion for the tetrachoric cor-item set. relation matrices of 10 n' -item ILPS tests, Fig. 2 is

59

O 6 4

M.H. Lin et al.

a presentation of eigenvalue scree plots against theoptimum number of extracted factors. Figure 3 showsthat the number of factors with A,1.00 falls within arange of 3-14 across the 10 data sets. It was foundthat A2 was significantly larger than 23 in both datasets for grade 9 biology and earth science.

Table 6 presents: (1) the eigenvalues of the firsttwo common factors AI and A,2, (2) the associatedpercentage variances, and (3) the ratios of Ai to 2,2.The table shows that the percentage variance of thefirst factor ranged from 20% to 37%, with a AI tOratio ranging from 4.41 to 8.97 across the 10 data sets.However, we found that 2,22.00 for the grade 9 bi-ology, earth science, and physics data sets. Together,these results indicate: (1) the 10 ILPS test item setswith ai1F1?.0.30 met the unidimensionality criterionset by Reckase (1979) since the percentage variancesof the first factor were all above 20%; (2) when analyzedin terms of Lord' s rule (Lord, 1980), the grade 9biology, physics, and earth science test item sets couldnot be considered unidimensional because the secondeigenvalue was much larger than any other.

3. Homogeneous Structure via ClusteringMethod

Figure 3(a) and (b) consist of the dendrogramsproduced following the application of the single link-age method to the 10 ILPS data sets; the horizontallines denote joined items and the vertical lines denotethe distance at which the items are joined. It shouldbe noted that the original scale values (coefficients)found in Lin (1995) have been rescaled to fall withinthe range of 1 to 25 under the SPSS Cluster procedure.

The dendrograms in Fig. 3(a) and (b) are quitesimilar in shape to that presented as Panel B of Fig.1. In addition, we failed to find any large change invalues in terms of distance at any combined item stage.The ladder-like shape and small distance values amongdendrogram stages suggest a lack of group structureto the data; this holds true for each of the 10 ILPSdata sets. These results indicate that the ILPS test itemsets which correspond to subject/grade areas arehomogeneous.

We examined the dendrograms to find the coreset of items having the greatest similarity. The 10ILPS test sets corresponding to their respective den-drograms in Fig. 3(a) and (b) are (54, 55, 46, 49), (60,64, 62, 63), (4, 5), (20, 21, 26, 27), (28, 29), (14, 17,35, 16, 34, 37), (12, 20, 9, 21), (13, 14), (17, 19, 18,22), and (20, 21, 19). When analyzing item stems,we found that the items within the parentheses mea-sured the same major and minor concepts corre-

sponding to the subject content areas previously de-scribed. On the other hand, respective sets of dissimi-lar items relative to those in the core sets were iden-tified as (5, 1, 18, 14), (37, 15, 13, 21, 68), (24, 27),(59, 47), (32, 14), (7*, 41*, 32, 33), (19, 60, 23), (21*,15), (5, 30), and (18, 15). With the exception of thethree items marked with an asterisk, we found that theai1F1 values for these items were all less than 0.40(Table 5).

4. Stout's Unidimensionality Test

Table 7(a) and (b) present a summary of theresults for the 10 test data sets, including score vari-ance estimate values (8-1), unidimensional varianceestimates (62u, k), the normalizing constant (Sk) foreach of K groups, and TL, TB and T values for eachsubject area; all values were obtained as the result ofprocedures described in the methodology section ofthis paper. The M value for item set assessment ofthe grade 9 subtests for biology and earth science was7, while for all other test sets M=5. This differencewas meant to compensate for the variance in test length(n') according to subject area and grade level. Readersinterested in examining detailed results with respectto the 84, k, 4, and it4,k values involved in obtainingSk will find them in Lin (1995).

As Table 7(a) and (b) show, the calculated valuesof T associated with the grade 9 biology and physicstests were 2.45 and 2.19 respectively both larger thanZ0.05=1.65. The calculated T values associated withthe other eight tests were all less than 1.65. Thus, twoof the 10 data sets reject 1/0: d=1 at a 0.05 level ofsignificance, implying that the two tests contain itemswhich load heavily on at least two dominant dimen-sions.

Deleting items having an a11F1<0.40 from thegrade 9 biology and physics tests resulted in therespective removal of 10 and 4 items. Results of anidentical computing procedure for analyzing the twotest data sets with n" are presented in Table 7(c); thebiology Tvalue was measured at 1.26 while the physicsT value was 0.49. Since both of the new T values wereless than 1.65, the biology (60 items) and physics (25items) test data both met the assumption of essentialunidimensionality.

VI. Conclusions and SuggestionsWe found that the four grade 9 biology subtests

matched the Quasi-Simplex model due to their havinga hierarchical order of complexity of T1 (knowledge)< T2 (understanding) < T3 (application) < T4 (integra-

60

65


Table 7(a). Stout's Unidimensionality Test for Biology, Chemistry and Earth Science ILPS Items

Biology-9(n'=70)

Biology-6(n =42)

Chemistry-9(n ' =41)

Chemistry-6(n'.27)

Earth Science-9(n'=53)

ifdt,k Sk zib,k Sk r2USk ai U.k Sk C 6-2u k Sk

k=1 0.07915 0.03374 0.01670 0.04303 0.03942 0.01356 0.05545 0.03260 0.01175 0.06250 0.04352 0.01277 0.11846 0.03511 0.014392 0.07453 0.03435 0.01425 0.04305 0.03723 0.01170 0.05597 0.03487 0.01245 0.05164 0.04077 0.00927 0.05043 0.03457 0.011753 0.09337 0.03472 0.01342 0.02182 0.04284 0.00922 0.03999 0.03360 0.01034 0.04665 0.03946 0.01000 0.08333 0.03318 0.010824 0.09601 0.03345 0.01327 0.05581 0.04155 0.01400 0.04157 0.03405 0.00728 0.04783 0.03946 0.00980 0.07307 0.02985 0.013645 0.07384 0.03269 0.01170 0.03496 0.03929 0.00735 0.05103 0.03314 0.00843 0.04408 0.03514 0.00849 0.09467 0.03096 0.013786 0.07399 0.03087 0.01480 0.04130 0.03656 0.00812 0.04928 0.03518 0.00775 0.04267 0.03294 0.00768 0.08009 0.02969 0.013717 0.07799 0.02849 0.01345 0.03862 0.03747 0.00837 0.04639 0.03113 0.00742 0.04403 0.03221 0.00735 0.05875 0.02318 0.014258 0.06316 0.02602 0.01200 0.04965 0.04229 0.00894 0.03951 0.03094 0.00693 0.03324 0.02656 0.00671 0.03880 0.02188 0.008199 0.05026 0.02729 0.01162 0.03711 0.04023 0.00678 0.04168 0.02927 0.00755 0.03833 0.02731 0.00762 0.03727 0.01693 0.01418

10 0.04973 0.02723 0.01311 0.04360 0.04224 0.00755 0.03986 0.02745 0.00883 0.03207 0.02344 0.00648 0.01757 0.01286 0.0052911 0.03676 0.02114 0.01253 0.03714 0.04210 0.00616 0.04222 0.02827 0.00837 0.02434 0.0,1969 0.00592 0.02857 0.01599 0.0087212 0.04954 0.02018 0.01905 0.04223 0.04227 0.00678 0.02466 0.02175 0.00548 0.02671 0.02036 0.00616 0.01206 0.00986 0.0052013 0.05192 0.02091 0.01738 0.04968 0.04105 0.00686 0.02428 0.02328 0.00714 0.01796 0.01636 0.00346 0.01322 0.00937 0.0052014 0.03782 0.01848 0.00894 0.04213 0.03841 0.00721 0.02063 0.02045 0.00490 0.01473 0.01352 0.00265 0.01239 0.00979 0.0053915 0.03188 0.01739 0.01208 0.04366 0.04054 0.00678 0.02318 0.01872 0.00883 0.01147 0.01155 0.00141 0.00830 0.00649 0.0034616 0.02656 0.01373 0.01015 0.05053 0.03886 0.00800 0.02028 0.01578 0.00837 0.01429 0.01374 0.00265 0.00470 0.00435 0.0024517 0.02031 0.01359 0.00837 0.04294 0.03696 0.00877 0.01565 0.01724 0.00480 0.00921 0.00701 0.0041218 0.02395 0.01455 0.00906 0.04257 0.04235 0.00781 0.01369 0.01124 0.00632 0.00307 0.00344 0.0014119 0.03456 0.01503 0.01612 0.03917 0.03597 0.00849 0.01612 0.01277 0.00700 0.00517 0.00397 0.0030020 0.01985 0.00829 0.01378 0.04479 0.03302 0.01020 0.01045 0.00852 0.00592 0.00621 0.00637 0.0026521 0.01779 0.00832 0.01140 0.00864 0.00990 0.00332 0.00361 0.00423 0.0017322 0.01710 0.01179 0.00656 0.00725 0.00690 0.00447 0.00315 0.00352 0.0020023 0.00633 0.00763 0.00316 0.00674 0.00969 0.00245 0.00241 0.00260 0.0017324 0.03619 0.01306 0.01386 0.00885 0.00955 0.00374 0.00497 0.00472 0.0028325 0.00387 0.00469 0.00224 0.00622 0.00635 0.00332 0.00237 0.00253 0.0017326 0.00794 0.00828 0.00374 0.00563 0.00484 0.00374 0.00491 0.00345 0.0031627 0.00502 0.00452 0.00332 0.00649 0.00671 0.00346 0.00273 0.00301 0.0020028 0.00378 0.00442 0.00224 0.00275 0.00269 0.00200 0.00595 0.00298 0.0052929 0.01106 0.00830 0.00480 0.00356 0.00393 0.00265 0.00165 0.00170 0.0017330 0.00434 0.00324 0.00361 0.00278 0.00285 0.0020031 0.00413 0.00508 0.00245 0.00070 0.00068 0.0014132 0.00286 0.00307 0.00245 0.00316 0.00212 0.0030033 0.00453 0.00554 0.00200 0.00097 0.00093 0.0017334 0.00435 0.00370 0.0031635 0.00518 0.00451 0.0037436 0.00123 0.00124 0.0017337 0.00456 0.00314 0.0042438 0.00073 0.00070 0.0014139 0.00073 0.00070 0.0014140 0.00120 0.00116 0.0014141 0.00093 0.00089 0.00173

71 7.44 1.19 3.37 3.71 6.10

k=1 0.04925 0.03163 0.01005 0.04397 0.03510 0.01010 0.03489 0.03099 0.00883 0.07957 0.04408 0.01526 0.05220 0.03364 0.009002 0.04345 0.03047 0.00954 0.03830 0.03738 0.00877 0.03358 0.03492 0.01015 0.04393 0.04302 0.00755 0.03007 0.03204 0.005293 0.04356 0.03376 0.00787 0.03016 0.03885 0.00927 0.04165 0.03564 0.00640 0.05361 0.04270 0.00964 0.04918 0.03345 0.007214 0.04945 0.03456 0.00894 0.03902 0.03484 0.00671 0.04548 0.03994 0.00735 0.04830 0.03885 0.00949 0.04762 0.03301 0.006485 0.04267 0.03381 0.00781 0.03512 0.03571 0.00707 0.04076 0.03858 0.00583 0.04506 0.03448 0.00812 0.05215 0.03088 0.007686 0.05680 0.03407 0.00686 0.04065 0.03673 0.00671 0.05395 0.04110 0.00632 0.04450 0.03150 0.00849 0.04043 0.03076 0.005837 0.04225 0.03536 0.00671 0.04514 0.03792 0.00806 0.05090 0.04001 0.00592 0.03714 0.02772 0.00624 0.05169 0.03021 0.007758 0.05365 0.03560 0.00949 0.03895 0.03962 0.00592 0.04984 0.03986 0.00700 0.02 73 0.02147 0.00480 0.05551 0.02954 0.009069 0.05022 0.03492 0.00794 0.04448 0.03861 0.00640 0.06818 0.03962 0.00964 0.01428 0.01432 0.00387 0.02805 0.02375 0.00548

10 0.04801 0.03348 0.00872 0.05172 0.03989 0.00686 0.05192 0.03801 0.00831 0.02218 0.01738 0.00510 0.03121 0.02366 0.0054811 0.04368 0.03480 0.00616 0.04461 0.03921 0.00671 0.05517 0.03832 0.01005 0.01699 0.01244 0.00469 0.02546 0.02466 0.0067112 0.03763 0.03135 0.00831 0.04463 0.03831 0.00693 0.04185 0.03487 0.00889 0.00943 0.00954 0.00245 0.02338 0.02115 0.0046913 0.02120 0.03017 0.00400 0.04363 0.04058 0.00735 0.04393 0.03202 0.00854 0.00445 0.00488 0.00173 0.01460 0.01465 0.0031614 0.02993 0.03104 0.00707 0.05064 0.03722 0.00906 0.04949 0.03346 0.01000 0.00554 0.00441 0.00300 0.01514 0.01708 0.0036115 0.02600 0.02660 0.00632 0.03116 0.03680 0.00539 0.03336 0.02446 0.00742 0.01660 0.01356 0.0058316 0.04704 0.02659 0.01625 0.03673 0.03434 0.00800 0.02519 0.02078 0.00557 0.01139 0.01168 0.0034617 0.03272 0.02479 0.00995 0.03636 0.03653 0.01153 0.02221 0.02096 0.00624 0.01443 0.01290 0.0077518 0.03111 0.02479 0.00800 0.01445 0.01971 0.00469 0.01133 0.01110 0.0041219 0.02653 0.02278 0.00671 0.01688 0.01410 0.00539 0.01142 0.00911 0.0046920 0.01537 0.02114 0.00458 0.01223 0.01107 0.00548 0.00883 0.00686 0.0034621 0.03135 0.02048 0.00686 0.00581 0.00735 0.00245 0.00836 0.00532 0.0053922 0.01859 0.01641 0.00583 0.00861 0.00891 0.00346 0.00669 0.00665 0.0031623 0.01810 0.01479 0.00583 0.00470 0.00398 0.00316 0.00603 0.00485 0.0033224 0.01455 0.01472 0.00480 0.00281 0.00285 0.00224 0.00602 0.00477 0.0034625 0.01239 0.01132 0.00424 0.00328 0.00346 0.00224 0.00116 0.00113 0.0014126 0.01587 0.01272 0.00500 0.00282 0.00136 0.0034627 0.01122 0.00700 0.00721 0.00064 0.00062 0.0010028 0.01242 0.01058 0.00447 0.00366 0.00426 0.0024529 0.01017 0.00845 0.0048030 0.01380 0.01004 0.0080631 0.00752 0.00835 0.0034632 0.00680 0.00745 0.0038733 0.00777 0.00438 0.0064834 0.00572 0.00570 0.0033235 0.00694 0.00513 0.0043636 0.00237 0.00242 0.0020037 0.00376 0.00426 0.0026538 0.00680 0.00629 0.00458

TB 3.98 1.81 3.79 2.93 4.10

T 2.45* -0.44 -0.29 0.55 1.41

Notes: * Reject; 110: d=1 if T>Z0.05(=1.65), T=(TL-TB)lif

61 -

66

M.H. Lin et al.

Table 7(b). Stout's Unidimensionality Test for Natural Science, SPS, and Physics ILPS Items

Natural Science-9(n'=33)

Natural Science-6(n'=28)

Science Process Skill-9(n =29)

Physics-12(n'=23)

Physics-9(n'=29)

s, Ø Ekli,k s, ö 6it.k Sk 1311 k s. 6i 6-6 s,k=1 0.04113 0.03930 0.00985 0.03502 0.03604 0.01086 0.09714 0.04696 0.01597 0.06277 0.04223 0.01631 0.06731 0.03769 0.01836

2 0.04444 0.03864 0.01109 0.02683 0.03515 0.00906 0.08567 0.04733 0.01118 0.04493 0.03477 0.01208 0.08586 0.04326 0.013493 0.04106 0.04103 0.00700 0.06635 0.04212 0.01746 0.06953 0.04425 0.00954 0.05129 0.03489 0.01281 0.08406 0.04650 0.010684 0.05102 0.03709 0.00964 0.04710 0.04327 0.00794 0.07779 0.04460 0.00748 0.04306 0.03944 0.00883 0.07223 0.04640 0.008605 0.04391 0.03972 0.00728 0.05240 0.04166 0.00877 0.08325 0.04211 0.00755 0.04953 0.04209 0.00762 0.07306 0.04157 0.010006 0.03783 0.04043 0.00608 0.04448 0.04390 0.00748 0.05624 0.03810 0.00548 0.06338 0.04250 0.00806 0.04227 0.03612 0.006867 0.03153 0.03811 0.00529 0.05804 0.04676 0.00819 0.04818 0.03184 0.00469 0.07016 0.04498 0.00831 0.04764 0.03515 0.008498 0.03129 0.03281 0.00671 0.04946 0.04573 0.00812 0 .04048 0.02618 0.00539 0.05338 0.04341 0.00624 0.03889 0.03678 0.005839 0.04731 0.03730 0.00866 0.05960 0.04672 0.00721 0.02851 0.02203 0.00400 0.06015 0.04405 0.00735 0.04287 0.03496 0.00794

10 0.04058 0.03597 0.00806 0.05469 0.04733 0.00583 0.02368 0.01689 0.00436 0.06250 0.04461 0.00728 0.04524 0.03720 0.0087211 0.04652 0.04405 0.00894 0.06603 0.04799 0.00648 0.01404 0.01272 0.00245 0.06427 0.04062 0.00748 0.03935 0.03866 0.0062412 0.05485 0.03992 0.00985 0.06467 0.04868 0.00583 0.01278 0.01090 0.00300 0.04278 0.03617 0.00686 0.04076 0.03877 0.0076213 0.03568 0.03609 0.00787 0.06735 0.04705 0.00608 0.00806 0.00719 0.00200 0.04519 0.03484 0.00927 0.05841 0.04249 0.0112214 0.05404 0.04181 0.00949 0.07138 0.04640 0.00735 0.00712 0.00696 0.00173 0.04750 0.02986 0.01086 0.06203 0.04161 0.0114915 0.04673 0.03901 0.01100 0.05595 0.04030 0.00755 0.00750 0.00640 0.00200 0.04302 0.02899 0.01127 0.05277 0.03991 0.0102516 0.05526 0.04384 0.01044 0.05213 0.04017 0.00831 0.00664 0.00600 0.00141 0.03429 0.01887 0.01510 0.05583 0.04151 0.0087717 0.06141 0.04353 0.01054 0.05020 0.04002 0.01030 0.00434 0.00396 0.00141 0.05753 0.04052 0.0110918 0.05089 0.04109 0.00849 0.00324 0.00346 0.00100 0.06232 0.03771 0.0105819 0.08015 0.04533 0.01281 0.00254 0.00267 0.00100 0.05127 0.03170 0.0123720 0.06812 0.03687 0.01375 0.00278 0.00266 0.00100 0.04495 0.02748 0.0119221 0.03570 0.02498 0.01136 0.00283 0.00257 0.00141 0.04114 0.02283 0.0109122 0.04854 0.02647 0.01539 0.00383 0.00322 0.00265 0.02445 0.02051 0.0092723 0.04182 0.02318 0.0176124 0.01328 0.01597 0.00648

71 3.73 5.78 7.43 6.44 7.37

k=1 0.04036 0.03430 0.01330 0.02507 0.03597 0.00900 0.05400 0.04705 0.00825 0.05460 0.03990 0.01970 0.05222 0.02250 0.028902 0.03621 0.03801 0.00728 0.05676 0.04250 0.01612 0.07386 0.04790 0.00748 0.04790 0.04029 0.00964 0.03120 0.02464 0.007873 0.04122 0.03759 0.00794 0.04833 0.04136 0.01127 0.06712 0.04851 0.00510 0.05471 0.04593 0.00854 0.03942 0.02610 0.008254 0.03813 0.04009 0.00648 0.05088 0.04338 0.00980 0.06711 0.04883 0.00436 0.05609 0.04661 0.00728 0.03944 0.03469 0.008435 0.03371 0.03988 0.00510 0.03902 0.04615 0.00768 0.07542 0.04805 0.00424 0.04911 0.04497 0.00632 0.05041 0.03641 0.008006 0.03984 0.03709 0.00700 0.05268 0.04015 0.00943 0.06090 0.04417 0.00361 0.05472 0.04525 0.00600 0.04969 0.04269 0.006327 0.03822 0.04051 0.00640 0.04814 0.04660 0.00640 0.05793 0.03762 0.00412 0.06163 0.04462 0.00678 0.05847 0.04605 0.007008 0.04212 0.04343 0.00663 0.06089 0.04897 0.00781 0.04721 0.03325 0.00424 0.05205 0.04278 0.00574 0.06338 0.04850 0.007219 0.04441 0.04233 0.00748 0.06104 0.04835 0.00671 0.03960 0.02916 0.00361 0.04633 0.03834 0.00539 0.05880 0.04851 0.00755

10 0.04290 0.04460 0.00693 0.06326 0.04938 0.00500 0.03322 0.02586 0.00332 0.04808 0.03914 0.00648 0.06381 0.04733 0.0085411 0.05925 0.04527 0.00812 0.05958 0.04814 0.00510 0.01902 0.01514 0.00283 0.03757 0.02854 0.00843 0.05750 0.04295 0.0083712 0.04554 0.04694 0.00693 0.05454 0.04541 0.00480 0.01474 0.01236 0.00245 0.02888 0.02237 0.01020 0.05695 0.04005 0.0086013 0.05421 0.04525 0.00812 0.05406 0.04347 0.00529 0.01096 0.01002 0.00200 0.04012 0.03460 0.0084314 0.05002 0.04033 0.00748 0.03658 0.03822 0.00548 0.00933 0.00848 0.00173 0.02519 0.02476 0.0050015 0.05661 0.04435 0.00883 0.04015 0.03378 0.00911 0.00693 0.00671 0.00141 0.02321 0.02010 0.0058316 0.04152 0.04256 0.00608 0.00486 0.00471 0.00141 0.01564 0.01982 0.0051017 0.04243 0.03828 0.00787 0.00346 0.00364 0.00141 0.01201 0.01073 0.0063218 0.03912 0.03190 0.00831 0.00259 0.00264 0.00173 0.01422 0.01142 0.0067119 0.03118 0.02825 0.00735 0.00736 0.00818 0.0050020 0.02367 0.01866 0.00819

TB 1.47 3.72 9.41 4.27 4.27

T 1.60 1.45 -1.40 1.54 2.19*

tion). On the other hand, the four grade 6 biologysubtests deviated from this hypothesized order, thehierarchical order in that case being TI<T2<T4<T3. Interms of difference in magnitude of complexity load-ing, we found that Form A of the grade 9 biology testcontained items measuring at least two cognitive levels(knowing and understanding), whereas its grade 6counterpart contained items measuring at least three(knowing, understanding, and integration/application).

The implications of these findings are as follows:(1) for the grade 6 and 9 biology ILPS tests, in bothcases Form A measured more than simple factualknowledge; (2) panel experts were less capable ofcreating effective items for the measurement of ap-plication and integration skills than they were foritems which measure knowledge and understanding;and (3) the Simplex model is a useful tool for enhanc-ing test quality in that it offers a supplementary means

of verifying which cognitive processes are measuredby individual test items. However, obtaining moregeneralized results by means of the Simplex theoryrequires at least seven subtests of the same kind, withthe lengths of the subtests being as equal as possible.

In this paper we proposed the use of dendrogramsto investigate whether ILPS test items with first factorloadings 0.30 maintain a homogeneous structure.Results show that the dendrograms associated with allof the 10 ILPS tests presented in Fig. 4(a) and (b)revealed a lack of distinct group structure. This find-ing suggests that: (1) the traditional criterion ofa,1F10.30 can be established during the initial stageof scale construction for the purpose of selecting amore homogeneous set of test items; and (2) dendro-grams can be used as graphical representations ofsimilarity structure among homogeneous item sets. Itis also possible to further distinguish a fairly homo-

62

.


Table 7(c). Stout's Unidimensionality Test for Biology and Physics Items with a ai1F10.40

K

TLBiology-9 (n'=60)

TB TL,

Physics-9 (n'.25)TB

6.2U,k Sk ai Sk al, 2U,k &2U,k Sk

1 0.05699 0.03753 0.01114 0.04655 0.03292 0.01058 0.02898 0.02027 0.01145 0.01158 0.00866 0.005102 0.06713 0.03915 0.01285 0.04364 0.03970 0.01068 0.02907 0.02460 0.00721 0.01886 0.01498 0.004903 0.07526 0.03729 0.01068 0.04844 0.03903 0.01015 0.02926 0.02892 0.00510 0.02900 0.02185 0.004124 0.07544 0.03653 0.01729 0.06454 0.04028 0.00860 0.03189 0.03189 0.00447 0.03142 0.02550 0.004365 0.06781 0.03639 0.01153 0.04868 0.04006 0.00927 0.03585 0.03249 0.00469 0.04107 0.02981 0.006166 0.07916 0.03350 0.01367 0.04910 0.04054 0.00906 0.02573 0.02884 0.00436 0.04489 0.03175 0.006487 0.05665 0.03094 0.01140 0.04148 0.03951 0.00707 0.02860 0.02848 0.00458 0.04132 0.03578 0.005108 0.04778 0.02820 0.01005 0.05012 0.03988 0.00938 0.03070 0.02829 0.00693 0.04966 0.03715 0.006089 0.06666 0.03302 0.01396 0.05280 0.03662 0.00975 0.04296 0.03088 0.00825 0.04633 0.03961 0.0059210 0.04675 0.02376 0.01581 0.05281 0.03606 0.00883 0.04694 0.03360 0.00819 0.04917 0.03853 0.0060011 0.03500 0.02514 0.00748 0.03991 0.03744 0.00714 0.04607 0.03419 0.00671 0.04447 0.03512 0.0060812 0.03387 0.02298 0.00849 0.03606 0.03308 0.00707 0.06337 0.03550 0.00933 0.03333 0.03142 0.0049013 0.02752 0.01957 0.00648 0.03966 0.03341 0.00693 0.05831 0.03697 0.00860 0.03674 0.02918 0.0052914 0.02816 0.01871 0.01054 0.02702 0.02593 0.00748 0.06842 0.03552 0.01077 0.02969 0.02535 0.0067115 0.02436 0.01808 0.00872 0.02928 0.02735 0.00656 0.05963 0.03526 0.0098516 0.03530 0.02137 0.01058 0.03072 0.02384 0.00748 0.05920 0.03044 0.0084317 0.01811 0.01608 0.00600 0.02642 0.02309 0.00671 0.04294 0.02393 0.0088318 0.00832 0.01043 0.00346 0.01886 0.01830 0.00539 0.02420 0.01766 0.0056619 0.01157 0.01081 0.00447 0.02730 0.01685 0.00985 0.01538 0.01274 0.0051020 0.00855 0.01027 0.00374 0.01280 0.01245 0.00469 0.00968 0.00901 0.0050021 0.00910 0.00797 0.00436 0.02239 0.01299 0.0092222 0.01233 0.00952 0.00510 0.01803 0.01607 0.0066323 0.01752 0.01489 0.00806 0.01823 0.01280 0.0074224 0.00955 0.01134 0.00361 0.01911 0.01524 0.0064825 0.01036 0.00676 0.00592 0.01379 0.01353 0.0072826 0.00710 0.00680 0.00443 0.00942 0.00887 0.0048027 0.00319 0.00340 0.00245 0.01031 0.00928 0.0051028 0.00527 0.00634 0.00332 0.00628 0.00826 0.0033229 0.00570 0.00449 0.00436 0.00810 0.00862 0.0045830 0.00437 0.00483 0.00300 0.00359 0.00392 0.0028331 0.00397 0.00191 0.00490 0.00822 0.00884 0.0042432 0.00342 0.00370 0.00300 0.00603 0.00775 0.0031633 0.00202 0.00208 0.00224' 0.00530 0.00419 0.0041234 0.00368 0.00388 0.00265 0.00784 0.00508 0.0052035 0.00198 0.00198 0.00245 0.00542 0.00403 0.0045836 0.00075 0.00073 0.0014137 0.00084 0.00082 0.0014138 0.00173 0.00168 0.00224

71=5.09 T5=3.30 T L=5.60 T 8=4.9

T=1.26 T=0.49

geneous item set from a fairly dissimilar one by meansof examining a dendrogram's distance coefficient val-ues.

We also proposed the use of Stout's method(Stout, 1987) for investigating the extent to whichILPS test data holds unidimensionality via a compari-son of results based on the rules of thumb stated byReckase (1979) and Lord (1980). We found that the10 ILPS tests with items showing an ai1F1_?.0.30 sat-isfied Reckase's unidimensionality criterion. On theother hand, the data for the grade 9 biology, physics,and earth science tests failed to comply with Lord'scriterion, with the first two also failing to meet Stout'sdefinition of essential unidimensionality. However,the deletion of items with ailF1<0.40 resulted in thegrade 9 biology and physics tests meeting Stout'scriterion.

It therefore appears that, relative to Stout's cri-terion, Reckase's criterion is lenient while Lord's isconservative. However, it must be pointed out thatStout's definition of unidimensionality is based on

1 Ui 101' 0 , which implies thatn(n 1 ) l n I C°V(Uiitems are of essential unidimensionality if testdata support 1/0: d=1. On the other hand, for thosewho prefer the traditional definition of unidimension-

ality based on g(u)=.1- g(u 10).f(0)d(9) for all u, it may

be preferable to apply Lord's rule to assess a test'sunidimensionality.

Two other noteworthy suggestions are implied byour findings. The first is that the assessment ofunidimensionality by means of inspecting whether testitems measure the same content area (as panel expertsare inclined to do) is unacceptable. For example, ournatural science test data sets composed of items whichmeasure six diverse content areas (shown in Table 1)held an assumption of unidimensionality regardless ofwhich assessment criteria was used. Thus, from alatent-theory point of view, test unidimensionality asdefined by (5) or (7) requires a statistical assessmentprocedure rather than one based on a panel study. Thesecond suggestion is that the iterated deletion of test

63

68_

M.H. Lin et al.

items is necessary to assure test unidimensionality;this requires that the first drafts of all tests be two tothree times longer than their final versions. It thenbecomes important to note that Stout' s procedure canonly be used to analyze data when test length is aminimum of 25 items and sample size is a minimumof 750.

Finally, we described a procedure for showinghow the four statistical methods discussed in this papercan function in the development of ILPS tests whichcomply with the criteria set by the commissionedproject (Shin & Cheng, 1994). According to thisprocedure: (1) a rigid panel study is required for evalu-ating and identifying the cognitive process to bemeasured by specific items. (2) During the preteststage, Simplex models can be used to examine whethera set of subtests or a set of individual items conformsto a quasi-simplex hypothesis. (3) The a11f'1?_0.30component analysis criterion may be used for findingand eliminating heterogeneous items, after which theremaining item data are subject to a cluster analysiswhich produces a dendrogram for ascertaining homo-geneity. (4) Stout's procedure may be used for testingthe unidimensionality of test data; if the data rejectthe null hypothesis that 1/0: d=1, then it is possibleto further delete items with ai1F1<0.40. (5) If test datasurvive Stout' s significance test for unidimensional-ity, it is possible to use IRT models for calibratingILPS test items.

Acknowledgments

This research was supported by a grant from the NationalScience Council, R.O.C. (NSC 84-2511-S-001-001). Sections ofthis paper were presented at the Second Chinese Conference onProbability and Applied Statistics, Taipei, 1996.

References

Cheng, Y. J. & J. T. Lin, (1994) Study of indicators of scienceeducation: Learning progress in biology (NSC Report No. 83-011 l -S-001-001). Taipei, Taiwan: National Science Council,R.O.C.

Cormack, R. M. (1971). A review of classification. J. R. Statist.Soc., Series A, 134(3), 321-367.

Everitt, B. (1980). Cluster analysis: An SSRC review of currentresearch. New York, NY: John Wiley and Sons,.

Fang, T. S. (1994). Study of indicators of science education:learning progress in chemistry (NSC Report No. 83-0111-S-001-021). Taipei, Taiwan: National Science Council, R.O.C.

Guo, H. M. (1994). Study of indicators of science education:

learning progress in physics (NSC Report No. 83-0111-S-001-020). Taipei, Taiwan: National Science Council, R.O.C.

Guttman, L. (1969). A new approach to factor analysis: The Radex.In: P. Lazarsfeld (Ed.), Mathematical thinking in the socialsciences. New York, NY: Russell and Russell.

Harman, H. H. (1976). Modern factor analysis (3rd Ed.). Chicago,IL: University of Chicago Press.

IAEP Sampling Manual (1990). Rockville, MD: Westat, Inc.Jardin, N., & Sibson, R. (1971). Mathematical taxonomy. New

York, NY: John Wiley and Sons.Rireskog, K. G. (1970). Estimation and testing of simplex models.

The British Journal of Mathematical and Statistical Psychol-ogy, 23, 121-145.

Lazarsfeld, P. F., & Henry, N. W. (1968). Latent structure analysis.Boston, MA: Houghton-Mifflin.

Lin, C. S. (1992). Psychological and educational statistics. Taipei,Taiwan: Dong Haw, Inc.

Lin, J. H. (1994). Study of indicators of science education: Learningprogress in earth science (NSC Report No. 83-0111-S-001-022). Taipei, Taiwan: National Science Council, R.O.C.

Lin, M. H. (1995). Investigation of the structure and dimensionalityof ILPS tests items. (NSC Report No. 84-2511-S-001-001).Taipei, Taiwan: National Science Council, R.O.C.

LISREL 8: User's Reference Guide (1993). Chicago, IL: ScientificSoftware International, Inc.

Lord, F. M., & Novick, M. R. (1968). Statistical theories of mentaltest scores. Reading, MA: Addison-Wesley.

Lord, F. M. (1980). Applications of item response theory to prac-tical testing problems. Hillsdale, NJ: Lawrence Erlbaum As-sociates.

Mao, S. L. (1994). Study of Indicators of Science Education:Learning Progress in Science Process Skills (NSC Report No.83-0111-S-001-023). Taipei, Taiwan: National Science 'Coun-cil, R.O.C.

Murnane, R. J., & Raizen, S. A. (1988). Improving indicators ofthe quality of science and math education in grades K- 12.Washington, D.C.: National Academy Press.

Odden, A. (1990). Educational indicators in the United States: Theneed for analysis. Educational Research, 19(5), 24-29.

Reckase, M. D. (1979). Unifactor latent trait models applied tomultifactor tests: Results and implications. Journal of Educa-tional Statistics, 4, 207-230.

Shavelson, R. J., McDonnell, L., Oakes, J., Carey, N., & Picus, L.(1987). Indicator systems for monitoring math and scienceeducation. Santa Monica State, CA: RAND Corporation.

Shavelson, R. J., Carey, N. B., and Webb, N. M. (1990). Indicatorsof science achievement: Options for a powerful policy instru-ment. Phi Delta Kappan, 71, 692-697.

Shin, S. H. and Cheng, Y. J. (1994). Study of indicators of scienceeducation (NSC Report No. 83-0111-S-003-052). Taipei, Tai-wan: National Science Council, R.O.C.

Smith, M. S. (1988). Educational indicators. Phi Delta Kappan,69, 487-491.

Stout, W. (1987). A nonparametric approach for assessing latenttrait dimensionality. Psychometrika, 52, 589-618.

Stout, W. F.(1990). A new item response theory modeling approachwith applications to unidimensionality assessment and abilityestimation. Psychometrika, 55, 293-325.

64


AR MPS it! Mf#Bdift

40** OM!** -Vv/M** -#-Imm** 407-x--**

**

* ,=PAfF9106011431-4-4AA** AfffiNat*wq-41

*3aill-f31-4-4-4-PRARte ANiiRM-43%It't , IJILPSIjgE 411 a mts ggrimit ff sit yattst pj Quasi-Simplex model tfklAtt`o ILPS ajgOz,oubtesto wisa , OfIEMAYV-C-01 ill It FP aNI.11(MECIMA-A-014 < T<ff < eVZICEPIA

ffiliZ/J\ FP aftIJAA-VA < TifiT<ev <1011113WV)Inrf tliATiFFIEMXIM7T-4ffiReV3-)-tilJUVWff -4. Z.fill.VM-AaTITtRATRNIJAWEtaft(35A3 1101, RAMAVA-07T-\. 4-'444AINMEil<i0l4RT

AVIMC)-(OtiltAli':P1M-A/EZSAIReVIN113* ILPS illleZriftgal:&I/j3ti3-111Z3-l-thZ*VEORTiffilf (a1lF10.30)PM*Mwakg It1S3-1-tfifi,14

woo); mignromw, Ni0:-TREAXIAIRItM4M iEt_4111,3,M*T (1)113* ILPS NOWA/T. Reckasevttfil-Rm , EZZM VVittAVI&IVAIZ. 20% ; (2) filiA 43A- EfftimetzT/-

Lord ZVeTHA lEIM=IC3).111*ZIlltiMEXititicti:IttknIAI*Z4VAtift ; (3) fkil stout riMMA.-44 , qltfoRtro ILPS 0,10U 8 IftfflIJOA micuatati-3-a ; (4) it=fft ILPS 9110 1).(

a11FIN).4o wigm nam flIMAZYPAIrla4013)-45MX 21Kff913.ta#AN.R3)-tRILPS Z.UrfRNE* IS ultm*frim rota* -Sfil IRT 444-t

65

70

Announcement

4-f1414-M-4(--Ak_k_A-4--_=-4tA 9

_EA )01 )tA_ 0

The Proceedings of the National Science Council

(Part D) will be changed from publishing twice

per year to three times per year (January, May,

and September) since Vol. 7, No. 1.

rf A :W 4101;6111*4f-ri.ts4

* A PrTARO_

tstJames BensenBrmidji State University,Bemidji, MN, U.S.A.






: IAM10636*-nt-V4=g1-1063N,174: (02) 737-7594 737-7595 Email

'WA : (02) 737-7248

JJiJ:#4[ti&EP

f-`1ALA1011-V4i=&293-33A-1-11Z

VI : -I A itht AlLMIMMPrf1619A1I 111 Y541040843-A

rti*A OJMARAM2OW,H-PC'ft** di MT*MA *V% fait VIMAIYM17713-A***A ArkfM**M1413N,

1'4

DIA Ixl:OAVA0

4FYI

LgAlan J. BishopMonash UniVersityVictoria, Australia


Wribrrt

Os _IA





: [email protected]@nscpo.nsc.gov.tw

: (02)732-1234

,t4 : (02)361-7511(02)361-7511

: (02)3821394: (04)2260330: (04)725-2792

: (07)332-4910

2T NT : rc3 P11*

liits*.flIMk .X 44V,Ehtiiii-MiiiifilM*H3--t 4AREWthEr0111-MiiiifilMINTGriMg 1. lalNifPrIfil)flOOP,IIROIRP0101222-5V

2. .77,E1 .LAMM*3-c./K-ft °

(02)737-7764

4=i0V-ET-fififiRW&*02-an&Orlf, 4--MJLEEIAq:1 JL --L Pf4, fiE lit

;imam

7 2

ff,th10. 0E4A-02.2068860025

ItO #cgAREIM-*#A

11#1 WrattaMrAitilMnINt EIS (4ict 1

&710-MaiiffilkfrAT 1:11±0WAStinbantat.WWPIA 10

121/J 2.f3F5-i.

*AP Pr 24

±0101M-TA4A3igflIOg nig MgA a James A. Shymansky *A4): 4-2* 38

tedm ILPSTIMUffit Ma tilt 14415-* -ti/kA A4t 2Y-f--J4 i*Ji3c1; 46

WPM *SII9V21

hrca

JIM IMDMOM

e Ya(DED6 &MC% COMA 1mTunIk

at emat Science,7-1

echo ogf

huaa

CONTENTS

Linking STS Teacher Development and Certification (Invited Review Paper)Cheng-Hsia Wang 67

Continually Expanding Content Representations: A Case Study of a Junior High SchoolBiology Teacher

Sheau-Wen Lin 77

Expert Opinions Concerning a Taxonomic Structure for the Curricular Organizationof Biotechnology

Jerming Tseng

A Study of the Concept "Living Organism" and Living Organism Classification in AboriginalChildren

Shih-Huei Chen and Chih-Hsiung Ku

An International Investigation of Preservice Science Teachers' Pedagogical and SubjectMatter Knowledge Structures

Norman G. Lederman and Huey-Por Chang

PERMISSION TO REPRODUCE ANDDISSEMINATE THIS MATERIAL HAS

BE N GRANTED BY

TO THE EDUCATIONAL RESOURCESINFORMATION CENTER (ERIC)

1

86

96

110

U S. DEPARTMENT OF EDUCATIONOffide of Educational Research and Improvement

ArE UCATIONAL RESOURCES INFORMATIONCENTER (ERIC)

is document has been reproduced aseceived from the person or organization

originating it.



PUBILBONED BV

HEI@HEi RaGiD@@TAPE, Ir&OWN

PEPUBLP© CrgAd

Coaf.0@

EDITING: Editorial Board of the National Science CouncilRepublic of China

EDITOR IN CHIEF: Chorng-Jee GuoEDITORS:Cheng-Hsia Wang Ovid J.L. Tzeng Tien-Ying LeeFou-Lai Lin Tze-Li Kang Jon-Chao HongYung-Hwa Cheng Yeong-Jing Cheng Tak-Wai Chan















EDITORIAL OFFICE: Room 1701, 106 Ho-Ping E. Rd., Sec. 2, Taipei 10636, Taiwan, R.O.C.Tel. +886 2 737-7594; Fax. +886 2 737-7248; E-mail: [email protected], [email protected]

Send subscription inquiries to: Room 1701, 106 Ho-Ping E. Rd., Sec. 2, Taipei 10636, Taiwan, R.O.C.Tel. +886 2 737-7594; Fax. +886 2 737-7248; E-mail: [email protected], [email protected]

SUBSCRIPTION RATES: NT$ 240.00 for one year in R.O.C.; US$12.0 for one year outside theR.O.C.







The list of contents is available at: http://www.nsc.gov.tw/sciedu

t


Humanities and Social Sciences, and (D) Mathematics, Science, and Technology Education. Part D is devoted to the publication ofresearch papers in Mathematics, Science, and Technology Education, covering domain/content areas such as learning and learner,curriculum and materials, instruction (including computer assisted instruction), assessment and evaluation, history and philosophy ofscience, teacher preparation and professional development, etc. Research papers of similar interests in other disciplines including:environmental, special, health, medical, and information education, etc. may also be submitted. Complete papers describing the resultsof original research that have not been and will not be published elsewhere will be considered for acceptance. Review papers providingsystematic evaluative reviews or interpretations of major substantive or methodological issues in certain disciplines or fields may also besubmitted. Manuscripts are welcome from any country, but must be written in English.


Proceedings of the NSC, R.O.C., Part DRoom 1701, 17F, 106 Ho-Ping East Road, Sec. 2Taipei 10636, Taiwan, Republic of ChinaTEL: (02) 737-7594FAX: (02) 737-7248Email: [email protected]

[email protected]



1. Title page should contain the following information:(1) Title


The department, institution and country of each author should be furnished. If there are several authors with different affiliations,authors should be paired with their respective institutes by means of superscript symbols after the author's names in this order*, **, ***, etc. Unless indicated, a single star refers to the author to whom correspondenCe should be addressed.



3. UnitThe metric or international standard system should be used fOr recording length, area, volume and weight. When used in conjunctionwith numbers, units should be abbreviated and punctuated (e.g. 3 ml, 5 gm, 7 cm).

4. FootnotesFootnotes should be kept to a minimum and indicated in die text by superscript numbers.



7. Figures and Graphs(I) There should be only one figure per sheet, and its position should be indicated in the text by typing Fig. on a separate line.(2) Figures and graphs should be of the smallest possible printed size that does not hinder clarity.(3) High-contrast computer-generated artwork, drawings in black ink on white paper or tracing paper, or high quality photographic




76



be indicated by typing Table on a separate line. Explanatory matter should be in footnotes, not as a part of the legend.(3) Short or abbreviated column heads should be used, and nonstandard abbreviations should be defined in the legend if not defined



10. References(1) References should be cited in the text by the last name of the author and year. If there is more than one reference to thesame

author(s) and the same year of publication, the references should be differentiated by a, b, etc. e.g., (Jones, 1996; Jones &Smith, 1996; Jones, Smith, Chang, & Lee, 1992; Reed, 1994a, 1994b); Jones (1996) suggested that ; Jones & Smith (1996)reported that ; reported by Smith, Jones, & Chen (1995) [first citation]; as described by Smith et al. (1995)[subsequent citations].

(2) References should be listed in alphabetical order according to the last name of author(s), patentee(s), or editor(s). Give completeinformation with the title of the paper, report, patent, etc. Do not number references.



Johnson-Laird, P. N., Byrne, M. J., & Tabossi, P.-(1989). Reasoning by model: The case of multiple quantification.Psychological Review, 96, 658-673.

(ii) For books*Kuhn, T. S. (1970). The structure of scientific revolutions (2nd ed.). Chicago, IL: University of Chicago Press.Gentner, D., & Gentner, D. R. (1983). Flowing waters or teeming crowds: Mental models of electricity. In: D. Gentner


Brush, S. (1993). A case study of learning chemistry in a college physical science course developed for prospectiveelementary teachers. Unpublished doctoral 'dissertation, Florida State University, Tallahassee, FL.


meeting of the Third International Conference on Fluidization, Henniker, NH.Edmondson, K. M., & Novak, J. D. (1992). Toward an authentic understanding of subject 'matter. In: S. Hills (Ed.),


(v) For technical reports*Tiller, F. M., & Leu, W. F. (1984). Solid-liquid separation for liquefied coal industries: Final report for RP-1411-1 (EPRI

AP-3599). Palo Alto, CA: Electric Power iesearch Institute.(vi) For patents*


III. PROOFS AND REPRINTS1. Authors will receive galley-proofs of their papers to check for errors. Proofs should be checked against the manuscript and return

as soon as possible. Keep alternations to a minimum, and if any, mark them on both galley and manuscript.2. Fifty reprints of each paper will be supplied to contribUtors free of charge. Additional copies may be supplied at the author's own

expense.

IV. PAGE CHARGEThere are no page charge for publication in the Proceedings of the NSC.

V. COPYRIGHTUpon acceptance of a manuscript for publication in the Proceedings of the NSC, the author will be asked to sign an agreement

transferring copyright to the Publisher, who reserves the copyright. No published material may be reproduced or published elsewherewithout the written permission of the Publisher and the author.

.,77

Revised April, 1997

Proc. Natl. Sci. Counc. ROC(D)V ol. 7, No. 2, 1997. pp. 67-76


Linking STS Teacher Development and Certification

CHENG-HSIA WANG

Department of ChemistryNational Taiwan Normal University

Taipei, R.O.C.

(Received March 27, 1997; Accepted May 3, 1997)

Abstract

This paper describes a performance-based STS teacher education system developed by the author.The system consists of the following components linking teacher development and certification: (1) acurriculum framework for STS teacher development; (2) a constructivist model of STS teacher devel-opment with formative evaluation; (3) a set of performance criteria of STS teachers; and (4) a performance-based certification method for STS teachers. These components were designed and developed since 1993based on a series of action research studies conducted in real teaching and learning contexts. Thecurriculum framework provides a way of thinking for teacher educators and/or teachers to identify, select,organize, and sequence STS activities for professional development. Novices need guidance; and pro-fessional development is more efficient using anchored, scaffolded, and independent practices withformative evaluation. A set of performance criteria for STS teachers was identified from a set ofinterpretive research studies; data were collected from classroom observations, in-depth interviews,teachers' journals, and portfolios. The STS teacher certification method is based on evidence obtainedfrom portfolios presented by teachers; the method is relevant to a set of stipulated criteria. These datawere collected during one or two years of professional development programs and coupled with actualteaching observations and interviews by evaluators. Performance-based STS teacher education linksteacher development and certification. The system can improve instruction for STS learning, identifyqualified teachers, and stimulate domestic educational reform.

Key Words: certification, performance-based evaluation, science teachers, STS, teacher development

I. Introduction

In the Republic of China, the standardized cur-ricula and the joint entrance examinations have hadfixed learning and teaching styles for many years. Thetraditional science teaching methods met with manyproblems (Wang, 1995a). Learners could transfer onlylittle of their learning to new real life situations (cf.Wheatley, 1991). New forms of education are needed.Experiencing in world situation is the fundamentalway for educating citizens who will take responsibilityin promoting social welfare and handling life-relatedproblems. STS education is to develop science literatecitizens with capabilities of creativity, higher orderthinking, critical thinking, value analyses, ethical andmoral consideration, decision making, and problemsolving. STS literacy cannot be cultivated by teachingmethods of the cook-book type. Learners must learnnot only the elements needed, but also make combined

use of these learning elements to solve newly encoun-tered real life problems (Rubba, 1991). Teachers arethe key to success in science education reform. Inpursuing STS education, teachers must internalize STSbeliefs and clearly understand the teaching goals. Theymust also have STS professional capabilities and bewilling to develop their own STS modules for theirstudents without relying on textbooks (Yager, 1990,1992; Wang, 1995a).

Recent educational reforms in the R.O.C. havedecentralized authority of teacher education andlicensure, so the goals of teacher education can beaccomplished only through unified approach. Learn-ing standards for students must be matched with stan-dards for teachers. Licensing requirements must assurethat all students are taught effectively. A widelyaccepted set of standards for performance of teachingprofessionals is needed to ensure consistency and com-patibility.

C.H. Wang

How should we develop teachers' professionalcapabilities? How should we monitor and reportprogress? How should data be collected and recorded(e.g., checklists, narratives, video recorded, or port-folios)? How should we evaluate teacher qualifica-tion? How should we handle bias? How should wehelp the evaluatee improve after we have identifieda profile of strengths and weaknesses?

To promote STS education requires a large numberof STS teachers. In April 1994 the author visited theUniversity of Iowa to observe a Chauttauqua STSWorkshop. In August 1994, Dr. RobertE. Yager wasinvited for a five-day workshop at Naiional TaiwanNormal University. A collaborative STS researchproject (August 1995 - July 1998) was then promotedby the NSC Science Education Division. Systematic,well planned, and integrated collaborative researchendeavors are needed to achieve the goal within areasonably short period of time, because many com-ponents of STS teacher education are interrelated. Thecollaborative project has included 10 individual re-search studies for Grades 1-12 teaCher preparation.Activities were all sponsored by the NSC ScienceEducation Division. Each investigator has (1) created(i) a typical STS module for a specific issue, subjectmatter, and type of learners and locality;' (ii) relevantSTS data-banks (including STS units, assessment tools,STS resource materials, media, etc.); () evaluated themodule's effectiveness in achieving STS learning goals;(3) examined (i) the teaching and learning process; (ii)the teacher abilities required; (iii) efficiency in pro-moting professional abilities; (4) assessed learningachievement; (5) designed a curriculum frameworkfor STS teacher preparation; and (6) considered prob-lems encountered during the research. All the mem-bers of the project have met regularly to exchangeexperiences and solve problems encountered duringthe studies.

II. A Performance-based STSTeacher Education System

STS education is different from the usual typeof curriculum, instruction, and evaluation found intraditional science classrooms. It is different fromwhat most teachers were prepared to deliver. Beliefsand the values of science teachers have been con-structed out of personal school experiences, scienceeducation they received, and their own scienceteaching experiences. Before teachers can developappropriate STS concepts (see Appendix) and putthem into practice, the teachers need opportunitiesto (1) examine their beliefs and values about res-

ponsible citizen action on STS issues, (2) find aplace of STS in school science education, (3) con-front inconsistencies in their beliefs and values aboutSTS action as a science education goal, and (4)construct more appropriate beliefs, values, andcorresponding science teaching practices (Rubba,1991).

The author proposes that STS teachers be viewedas a category in terms of a prototype structured by thesimilarity of characteristics of STS teachers to oneanother (cf. Sternberg & Horvath, 1995). A perfor-mance-based STS teacher education system wasdesigned and developed by the author to provide op-portunities and evidence of professional growth andcertification for preservice/beginning teachers fromreal teaching situations. This system consists of fourcomponents: (1) a curriculum framework for STSteacher development; (2) a constructivist model ofSTS teacher development with formative evaluation;(3) a set of performance criteria of STS teachers; and(4) a performance-based certification method for STSteachers. The relationships among the four compo-nents are shown in Fig. 1. Curricula frameworks andthe criteria for a prototype STS teacher are guidelinesfor teacher development. In turn, portfolios developedduring teacher preparation are used as a part of theevidence for certification.

The four components were developed based ona series of action research studies in authentic contexts(Fig. 1). A primal curricula framework needed forprofessional development was designed based on therelevant literature. Using the primal framework, STSmodules "Cooking Oil" (Wang & Yu, 1994) and"Detergent" (Wang & Tsai, 1994a) were developed bythe preservice teachers (Wang, 1993). The primalframework underwent continual revision and improve-ment during the development of these two modules,to become the prototype framework. "Ozone Deple-tion" (Wang & Liu, 1995), "Greenhouse Effect" (Wang,1994a), "Acid Rain" (Wang & Hsieh, 1997a, 1997b),and other STS modules were developed using theprototype framework.

The prototype framework provides a way ofthinking for educators to establish the goals for teacherpreparation, identify curriculum contents and repre-sentative modules (on global, regional, and daily lifeissues), and cultivate STS teachers' capabilities. Aset of performance criteria for -STS teachers wasidentified from group of exemplary STS teachersthrough a series of interpretive research studies in-cluding studies previously identified. Data werecollected from classroom observations, in-depth inter-views, teachers' journals, and portfolios. Twelve types

68

,79

CurriculumFramework

4,1s

Goals of TeacherPr paration


Novice

Curriculum Content

RepresentativeModules

TeacherDevelopment

- Anchored

Scaffolded

- IndependentStudy

Types of Activity

Criteria(ProfessionalCapabilities)

Teaching & LearningActivities

Teacher(Prototype View)

ICertification Method I

TeacherDevelopment

Fig. 1. Development of an STS teacher education system.

of STS activities (e.g., role playing, exploratory ex-periment; see Appendix) sufficient to cover teachingand learning activities for STS teacher developmentwere identified (Wang, 1996). It was observed thatnovices need guidance; professional development forthem was more efficient using anchored, scaffolded,and independent practices guided by formative evalu-ation using the criteria. The STS teacher certificationmethod is based on evidence obtained from portfoliosdeveloped by teachers during one or two years ofprofessional development (Wang & Hsieh,1997a,1997b), coupled with an actual teaching observationand interviews by evaluators (Stanley & Popham,1988).

Ill. A Curriculum Framework for STSTeacher Development

Because STS education is learner-centered andissue-oriented learning linked to the real lifeoutside the classroom, STS teachers need a varietyof "teach, learn, and assess" strategies to dealwith many different problems, contexts, and con-tents related to many branches of science and tech-nology.

A dynamic framework (Fig. 2) was synthesizedfrom the literature and developed through a series ofaction research studies. The framework provides away of thinking for teacher educators and/or teachersto identify, select, organize, and sequence STS activi-ties for Professional development. The frameworkconsists of the following components (Appendix)(Wang, 1994b, 1996a): (1) a rationale for STS learn-ing; (2) learning goals and elements (Fig. 3); (3) majorSTS concepts and assumptions; (4) a procedure fordeveloping learning units and modules (Fig. 4); (5)learning Phases; (6) a procedure for professionaldevelopment (Fig. 5); (7) types of STS activities; (8)learning activities; (9) instructional strategies; (10)assessment strategies; (11) performance criteria forSTS teachers; and (12) representative modules andguidelines for selecting STS activities.

Framework

Rationale

Goals

Blocks

Modules

Units

11

SupportingSystem

Cumculum De

Education System

Learning PhasesInstructional strategiesAssessment StrategiesActivity modelLearning context

ResourcesInstructional toolAssessment toolInternational/NationalCommunityUniversityBusinessResearch

No print

MediaPrint

Fig. 2. A curriculum framework for STS teacher development.

STSIssues

41,-111.

ConceptKnowledge

ProcessSkill

ExperimentalSkill

Ethical/MoralValue

Attitude

ProblemSolving

DecisionMaking

Investigation

Social/Global View

CareerAwareness

7M11111=Performance

Self-Learning

Self-Development

STS Literacy

Higher OrderThinkingCognitiveStrategies

Internal

69

"t 0

Fig. 3. Learning goals and elements.

C.H. Wang

Explore STS course Goals ofthe unitframework

1

Developfigure/textdatabase

Goals ofteaching

Learningcontext

- Or

Concepts ofthe unit

Transparencies

STS unit 4 Design 4 Slidesactivitydesign Videotapes

informationCollect

IOrganizedatabase

I. Assessment

strategies

I

dimineg

Vignettesnsions

I\

Instructional

i oFfinle.

1 strategiesPerform Videotape

activity processunit + the whole

ILearningactivity '

Create documentfiles

Manage softwarefor figure & text

Scanintocomputer

Createtextfiles

AnalyzeDiscussresults

Conclusions

suggestions

Self-assessmentPeer-reflection

1.

Higher-order thinking

Decision making

Problem solving

Creativity

Social views

Global views

Career planning

Fig. 4. Procedure for developing STS unit activities.

STS unit

(Prototype unit)

Applicatio in other Micro ching onSTS aching gro A

Teacher enhancement

Inte iew

Classroom teachingor on

group B

Fig. 5. Procedures for professional development.

IV. A Constructivist Model of STSTeacher Development

The model was developed based on the followingrationale (Wang, 1994b): (1) Teachers must havepersonal involvement with STS-related social issues

and problems, because people do not solve problemsby the logical application of decontextualized knowl-edge (Schuman, 1987). If solving a problem is de-termined by the specific situation in which the problempresents itself, then learning how to solve the problemalso has to be developed in that situation. (2) Thecurriculum development model using RDDA (research,development, diffusion, and adoption) model has beenfound to be ineffective. The division of labor betweengoal setters and goal implementers, and betweenpractioners and valuators should be eradicated (Yager,1992). (3) The teacher has to learn as a learner; learnhow to learn, how to teach, and how to constructknowledge (Cunningham, 1991). Instruction must movethe novice toward independence for professional prac-tice. This can be accomplished through anchored,scaffolded (Rosenshine & Meister, 1992) instructionor the regulation of task difficulty, gradually reducingthe instructors support as the learner' s proficiencyincreases (cf. Bruner, 1978; Pearson and Fielding,1991).

From our action research, we have observed thatSTS teacher development should go through anchored,scaffolded, and independent practices (Wang, 1995b).It is advisable that preservice teachers (hereafter calledteachers) in their senior year start to participate in thefirst and the second stages of the program; independentwork should be done during the internship period.Teachers can join in the program individually as asmall group or as a class.

The following description is for an STS class.(1)Stage 1. STS learning under anchored instruc-

tionThe teacher experiences the STS approach forteaching and learning in the context of infor-mation rich video-based anchors that serve as"macro-contexts" for STS activities; theseanchors encourage students and teachers topose and solve complex, realistic problems. Twomodules, "Cooking Oil" and "Detergent," fo-cused on health, pollution, values, science, andtechnology issues (Wang, 1994a, 1994c; Wang& Yu, 1994; Wang & Tsai, 1994a). The bestway for a teacher to understand the STS wayof learning and teaching is to actively be en-gaged in the activities that the video-basedanchors present. The activities using problem-centered cooperative learning consist of (i) in-dividual work, (ii) group discussion, (iii) roleplaying and debating, (iv) brain storming, (v)oral and written reports, and (vi) student-madegames. During the activities mapping is usedas a thinking tool for cultivating students' ability

70

til


to plan and solve problems through aconstructivist learning model. Meaningful learn-ing occurs when an individual (1) recognizesa connection between the new learning elementsand the elements previously learned, and (2)modifies his or her individual cognitive frame-work to accommodate the new ones. Studentsact as active builders, rather than passive recipi-ents, of information. Maps are constructed asmind maps for the teacher and the students toimprove instruction and learning in real con-texts. Teachers are assessed for their learningoutcomes and asked to provide feedback andself-evaluation on teaching and learning strat-egies in the STS activities (Wang, 1993; Wang& Tsai, 1994a; Wang & Hsieh, 1997a, 1997b).

(2) Stage 2. STS unit development under scaffoldedinstructionThe STS learning framework is used by theresearcher to scaffold and help novice scienceteachers working in small groups to developSTS modules (Wang & Tsai, 1994b; Wang,1995b). (i) The class selects an STS problemof their choice after brainstorming, and dividesthemselves into small groups with a sub-prob-lem for each. (ii) Each small group designs andworks on an STS unit. (iii) Each small grouppresents the unit they designed to the class fordiscussion and then leads the class activities.The activities are videotaped. Class reactionstoward the activities are collected. (iv) Teach-ers share their problems and solutions in a forumto construct explanations for their reasoning.(v) Through the exchange of ideas with peersand the teacher educator, teachers develop sharedmeanings that allow communication. Class dis-cussion transforms intuitions into public prod-ucts. (vi) Teachers perform self assessment andpeer reflection on teaching and learning strat-egies in the videotaped activities.

(3) Stage 3. Independent practiceDuring the internship teachers should be givenopportunity for independent practice in STSactivities.

During the teacher development, Stages 1 to 3,formative evaluation was performed as follows.

(1) Data collectionDuring activities the following four items arecollected: (i) Degree of participation is assessedon the basis of students' information collectionand organization, frequency of comments, andquality of comments. (ii) Achievement in subjectmatter knowledge is assessed with paper-pencil

;1 0

tests. (iii) Open-ended thinking is assessed withthe McDaniel scale of five thinking levels(McDaniel & Lawrence, 1990). (iv) Learningachievement is assessed in terms of the fivephases of STS (Wang, 1994b; see Appendix 5).These four items were collected using the fol-lowing four methods: (i) observing microteach-ing (in person and on video-tapes); (ii) reportsof peer reflection toward microteaching; (iii)interview notes; and (iv) portfolio development(the STS modules developed, the STS moduledesigns, the STS units developed, the unit taskanalyses, teachers' journals, cognitive maps,and evidence of learning achievement).

(2) Data analysisThe collected information (observation, video-tapes, audio tapes, written materials) are ana-lyzed, organized, and induced to determine theabilities constructed by the learners from thatparticular activity. An interpretive researchmethodology is used to analyze the changes inthe STS related knowledge, skills, values, prob-lem-solving ability, and teaching practices ofa teacher through an STS unit development.

The professional capability of the teacher isevaluated by the usefulness of the units designed whenused with peers, with respect to how well they promote(1) learning achievement, and (2) professional capa-bilities. If a unit is found to be useful, then the teacherdevelopment is effective.

(1) Learning achievementResearch data (Wang, 1997; Wang & Liu, 1995;Wang & Tsai, 1994a; Wang & Yu, 1994) showedthat after participating in the STS activities ofthis model, with the aid of scaffolds, almost allteachers advanced to higher levels of thinking,becoming able to present ideas on the STS issues.They gained deeper understanding, problemsolving ability, creativity, and appropriate so-cial values. There were strong positive corre-lations, expressed by percentage consistencies,among degree of participation, paper-penciltests, levels of thinking, and levels of learning,which demonstrated satisfactory learningachievement.

(2) Professional developmentThe STS teacher development was found effec-tive (Wang, 1997a) by analyzing the following:(i) the unit developed at the entry leveland the improved unit at the exit level; (ii) theunit task analysis and interpretation made by theteacher at the entry and the exit levels; (iii)learning achievement; (iv) the peer reflection

71

82

C.H. Wang

reports on micro and class teaching perfor-mances; (v) video-taped activities; and (vi) theresearcher's observation and interview docu-ments.

The research studies indicate the model improvedthe participants' professional performance (Wang,1993, 1997; Wang & Hsieh, 1997a, 1997b; Wang &Tsai, 1994b).

V. A Set of Performance Criteria forSTS Teacher

Effective science teachers have the ability tointegrate their knowledge of science content, curricu-lum, learning, teaching, and students. Teachers shoulduse their knowledge of learning to make effectivedecisions about learning objectives, teaching strate-gies, assessment tasks, and curriculum materials(National Research Council, 1996). In addition to theprofessional abilities of effective general scienceteachers, STS teachers must clearly understand theSTS philosophy and teaching goals. Also, they musthave STS knowledge, metacognitive skills, and STSprofessional ability to deal with issue-oriented andlearner-centered learning. Qualified STS teachers areviewed as a category perceived to have a familyresemblance to one another (Sternberg & Horvath,1995). STS teacher performance criteria which wereinduced from job analysis, direct observation, inter-views, and portfolios, including video-taped STSactivities, were the following (Wang, 1996b, 1997a):(1) use local resources to design and use appropriateSTS modules having a high degree of self-modifica-tion; (2) promote situated learning in an authenticholistic context; (3) arrange activities to cultivate stu-dents' habits of mind (i.e., self-regulation, criticalthinking, creative thinking, and readiness to act onSTS issues); (4) use guided instruction before inde-pendent practice and arrange learning in proper se-quence; (5) be able to provide instruction appropriatefor capabilities and learning styles of students; (6)organize students for effective learning; (7) promotemultidimensional thinking ability and problem-cen-tered constructivist learning; (8) prepare appropriateevaluation activities; (9) provide students withspecific evaluative feedback; (10) communicate effec-tively with students and encourage them to ask ques-tions; (11) monitor learning progress closely and adjustteaching strategies such as scaffolding, cuing, ques-tioning, and fading; (12) demonstrate metacognitiveskills and sensitivity in relating to learning goals withcritical, ethical, and world views; (13) promote coop-erative learning including mutual stimulation and

recognizing good points of others; (14) exhibit enthu-siasm toward students and teaching, expect learnersto achieve well, and provide an active and a positivelearning climate; and (15) act as teacher/learner/re-searcher through feedback and self-regulated learningstrategies.

VI. A Performance-based Certifica-tion Method for STS TeachersLargely in response to dissatisfaction with some

traditional fact-oriented and multiple-choice tests, per-formance-based certification methods have been de-veloped and used for teacher licensure and others.This type of certification emphasizes testing complexteaching knowledge and skills in the real class contextin which they are actually used (Swanson, Norman,& Linn, 1995).

This project has developed performance-basedcertification methods for STS teachers on the basis ofSTS teacher performance criteria. Performance-basedcertification criteria and guidelines are developed inthe forms of printed and video materials which areused to train evaluators to ascertain inter-rater reli-ability. Evaluation by licensure is summative andbased on evidence in a portfolio format provided byteachers with pieces of evidence relevant to a set ofstipulated criteria, collected during one or two yearsof a professional development program (Wang & Hsieh,1997a, 1997b), coupled with actual teaching,observa-tions and interviews by licensures. The portfolio isdeveloped in a value-added process. Through theportfolio, one can review growth and achievement, theability of teachers to reflect on work, and the abilityto establish goals for future learning to meet thecertification criteria. The portfolio includes a com-mentary in which the teacher reflects on his or herteaching as documented for formative evaluation. Theportfolio identifies areas for professional developmentof that teacher: ( I) a commentary of the real teachingcontext for learning and teaching; (2) modules pre-pared by the teacher describing learning tasks, learn-ing environments, analyses of teaching, anchoring,scaffolding, and learning across a series of activitiesas well as plans for each of the individual activitiesin the series; (3) assessments of student learningoutcomes across several learning activities, includinga commentary; and (4) any additional information theteacher believes represents his or her work as a teacher.

Steps in the performance-based certification pro-cess in this study are as follows: Step 1. Teacher's self-evaluation developed in a portfolio format. Step 2.Teachers provide portfolios to licensure. Step 3.

72 -

83


Portfolio conference. The evaluator reviews the port-folio with the teacher (evaluatee) examining his or herprofessional growth based on the set of stipulatedcriteria. Step 4. The evaluator assigns the teacher(evaluatee) an STS topic for class observation. Teach-ers are evaluated through: (1) STS modules theyprepare; (2) oral presentation of ideas behind theirmodule design; (3) class teaching or microteachingperformances on a unit in the module; (4) students'learning outcomes; (5) peer teachers' assessment; and(6) an interview by evaluators (Wang, 1997).

Through an actual teaching observation, theevaluators expect to see how efficiently the teachercommunicates with students, monitors learningprogress, and demonstrates metacognitive skills toachieve learning goals (Criteria 10-14).

VII. Conclusion

There are many issues in teacher education withlittle agreement among authorities. Yet all of themagree on some ideas: (1) Teacher development is anecessary component of teacher certification. (2)Supervision must be ongoing for teacher development.(3) Evaluation is a rigorous process. (4) Evaluatorsmust be skilled and trained. (5) All affected partiesmust be involved in the evaluation process. (6) Thecriteria for judgment must be defined, communicated,and understood by affected parties (Stanley & Popham,1988). As the National Science Education standards(National Research Council, 1996) in the USA indi-cates: "Becoming an effective science teacher is acontinuous process that stretches from preserviceexperiences in undergraduate years to the end of aprofessional career. Science has a rapidly changingknowledge base and expanding relevance to societalissues, and teachers will need ongoing opportunitiesto build their understanding and ability."

Core training should be provided to all teachersto assure successful and lasting instructional enhance-ment efforts. In this project, the curriculum frame-work developed provides a way of thinking for teachereducators and/or teachers to identify, select, organize,and sequence STS activities for professional develop-ment. From a set of our action research studies, wehave observed that novices need guidance and profes-sional development is more efficient using anchored,scaffolded, and independent practices with formativeevaluation. A set of performance criteria of STSteachers was identified from interpretive researchstudies for which data were collected from classroomobservations, in-depth interviews, teachers' journals,and portfolios. The nature of criteria-based perfor-

mance assessment is such that evidence must begathered showing that the candidate meets the stan-dards consistently. A one-time assessment of perfor-mance is inadequate (Ambach, 1996). The STS teachercertification method is based on evidence in portfolioformat provided by teachers with pieces of evidencerelevant to a set of stipulated criteria, collected duringone or two years of a professional development pro-gram, coupled with actual teaching observations andinterviews by evaluators.

The performance-based STS teacher educationthus can link teacher development and certification.The system can improve instruction for STS learning,identify qualified teachers, and stimulate domesticeducational reform. The author believes the findingscan be used in teacher development and certificationin general.

Acknowledgment

The author wishes to acknowledge the financial support (NSC84-2511-S-003-095, NSC 86-2511-S-003-011) from the NationalScience Council, R.O.C.

References

Ambach, G. (November, 1996). Standards for teachers: Potentialfor improving practice. Phi Delta Kappan, 207-210.

American Chemical Society (1988). Chemistry in community(ChemCom), Teacher's Guide (pp. xxviii-xxx). Dubuqua, IA:Kendall Hunt Publishing Co.

Bruner, J. (1978). The role of dialogue in language acquisition. In:A. Sinclair, R.J. Jarvella, & J.M. Levelt (Eds.), The child'sconception of language (pp. 241-256). Berlin: Springer Verlag.

Cunningham, D. J. (May, 1991). Assessing constructions andconstructing assessments. Educational Technology, 13-17.

Hines, J., Hungerford, H., & Tomera, A. (1987). Analysis andsynthesis of research on responsible environmental behavfor:A meta-analysis. Journal of Environmental Education, 18(2),1-8.

Marzano, R. J. (1992). A different kind of classroom: Teachingwith dimensions of learning. Alexandria, -VA:- Association forSupervision and Curriculum Development.

McDaniel, E., & Lawrence, C. (1990). Levels of cognitive com-plexity, an approach to the measurement of thinking. New York,NY: Springer Verlag.

National Research Council (1996). National science educationstandards. Washington D.C.: National Research Council.

Pearson, P.D., & Fielding, L. (1991). Comprehension instruction.In: R. Barr, M.L. Kamil, P. Rosenthal, & P.D. Person (Eds.),Handbook of reading research (Vol. 2, pp. 815-860). New York,NY: Longman.

Rosenshine, B., & Meister, C. (April, 1992). The use of scaffoldsfor teaching high-level cognitive strategies. Educational Lead-ership, 26-33.

Rubba, P. A. (1991). Integrating STS into school science andteacher education: Beyond awareness. Theory into Practice,30(4), 303-308.

Shulman, L. S. (1987). Knowledge and teaching: Foundations of

73

84

C.H. Wang

the new reform. Harvard Educational Review, 57(1), 1-22.Sia, A., Hungerford, H., & Tomera, A. (1986). Selected predictors

of responsible environmental behavior: An analysis. Journalof Environmental Education, 17(2), 31-40.

Stanley, S. J., & Popham, W. J., Eds. (1988). Teacher evaluation:Six prescriptions for success. Alexanderia, VA: Associationfor Supervision and Curriculum Development.

Sternberg, R. J., & Horvath, J. A. (1995). A prototype view of ex-pert teaching. Educational Researcher, 24(6), 9-17.

Swanson, D. B., Norman, G. R., & Linn, R. L. (1995). Performance-based assessment: Lessons from the health professions. Edu-cational Researcher, 24(5), 5-11.

Waks, L. J. (1992). The responsibility spiral: A curriculum frame-work for STS education. Theory into Practice, 31(11), 13-19.

Wang, C. H. (1993). A design for a model to develop STS module:Cooking oils and detergents (in Chinese). Research Report.Taipei, Taiwan: The National Science Council, R.O.C.

Wang, C. H. (1994a). An instructional design for greenhouse effectSTS activities. Proceedings of the National Science Council,Part D, 5(2), 1-12.

Wang, C. H. (1994b). A model for teacher enhancement throughinstructional research in STS unit development. EducationalTechnology Research, 17, 45-55. Tokyo: Japan Society forEducational Technology.

Wang, C. H. (1995a). Learning and teaching in STS activities.Chinese Journal of Science Education (in Chinese), 3(1), 115-137.

Wang, C. H. (1995b). A model of STS teacher development. In:Proceedings of the 28th Seminar on Science and Technology:Science Education (pp. 71-80). Tokyo: Interchange Associa-tion, Japan.

Wang, C. H. (1996a). A curriculum framework for STS teacherpreparation. Chemistry (in Chinese), 54(2), 103-114.

Wang, C. H. (1996b). STS teacher preparation, qualification, andcertification: a collaborative research project. In: Proceedingsof the 1996 International Conference on Science Teachers'Qualification and Certification (pp. 78-106). Taipei, Taiwan:National Taiwan Normal University.

Wang, C. H. (1996c). Representative types of STS activities:Analysis (in Chinese). In: Proceedings of the 1996 PhysicsEducation Conference (pp. 15-19). Taipei, Taiwan.

Wang, C. H. (1997). STS teacher development. Chinese Journalof Science Education (in Chinese), 5(1), 1-36.

Wang, C. H., & Hsieh, C.H. (1997a). Using cognitive maps to assesslearning outcome: The acid rain STS activities. Submitted forpublication. Bull. National Taiwan Normal University (inChinese). Taipei, Taiwan: National Taiwan Normal University.

Wang, C. H., & Hsieh, C. H. (1997b). Portfolio assessment forSTS teacher development (in Chinese). In: Proceedings of theSTS Teacher Preparation Conference. Taipei, Taiwan.

Wang, C. H., & Liu, I. H. (1995). Developing STS activities onthe ozone layer depletion. Bull. National Taiwan Normal Uni-versity (in Chinese), 40, 331-364.

Wang, C. H., & Tsai, H. H. (1994a). Promoting open-ended thinkingon the STS topic "detergent." Proceedings of the NationalScience Council, Part D, 3(1), 37-47.

Wang, C. H., & Tsai, H. H. (1994b). A model for STS professionaldevelopment through scaffolded strategies. Bull. National TaiwanNormal University, 39, 429-454.

Wang, C. H., & Yu, P. C. (1994). STS unit activity: What shouldwe do with the used cooking oil? Chemistry (in Chinese), 52(4),415-428.

Wheatley, G. H. (1991). Constructivist perspectives on science and

mathematics learning. Science Education, 75(1), 9-21.Yager, R. E. (1990). The science/technology/society moVement

in the United States: Its origin, evolution, rationale. SocialEducation, 54, 198-201.

Yager, R. E. (1992). The constructivist learning model: A mustfor STS classrooms. ICASE Yearbook (pp. 14-17). Arlington,VA: National Science Teachers Association.

Appendix:

Components of the Curriculum Framework

74

(1) Rationale for STS learningIn some traditional schooling, learners can transfer only a littleof their learning to new situations (Wheatley, 1991). Expe-riencing in a real world situation is the fundamental way foreducating citizens who take up responsibility in promotingsocial prosperity and handling life-related problems. STSliteracy cannot be cultivated by teaching methods of the cook-book type. Learners must not only learn the elements neededbut also take combined uses of these learning elements to solvenewly encountered real life problems.

(2) Learning goals and elements (See Fig. 3)(3) Major STS concepts and assumptions

Taken in part from Chemistry in Community (ChemCom)(American Chemical Society, 1988).(i) Science and technology have contributed in changing thetype of society, advancing civilization, and promoting pros-perity. (ii) Wise resource management on the planet earthaims at maximizing both short and long term benefits whileminimizing environmental and human health burdens/risks/costs. (iii) Many complex problems can be made manageableby breaking them into simpler subproblems or components.(iv) Learners may become better citizens if they are givenopportunities to (a) learn and apply fundamental scientificconcepts and methods, and (b) practice group decision makingskills in cooperative learning. (v) The learners understandthemselves as interdependent members of society and anecosystem of nature. (vi) Learners must cultivate abilities indecision making. (vii) STS activities arouse learners' interest,promote career awareness, and emphasize careers in scienceand technology they might pursue. (viii) STS activities helplearners develop and maintain effective mental habits such asself-regulating, critical, and creative thinking and learning.(ix) STS activities develop substantive understanding andbroad fundamental concepts, which cut across all disciplinessuch as system, interaction, relativity, change, stability, andequilibrium. (x) The present state of knowledge about anygiven STS issue is likely to contain some imprecision, inac-curacy, and uncertainty. Society must act upon the bestavailable information with the understanding that additionalinformation may call for subsequent reevaluation of a previoussolution.

(4) A procedure for developing learning units and modules(See Fig. 4)

(5) Learning phasesFive learning phases were synthesized by Wang (1994b) basedon the literature (Sia, Hungerford, & Tomera, 1986; Hines,Hungerford, & Tomera, 1987; Marzano, 1992; Waks, 1992).Phase 1. Awareness: Positive attitude and awareness of issues.Phase 2. Understanding: Acquiring and integrating knowl-edge. Phase 3. Proposing solution with explanations. Phase4. Responsible action: using knowledge meaningfully. Phase

85


5. STS literate creative personality: productive habits of mind.(6) Procedure for professional development (See Fig. 5)(7) Types of STS activities

(i) Information search, (ii) STS module design (includingsimulation), (iii) Brainstorming, (iv) Mapping, (v) STS game/play, (vi) Role playing, (vii) Debate, (viii) Discussion, (ix)Decision making, (x) Problem solving, (xi) Field experience,and (xii) Experimental (inquiry) investigation.

(8) Learning activities(i) Study science and/or social science concepts. (ii) Conductlibrary research. (iii) Secure data and information from gov-ernment and private agencies. (iv) Collect data from experi-mental inquiry. (v) Use social science research techniquessuch as questionnaires. (vi) Analyze information and use datato propose alternative resolutions or the issue. (vii) Weighthe pros and cons of each resolution. (viii) Clarify values.(ix) Select a resolution and decide courses of action. (x)Decide which actions individuals or members of a group mighttake and carry through with plans, action, and evaluation.

(9) Instructional strategies (anchored and scaf-folded)(10) Assessment strategies

(i) Group and individual presentation, (ii) Interviewing stu-dents and groups of students about project design, (iii) Ob-serving students in role playing and simulations, (iv) Writtenreports, (v) Student journals, (vi) Student action plans, (vii)Standardized achievement tests, (viii) Criterion referencedtests, (ix) Home work, (x) Home tests, (xi) Extended problemsolving projects, (xii) Field investigation, (xiii) Essay exami-nation, (xiv) Multiple-choice item assessment, (xv) Cognitivemaps, (xvi) open-ended thinking assessment, and (xvii) Port-folio.

(11) Performance criteria of STS teachers (See Section V)(12) Representative modules and guidelines for selecting STS

activitiesRepresentative Modules should cover global, national, local,and life-related issues including representative types of STSactivities.The learning activity designs are characterized by the fol-

lowing features: (1) student centered; (2) problem centered; (3)based on student prior knowledge; (4) linked to the world outsidethe classroom; (5) linked to social, cultural, and environmentalissues; (6) cultivate higher order thinking for decision making andproblem solving; and (7) encourage individual as well as coopera-tive learning to promote ethics and social values.

It is important that the learning can improve views on STSissues and that learners are encouraged to (1) link STS issues withother knowledge within a discipline, with other disciplines, and withsocial and global problems, (2) understand the overall structuralrelationships among the elements in the issue, (3) understand therelationships among the prerequisites already known and the presentlearning elements, (4) apply the process skills already learned toa new topic, identify sithilarities and differences, and generalize,and (5) build cognitive networks to perceive, define, analyze, support,and elaborate on a situation.

After the learning activities the teachers will be able to (1)select STS issues and contents, (2) determine the most appropriatefit for STS issues in a curriculum, (3) develop an integrated approachfor teaching STS issues, and select the STS materials, (4) selectstrategies of instruction and assessment, (5) construct STS conceptsand positive attitudes toward science and technology, (6) developthe skills of analyzing STS issues, and (7) assess the design of studyunits.

758 6

C.H. Wang

TAR.

MAeMROatt-tf*Di

fgATOR-VMV/fAM ri4J STS TFff; (USTS 0-05igit-MVPM ; (2) g*I-ANAZ STS ; (3) 4ag STS Yii;rfr4

* ; (4) J44 40 STS ° frZ. ft 1993 *I)..k5K tc -T111,11 Ift**WfiNiffAvffiOMI[tA4W* it-AVV4N1AVIEJAA iMilffit-VOIR01144t MO #1114R41M STS fPMAR-6914014-RtMgi 214f31151Alt-MWO4 ; eV 11M4AIEJA a**At*, H-Fi4s5tutifmtnt , pieumt o STS VcrOZTIff4*A *NO ffltVf91 t

ttoln,_,_Fiti-eArc4tlevig4*#40#.Nffil4(1) -VcRffNiAntt*EVC*,M-Vil (2) 115MAVMSIV- (3) ifiANOIK ° ttOffalfta43211

-,--W<SAOMJi*Rmitt,r1 virm , STSiftSTS garlitIri*ME&A STS SI-4-s- aigNilf RI AMER rAi Ii*fh

-76--


Continually Expanding Content Representations:A Case Study of a Junior High School Biology Teacher

SHEAU-WEN LIN

Department of Mathematics and Science EducationNational Pingtung Teachers College

Pingtung, Taiwan, R.O.C.

(Received April 20, 1996; Accepted April 10, 1997)

ABSTRACT

The purpose of this study was to explore how an experienced biology teacher continually expandedrepresentations her teaching career. Participant observation, interviews and various related documentswere used to collect data. The findings were presented in terms of five assertions: (1) The evaluationof representations made the teacher feel the need for change. (2) The teacher arranged representationsin an expanding sequence. (3) The teacher searched for data to form alternative representations. (4)New representations, resulted from modification, specification, combination, or invention. (5) Repre-sentations were mo&fied through pretests and classroom teaching. The incompleteness of representationsmade the content transformation act a dilemma. As Lucy managed dilemmas, she reflected on herrepresentations and continually expanded the representations. This process was based on her extensiveknowledge, and this knowledge was developed during the process. The implication of this study is thatscience teachers can reflect on their representations to promote professional growth.

Key Words: content representation, professional knowledge growth, interpretive research

I. Introduction

The entrance examination has become a majorfactor influencing the teaching practices and qualityof learning in junior high schools in Taiwan. In orderto help students get high scores on the entrance ex-amination, teachers often put emphasis on repeatedexercises that aid the retention of facts. The contentof the curriculum is usually based on textbooks andthe questions asked on examinations. Students' learn-ing life is full of examinations and memorization offacts from textbooks.

Increasingly, people have begun to reflect on theeducational quality of primary and secondary schoolsand calls for educational refotm have been proposed.They emphasize the importance of learning equalityfor all children and are asking for a higher qualitylearning environment. Despite this situtation, why doteachers still slide into.traditional routines and repeatthe things they have done before? Can teachei-s developcritical ways of understanding their teaching contextsand alter their teaching practices to enhance student

learning?Studies have proposed that content transforma-

tion is the central intellectual work of teaching(Shulman, 1986, 1987). To foster understanding ofscience, successful teachers cannot only have anundrestanding of a concept, principle, or theory ofscience. They must develop the ability to representthe content in different ways to communicate knowl-edge to their students. Through a lens of represen-tation, this study explored how an exemplary juniorhigh school biology teacher, who had succeeded inremaining enthusiastic over the course of a 13-yearcareer, continually expanded her repertoire of repre-sentations to help her students learn biology. Rep-resentations are defined in this article as ways ofrepresenting the subject to make it comprehensible tostudents in real teaching contexts.

Examining teaching in terms of representationsinstead of methods or strategies is intended to focusattention not just on activities of teachers and studentsin a classroom but on the relationship between activi-ties and the science knowledge taught. The concept

77

S.W. Lin

of representations illuminates the wholeness of teach-ing content and strategies (McDiarmid, Ball, & Ander-son, 1989), which both influence student learning. Thenotion of representations also explains subject-spe-cific properties of teaching (Shulman, 1986). Teach-ers need to consider different issues in different sub-ject mater when they transform subject matter andselect and evaluate representations. Using represen-tations appropriately demands specific knowledge(McDiarmid et al., 1989).

The results can contribute to a better understand-ing of the nature of life-long learning in science teach-ing, especially for science teacher educators and sci-ence teachers who believe that continual learning fromteaching practices plays an important role in profes-sional growth.

II. Review of LiteratureWithin the learning-to-teach literature, research-

ers examine the developmental concerns of studentteachers (Fuller, 1969), developmental stages (Ber-liner, 1988; Kagan, 1992), their roles and beliefs(Brickhouse, 1990; Duffee & Aikenhead, 1992; Munby,1984, 1986; Tobin, 1990a, 1990b) and their knowledge(Geddis, Onslow, Beynon, & Oesch, 1993; Lederman,Gess-Newsome, & Lantz, 1994; Shulman, 1986, 1987;Wilson, Shulman, & Richert, 1987). Researchers havealso tried to understand the nature of beginning teach-ers' critical transition from college students to schoolteachers and support beginning teachers' teachingneeds.

Most of the above literature puts emphasis onbeginning teachers' professional growth. However,learning to teach is a lifelong process and as oneteaches, one learns (Feiman-Nemser, 1983). Passingthe survival stage does not guarantee that teachers willsuccessfully continue to grow. The stage theories donot imply that earlier stages lead naturally to laterstages (Grossman, 1992). In a review article Britzman(1986) described the way in which teachers' personalhistories interacted with common myths of their cul-ture to maintain current teaching practices. Sheconcluded that without a critical perspective, studentteachers would lose their intention to enhance thepotential of students and slide into a cycle of repetitionof other teachers' teaching. Fefman-Nemser &Buchmann (1989) conducted interviews and class-room observations with six teacher candidates through-out two years. They reported that teachers were satisfiedwith their teaching and became less likely to criticizethe prevailing norms after mastering the sociallypatterned school routines. Some teachers in this study

paid attention to management problems, but didn' tnecessarily learn how to teach.

Science education researchers have indicated thatseveral constraints shape science teachers' develop-ment or prevent them from changing in their teachingenvironment (Abell & Roth, 1992; Brickhouse, 1993;Lantz & Kass, 1987; Loughran, 1994; Tobin, 1990a,1990b; Tobin, Briscoe & Holman, 1990; Tobin &Gallagher, 1987; Wallace & Louden, 1992; Wood,1988). These constraints included institutional andsocial expectations, accountability via test scores,curriculum, equipment, support, time, knowledge,experience, and beliefs. These researchers also sug-gested that critical thinking or reflection is an impor-tant first step if teachers are to change and improvetheir teaching. Sharon, an elementary teacher ofmathematics and science, was a participant in the studyof Tobin et al. (1990). Through reflection on herinteractions with colleagues and students, Sharonperceived a problem with the curriculum. She con-fronted constraints and sought the support she needed.Sharon became an agent for change for herself and forthe colleagues and children with whom she worked.

Researchers have acknowledged the importanceof content transformation in teaching. What is thetheoretical framework of the process of transforma-tions? In "Knowledge Growth in a Profession Project,"Shulman and his associates (e.g., Carlsen, 1991;Hashweh, 1987; Shulman, 1986, 1987) focused onhow the subject matter knowledge of novice teachersgrew and changed over time. As they observed andconversed with teacher collaborators, they found thatnovice teachers struggled to develop many differentways to explain the content of their disciplines in orderto help their students learn (Shulman, 1986, 1987).They constructed a theoretical framework for under-standing the reasoning process of transformations. Toaccomplish transformations, teachers draw on peda-gogical content knowledge. Pedagogical contentknowledge is "a blend of pedagogy and content whichincluded an understanding of how the topics in instruc-tion were related and how they were most effectivelyorganized and presented in the classroom context"(Shulman, 1987). Pedagogical content knowledge alsoincludes knowledge of alternative representations ofa particular subject and knowledge of understandingand misconceptions of a subject. This knowledge is

. developed in a cyclic process in which teachers com-prehend, transfer, instruct, evaluate, reflect, gain newcomprehension, and transfer again. The transforma-tion process involves four subprocesses: interpreta-tion, representation, adaptation, and tailoring to stu-dent characteristics (Shulman, 1986, 1987).

78

Continually Expanding Content Representations

In science teaching a growing body of recentresearch has utilized Shulman' s concept of pedagogi-cal content knowledge. Geddis et al. (1993), forexample, described two student teachers' attempts atteaching the subject of chemical isotopes. The re-searchers collected data by interviews and classroomobservations. In the course of analysis, examples offour distinct categories of pedagogical content knowl-edge were articulated. Knowledge of learners' priorknowledge, effective teaching strategies, alternativerepresentations, and curricular salienCy were all im-portant components of the pedagogical content knowl-edge that beginning chemistry teachers needed toacquire. The researchers claimed that the componentsof pedagogical content knowledge played importantroles in the task of effective transformation.

Most of these studies beginning with Shulman' ssuggested that pedagogical content knowledge playsimportant roles in transformations of subject matter.Researchers have tried to articulate the' concept ofpedagogical content knowledge by exploring knowl-edge growth in student teachers. They paid littleattention to the dynamic aspect of pedagogical contentknowledge and to continual growth in experiencedteachers. Interviews and observations were majormethods of collecting data. However, many of thestudies were conducted in laboratory settings (Hasweh,1987; Marks, 1990). Transformations are complicatedin real teaching contexts. There are many factors thatinfluence transformations of subject matter. It seemsthat a true picture of a teacher' s expanding processof representations can only be derived in the contextof a naturally occurring classroom setting.

This study collected data through extensive class-room observations, teacher interviews, and review ofdocuments. It was sought to explore the processthrough which an experienced biology teacher con-tinually expanded her representations and promotedgrowth in real teaching contexts. This process wasnot the trial-and-error approach that beginning teach-ers adopted. The teacher' s goal of expansion was nota struggle to survive, but to renew and search for themost effective way to attain worthwhile goals.

III. Method

1. Selection of Teacher

The teacher in this case study was selected froma list of candidates containing exemplary biologyteachers nominated by science education experts andscholars. An observation of potential participants'classroom teaching was also conducted to select a

teacher who displayed expert characteristics in biol-ogy teaching and an expansion process of contentrepresentations. "Criteria of Excellence: BiologyTeachers of Junior High School" (see Lin, 1994)describes expert characteristics in biology teaching asan accredited subject. The teacher and the schooladministrators were asked to cooperate. Through thisprocedure the teacher, Lucy, was selected.

2. Participant and Context

Lucy, a female teacher, had been a biology teacherfor 13 years. She had won many awards. Someexamples were the Taipei Municipal Annual Outstand-ing Teachers' Award, Award for OutstandingAchievemnt in the Taipei Municipal Science Fair, andAward for Outstanding Achievement in the TiapieMunicipal Teaching Aids Presentation.

Lucy worked in a school located in Taipei City.It was a co-educational public school where most ofthe students' families were of a working-class back-ground with middle socioeconomic status. There werefive biology teachers and 21 seventh grade classes inthe school. Thirty-eight students, 18 boys and 20 girls,of mixed abilities were in the participating biologyclass.

3. Data Collection

Data was collected by means of observation,teacher interviews, and review of documents. Obser-vations were made during twenty-eight 50 min lessonsof the participating class, which were held continu-ously from March to June of 1992. The teacherinterviews regarding content representations wereconducted before and after each individual observa-tion. An occasional lunch or tea time was arrangedso that Lucy could be interviewed in a more casualsetting. All observations and teacher interviews werevideo-recorded or tape-recorded. Related documents,including the teacher' s lesson plans, notes, transpar-encies, work' sheets, maps, charts, pictures, and tests,were preserved with photocopies or photographs.

4. Data Analysis

The data base consisted of field notes and tran-scriptions from observations, interviews, and relateddocuments. Excerpts were taken from the field notesand transcriptions to describe teaching practices andtentative assertions concerning the practices. Theexcerpts were discussed with Lucy regularly through-out the study to confirm the significance of her be-

79

9 0

S.W. Lin

havior, and Lucy was also asked to comment on anyideas that she thought to be misrepresented or incom-plete. Then, all the excerpts were coded and classified.Major categories included forms of content represen-tations, processes of expanding representations, andteacher knowledge. Within each category were manysubcategories, in the form of content representationssuch as metaphor, discussion, explanation, hands-onactivity ... etc. All the excerpts were examined fortrends and frequency. At this time tentative hypoth-eses about the processes of expanding content repre-sentations were formed. Then the specific trends wereexplored and more concrete hypotheses were formu-lated and tested with subsequent coded data fromdifferent data gathering methods. Contradictory datawas sought to revise the hypotheses. Reliability checkfor coding were conducted with other researchers.(For more information on the data analysis, see Lin,1994.)

IV. Findings

The findings of this case study are presentedbelow in five assertions.

1. Assertion 1. Evaluation of RepresentationsCaused Lucy to Feel Dissatisfied and ThinkThere Was a Need for Change.

Lucy constantly changed her representations forthe same concepts throughout her teaching career. Sheexplained the used representations during interviews.Why did she change them? Lucy described her mo-tivation to change her representation of "relationshipbetween genes and traits" as follows:

I had used different color glasses to interpret the relationshipbetween genes and traits for many years. One day I stood inthe back of the classroom, and I saw that the color of twooverlapping red glasses was deeper than the one made by atransparent one and a red one. I was shocked by this phenomena.

It might mislead students into thinking that the traits of homo-geneous and heterogeneous combinations were different. I

decided to correct this immediately.

If Lucy did not know the disadvantages of herrepresentations, she would not feel the need to changethem. There was not a repertoire of representations.

What was a "good" representation? For Lucy allrepresentations had to be feasible in the teachingcontext. A representation also had to provide authen-tic information 'and be helpful to student learning.Lucy tried her best to transfer content but in practice

circumstances were too complex to achieve all criteriasimultaneously. Scientific information or scientists'explanations were always too technical for her stu-dents to comprehend. Though stimulating students'learning interest and making learning easier, transfor-mations could not avoid distorting the meaning of thescientists' knowledge. For example, Lucy sometimesgave explanations in terms of purposes and intention(e.g., the male peacock blows its tail open to impressand attract the female) to help students learn. She said,"These anthropomorphic explanations may misleadstudent learning. Some students will think that ani-mals, like human beings, have intentions."

Reflection on representations made Lucy see thecharacteristics of her teaching. Lucy knew that dif-ferent representations could present different facetsand dimensions of science. She said that some rep-resentations were closer to scientists' knowledge andsome were easier to learn. She also mentioned thatdifferent representations could attract students' atten-tion and meet individual needs. Lucy stated, "Everyrepresentation has its own advantages and limitations.None of them are perfect." This incompletenesg ofrepresentations made Lucy feel they were unsatisfac-tory and there was a need to create other representa-tions.

2. Assertion 2. Lucy Arranged Representationsin an Expanding Sequence.

Every representation had its own advantages andlimitations. It seemed that there was no representationwhich could represent all the meanings that Lucywanted to present. Which one would Lucy choose topresent? This was a dilemma. Presenting multiplerepresentations for one concept was the strategy thatLucy selected to deal with this dilemma. Lucy said,

I try to transform the content into different forms just like acook who use the same meat or vegetables to prepare manydifferent dishes. My students can s.elect helpful forms to aidtheir understanding just as customers choose their favoritedishes to eat.. No matter which one or ones they choose, theyall learn the concept.

However, the expanding process was time con-suming. Lucy could not transfer all concepts intomultiple forms at the same time. Therefore she ar-ranged them in a sequence to form alternative repre-sentations that would help students learn.

Lucy gave priority to forming multiple represen-tations for difficult concepts. She noted students'common difficulties in learning biology. Most of these

80

91


concepts were within genetics, cell biology, and evo-lution. The common characteristics of these difficultconcepts were complexity, having too many terms,being abstract, and not being observable with the nakedeye.

About the unit of "genetics" Lucy said,

Genetics is one of the difficult topics for the students. Studentsneed the concepts of mitosis, meiosis, chromosomes, genes,sexual reproduction, asexual reproduction, genotypes, pheno-types, probability, ... etc. Students seldom have had a chanceto observe these phenomenona before. These concepts are alldifficult to learn. In particular, most of my students do notdevelop the ability to manipulate problems of probability.

As a consequence, it was a priority that difficultconcepts be presented in different ways. For example,when teaching "mitosis", Lucy explained six represen-tations. They were teacher explanation of the rulesof mitosis, hands-on activity with mitosis cards, illus-tration of onion mitosis, illustration of fish mitosis,reading paper, and discussion. Lucy used about 70pictures in the above teaching activities. She empha-sized visual learning and hands-on activity in this case.

3. Assertion 3. Lucy Searched for Data to FormAlternative Representations.

Lucy required a body of knowledge that wouldcontribute to her expanding collection of representa-tions. Examples were knowledge of subject matter,students and learning, curriculum, teaching media,alternative representations, and context. She obtainedthis kind of knowledge from her personal learningexperiences, teaching experiences, textbooks, otherteachers, volunteer worker training, in-service train-ing, and students' responses.

Subject matter knowledge was the content ofrepresentations. Lucy had received a bachelor degreein biology. She had also earned some master levelcredits in biology. Lucy already had extensive knowl-edge of biology. Most of the subject matter knowledgecam from her personal learning experiences. How-ever, Lucy did not think that she had enough subjectmatter knowledge for teaching secondary school bi-ology. She said, "Because of the diversity of biology,I cannot answer all questions students ask. I need tocheck it out in books." She told stories of biologists,for example Charles Darwin, Louis Pasteur, and GregorMendel. Lucy emphasized environmental issues in herbiology class. She showed an environmental roomfilled with ten environmental problems occurred atTaiwan. This kind of knowledge did not come from

her formal education but rather from books, journals,newspapers, volunteer worker training and in-servicetraining. Lucy' s subject matter knowledge, especiallyabout science-technology-social issues, everyday life,and stories of biologists, was developed through theprocess of expansion.

Lucy needed knowledge about students and learn-ing to communicate effectively. Lucy was concernedabout what students knew about science knowledge,reasoning skills, and life experiences. When teaching"genetics", Lucy reviewed related topics in elemen-tary school to find out what her students had learned.She asked a mathematics teacher about what her stu-dents had learned about probability. She also inter-viewed some students about the concepts of genetics.When these were ascertained, she designed activitiesfor students in order to relate the new content to theold.

Lucy required knowledge concerning intendedobjectives, activities and textbook problems. Sheselected the proper content, put it in sequential order,linked concepts from different topics or lessons, andmanipulated the pace. Most of this kind of knowledgecame from her teaching experience, textbooks, andteacher' s manuals.

Knowledge of contexts made Lucy adjust herideals to the situation in the real practice environment.Lucy was familiar with the expectations held by society,the school, parents, and students after thirteen yearsof teaching. For example, the expectation of highachievement in examinations influenced her instruc-tional emphasis and test content and made her increasetest frequency. She knew the hardware and softwarein school and out. This had an impact on Lucy' sdecisions to use materials and media.

Knowledge of alternative representations allowedLucy to conduct lively transformations. Lucy activelyattended teaching conferences, in-service training, andvolunteer worker training to get new ideas. She alsodiscussed her teaching with her colleagues and teacherfriends and thereby change her representations.

Lucy knew who and where would provide re-sources for teaching, and she would learn new methodsand knowledge for improving her instruction when shefelt the need. University libraries, science museums,educational data centers, botanical gardens, zoos, andnational gardens were the places where she or herstudents usually visited to get information. Lucy alsobuilt good relationships with university professors,consultants of teachers' in-service training centers,and colleagues who could provide assistance andsuggest alternati ves to help solve teaching pro-blems.

81

92

S.W. Lin

4. Assertion 4. New Representations Resultedfrom Modeification, Specification, Combi-nation, or Invention.

If Lucy felt the need to change her representa-tions, she tried to modify her representations fordifferent students and teaching conditions. There weretwo kinds of modifications: inter-form modificationsand intra-form modifications. The same content wastransferred to different forms in the inter-form modi-fication. For example, because of difficulty in han-dling students' responses, Lucy changed an illustra-tion of "flowers" with the aid of slides to a studentdiscussion with the aid of pictures. Lucy would onlyadd, delete, or replace content and not change theform in intra-form modification. Giving students afamiliar frog instead of an unusual salamander as anexample of the amphibians was a case of intra-formmodification. Modification seemed to be the mosteconomical way to increase the effectiveness of rep-resentations.

Besides modification, Lucy also tried to createnew representations. A new one could result from aspecification, combination, or invention.

A specification was when Lucy borrowed theform of an existing representation from instruction inother subjects. Then she transformed a biology con-cept into the same form to generate a new represen-tation. For example, a form Lucy used to help herstudents construct the concept of "classification ofliving things" came from a professor's representationfor "shape". Lucy said,

I taught the "classification of living things' according thecontent in the textbook for many years. I pointed out charac-teristics and names of animals and plants for each ICind. Students

were always confused by the names and charactefistics of somany living things. There were communication problemsbetween the teacher and students and between students andstudents. What students needed to do was to memorize I

remembered a lively professor who showed us many differentshapes but only answered yes or no questions. We all learnedthe classification scheme, characteristics and names of theshapes. Why shouldn't I use this form to teach "classificationof living things"? I prepared pictures of different living thingsand showed them in my classification class. I asked studentsto raise questions that could be answered by yes or no to helpthem construct the classification scheme.

Lucy would combine representations to generatea new one. In the "mitosis" unit, Lucy combined ateacher demonstration of a model of mitosis with astudent discussion on the characteristics of mitosis to

form a student manipulation of pictures of mitosis.Combination always occurred in the condition thatLucy taught the same or related concepts.

Lucy also invented new representations. Lucydescribed the development of her representation of the"relationship between genes and traits" as the follow-ing:

After deciding to change the representation of "relationshipbetween genes and traits", I thought it over all day, every day,even while walking and during daily activities. One day anidea flashed into my mind. 'Ah ha, I got it' I said.

The expanding process took time. Sometimes,Lucy did not have any new ideas and her thoughtsstayed in a latent stage. She did not like her strategyof repeated exercises to help students to learn theconcept of "probability in genetics", but she did notknow of any better strategy. Therefore, she concludedthat repeated exercises were her best choice and pre-sented them in her teaching. This did not mean thatLucy gave up. She worked hard to struggle with it.

5. Assertion 5. Representations Were ModifiedThrough Pretests and Real Teaching Prac-tices.

To make sure of the feasibility of new represen-tations, Lucy would ask some students to do the tasksin a simulated teaching environment before using them.She observed students' responses and asked questionsabout new representations. Then she tried to changethem. As an example, Lucy asked five students to doa worksheet on "asexual and sexual reproduction ofplants" while out of class. She collected informationabout its readability and how much time was needed.According to students' reactions, she modified someitems on the worksheet.

All the above actions were still during the plan-ning stage. Many other factors could affect the successof a representation in a real classroom. Representa-tions must be tested by direct teaching.

Teaching behavior was a focus that Lucy evalu-ated when she presented a new representation in theclass. Lucy said,

' 82

Whenever presenting the new materials or activities, I alwayspaid attention to myself. During the class I asked myself, "What

will I do next? Is the explanation clear enough?" I was busy

with looking for pictures or specimens that I wanted to illustrate.I spent much time on managing the class. After teaching thesame topic for the second or third time, I would have enoughtime to be attentive to student learning behavior.

93


After maintaining a fluid teaching flow, Lucyturned her attention to students' responses. Lucyanalyzed test results, homework, and worksheets, keptan eye on student behavior, and conducted studentinterviews to collect learning information. She wantedto make sure that the representation could help stu-dents understand.

Lucy reviewed her teaching behavior, the teach-ing context, students' behavior, classroom manage-ment, and time arrangement. Depending on her memoryand feelings about classroom events, she analyzed theevents and tried to find factors which influenced thesuccess of the representations. Later she reconstructedthem for future instructional action. These activitieshappened both during class and afterwards. Whilediscussing "the results of Hydra's asexual and sexualreproduction", there was no student response on thistopic. Then Lucy showed another four different rep-resentations to help them participate in the discussion.About this event Lucy said,

I saw dull expressions on the faces of my students and I knew

I was in trouble. I guessed that they were too tired to keeptheir attention on my teaching after a whole morning of classes.I repeated the same explanation again. But they still did not

respond. I decided to draw a picture and gave an example tohelp the students understand. Then I asked them to pretendthat they were the Hydra family, to state their position, and toexplain why they hold this position. This evoked a favorable

response.

This was an example of Lucy' s instructional actionwhich resulted directly from her reflection duringteaching.

Lucy's colleagues also participated in the evalu-ation of her representations. She always introducedand discussed new representations with her colleagues.They would give her feedback, such as if they likedit or disliked it and the reasons why.

V. DiscussionThroughout the cycle of teaching and reflection

on representations, Lucy expanded her repertoire ofinstructional representations. This was a continuousprocess of learning to teach.

Coming from a family of teachers, Lucy lookedforward to being a teacher and considered being ateacher her number one choice. For her, teaching wasan important thing which would influence thousandsof students. She thought that teachers played a keyrole in learning and could make a different world forstudents. She had high expectation of herself. She

believed that a good teacher must be a learner of howto teach, and no man was born a successful teacher.Lucy had rich resources for learning to teach. Personallearning, teaching experience, textbooks, other teach-ers, volunteer training, in-service training, universityprofessors, library services, consultants at teachers'in-service training centers, museums, parks, and stu-dents' responses were all probable sources of herknowledge base. Her belief in learning how to teachand rich resources for learning how to teach helpedher knowledge develop continually.

For Lucy, an ideal representation was close tothe scientific facts, was comprehensible to students,made learning interesting, and was feasible in theteaching context. However, teaching circumstancewere too complex to achieve all criteria of "good"representations simultaneously. Reflection on thecriteria made her believe that every representation hadlimitations. It was like a double-edged sword, as ithelped student learning, it might also lead to misun-derstandings (Duit, 1991). The incompleteness ofrepresentations provided space for her to continuallyexpand her representations.

Lucy needed to transfer the subject matter intoforms accessible to students and to represent the contentas authentically as possible. This made the transfor-mation act a dilemma, as other researchers have de-scribed (Geddis et al., 1993). Developing and pre-senting multiple transformations was a strategy sheused to deal with the dilemma. Then she could com-pensate for the limitations of representations. In themeantime she could show the many facets of scienceand meet the needs and interests of different students.As Lucy dealt with dilemmas, she reflected on her ownteaching and promoted her own professional growth.This kind of growth seemed to be the same as that inTomanek's (1994) study.

Shulman (1987) indicated that transformationsrequire a combination of preparation, representation,selection, adaptation, and tailoring to student charac-teristics. Among these processes, representationseemed to be the most obscure and difficult processof forming new representations. We know little about"how to represent". The findings in this study showthat this "how" was modification, specification, com-bination, or invention in Lucy's case. The processproceeded from Lucy's extensive knowledge; herknowledge also developed during the renewing pro-cess. Lucy's expanding representations and her knowl-edge hold a dialectic relationship. This seems to bethe same as Clark and Peterson's (1986) description.

Besides knowledge, time, effort, resources, andsupport, creativity was also a key factor in Lucy's

839 4

S.W. Lin

renewal of her representations. Invention, as knowl-edge reorganization (Schon, 1983), showed the cre-ative side of the teacher' s thinking. It allowed Lucyto used only simple and inexpensive materials andeasily arrive at procedures to represent concepts forspecial purposes, students, and contexts. This renew-ing process not only solved Lucy' s teaching problems,but also helped her enjoy the self-actualized feelingthat accompanied the process of creation. These innerand outer motivations seemed to be the force drivingLucy to renew representations continually.

Facing the existing environment, Lucy knew howto balance her ideals with the expectations held by thesociety, school administrators, parents, and students;bow to utilize hardware and software of the schbol andcommunity to enrich her representations; and how toask support from colleagues, school administrators,and the community for biology teaching. Teachingcontext shaped Lucy's forms of representations but didnot prevent her from improving them. For Lucy thesecomponents did not seem to act as constraints in thesame manner as beginning teachers' construction (Abell& Roth, 1992; Brickhouse, 1993; Tobin & Gallagher,1987). They acted more like learning conditions whereLucy could test the feasibility of her new represen-tations. Sometimes, they seemed to act as a facilitatorthat pushed Lucy to use her creative capacity towardsolving teaching problems.

VI. ImplicationThis case study of an experienced biology teacher

illustrates that reflection on representations can be aneffective approach for professional growth. A teacher' sthinking is often a reflection of her representations.The incompleteness of representations opens spacesfor her to continually expand representations. Throughthe expanding process the teacher demonstrates thedynamic, flexible, content specific, context depen-dent, interchangeable, incomplete, and personal teach-'ing style of representations. This process is based onthe teacher' s extensive knowledge, and knowledge isdeveloped during the process. The results enable usto further understand the nature of a teacher' s lifelonglearning.

Perhaps this suggests that teacher educators,especially those involved in in-service education, canencourage teachers to take chatge of their own pro-fessional growth by helping them to build criteria of"good" representations for reflecting on their repre-sentations. Discussion with teachers about reflectionon criteria of "good" representations should lie illus-trated in terms of dilemmas. Helping teachers focus

on the dilemmas may empower them to reflect on theirteaching, identify dilemmas, locate resources, andmanage them within themselves, rather than initiatingprofessional growth from outside their classroom world.Science teacher preparation programs should fosterreflection that enables teachers to make sense ofclassroom events and apply their knowledge to formrepresentations which fit their circumstances. Scienceteacher educators could then provide alternative sug-gestions and support for those who are unsatisfied andwant to make change, and those who need help.

Further research is required to articulate similardevelopment for other teachers in the same and dif-ferent subject areas.

References

Abell, S. K., & Roth, M. (1992). Constraints to teaching elementaryscience: A case study of a science enthusiast student teacher.Science Education, 76(6), 581-595.

Berliner, D. C. (1988). The development of expertise in pedagogy.Paper presented at American Association of Colleges for TeacherEducation, Washington, D.C.

Brickhouse, N. W. (1990). Teacher's beliefs about the nature ofscience and their relationship to classroom practice. Journalof Teacher Education, 41(3), 53-62.

Brickhouse, N. W. (1993). What counts as successful instruction?An account of a teacher's self-assessment. Science Education,77(2), 115-129.

Britzman, D. P. (1986). Cultural myths in the making of a teacher:Biography and social structure in teacher education. HarvardEthtcational Review, 56(4), 442-456.

Carlsen, W. S. (1991). Effects of new biology teachers' subjet-matter knowledge on curricular planning. Science Eudction,75(6), 631-647.

Clark, C. M., & Peterson, P. L. (1986). Teacher's thought process.In: M.C. Wittrock(Ed.), Handbook of research on teaching (3rded.) (pp. 255-296). New York, NY: Macmillan.

Duffee, L., & Aikenhead, G. (1992). Curriculum change, studentevaluation, and techer practice knowledge. Science Education,76(5), 493-506.

Duit, R. (1991). On the role of analogies and metaphors in learningscience. Science Education, 75(6), 649-672.

Feiman-Nemser, S. (1983). Learning to teach. In: L.S. Shulman& G. Skyes (Eds.), Handbook of teaching and policy (pp. 150-170). New York, NY: Longman.

Feiman-Nemser, S., & Buchmann, M. (1989). Describing teachereducation: A framework and illustrative findings from a lon-gitudinal study of six students. The Elementary School Journal,89(3), 365-377.

Fuller, F. (1969). Concerns of teachers: A developmental charac-terization. American Educational Research Journal, 6, 207-226.

Geddis, A. N., Onslow, B., Beynon, C., & Oesch, J. (1993).Transforming content knowledge: Learning to teach aboutisotopes. Science Education, 77(6), 575-591.

Grossman, P. L. (1992). Why models matter: An alternate viewon professional growth in teaching. Review of EducationalResearch, 62(2), 171-79.

Hashweh, M. Z. (1987). Effect of subject matter knowledge in the

84

95


teaching of biology and physics. Teaching and Teacher Edu-cation, 3(2), 109-120.

Kagan, D. M. (1992). Professional growth among preserving andbeginning teachers. Review of Educational Research, 62(2),129-169.

Lantz, 0., & Kass, H. (1987). Chemistry teachers' functionalparadigms. Science Education, 71(1), .117-133.

Lederman, N. G., Gess-Newsome, J., & Lantz, M. S. (1994). Thenature and development of preservice science teachers' concep-tions of subject matter and pedagogy. Journal of Research inScience Teaching, 31(2), 129-146.

Lin, S. W. (1994). An interpretive study on instructional. repre-sentations of a junior high school biology teacher. UnpublishedDoctoral Dissertation. National Taiwan Normal University,Taipei, Taiwan.

Loughran, J. (1994). Bridging the gap: An analysis of the needsof second-year science teachers. Science Education, 78(4), 365-386.

Marks. R. (1990). Pedagogical content knowledge: From a math-ematical case to a modified conception. Journal of TeacherEducation, 41(3), 3-12.

McDiarmid, G. W., Ball, D. L., & Anderson, C. W. (1989).. Whystaying one chapter ahead doesn't really work: Subject-specificpedagogy. In: M.C. Reynolds (Ed.), Knowledge base for thebeginning teacher (pp. 193-205). Oxford: Pergamon Press.

Munby, H. (1984). A qualitative approach to the study of a teachers'beliefs. Journal of Research on Science Teaching, 21(1), 27-38.

Munby, H. (1986). Metaphor in the thinking of teachers. Journalof Curriculum Studies, 18(2), 197-209.

Schon, D. A. (1983). The reflective practitioner: How professionals

SIV---kritreg

think in action. New York, NY: Basic Books.Shulman; L. S. (1986). Those who understand: Knowledge growth

in teaching. Educational Researcher, 15(2), 4-14.Shulman, L. S. (1987). Knowledge and teaching: Foundation of

the new reform. Harvard Educational Review, 57(1), 1-22.Tobin, K. (1990a). Belief about teaching and learning. In: K. Tobin,

Kahie, & B.J. Fraser (Eds.), Windows into science class-room: Problems associated with higher-level cognition learn-ing (pp. 36-51). London: The Falmer.

Tobin, K. (1990b). Changing metaphors and beliefs: A masterswitch for teaching? Theory into Practice, 29(2), 122-127.

Tobin, K., Briscoe, C., & Holman, J. R. (1990). Overcomingconstraints to effective elementary science teaching. Science

Education, 74(4), 409-420.Tobin, K., & Gallagher, J. J. (1987). What happens in high school

science classrooms? Journal of Curriculum Studies, 19(6), 549-560.

Tomanek, D. (1994). A case of dilemmas: Exploring my assump-tions about teaching science. Science Education, 78(5), 399-414.

Wallace, J., & Louden, W. (1992). Science teaching and teachers'knowledge: Prospects for reform of elementary classrooms:Science Education, 76(5), 507-521.

Wilson, S. M., Shulman, L. S., & Richert, A. E. (1987). "150different ways" of knowing: Representations of knowledge inteaching. In: J. Calderhead (Ed.), Exploring teachers' thinking(pp. 104-124). London: Cassell.

Wood, T. (1988). State-mandated accountability as a constraint onteaching and learning science. Journal of Research in ScienceTeaching, 25(8), 631-641.

4414-.

VOIS11,100z

AN*NR4AROOMIA-fft

N

21KfMAIMANV-7111*MIAI=PIWAN,Aft*N414,0M4MMARAft*Vik,U4*I.*)NIV041.iffA*MMAttfte#VOWitl'3c411,01MM'aliT-UU*5)-*;AUrElfMEJOAMXIIMMOMVO-OPPIA*Vit,OAORft*V4;01Aft0-MIVORMAIA&Tt;,Ug4I01 A;(3)16MVOMXMM*g,UJIARV,M;(4) finft*,

MAN$A0Wft-0.1'44nTh'EAMMMMA;(5)Ma0A*AMOffigAMMRIXWAY4WIT.diY:t*ATIAMT'AiMit,IAMV4MMRA,ft0ffikitMRME04,ffiaWV*AMIMISit,F1044f-XIAM,404±A"044-1.#*ttOWJAMIVA01404,MOAPis*N'IittMnftfM.21KaA0a71--,A,4*ftWiqUEhaqift*MASt,VAWTMAZ4W-4-14t%1MAR,M444*N*Wft MAMMffAt.-

85

96


Expert Opinions Concerning a Taxonomic Structurefor the Curricular Organization of Biotechnology

JERMING TSENG

Department of BiologyNational Taiwan Normal University


(Received May 13, 1996; Accepted April 10, 1997)

ABSTRACT

Biotechnology has been declared a "strategic industry" and made a national priority by thegovernment in Taiwan. In this study, the opinions of biotechnology experts were investigated and ataxonomic structure describing the range and content of the field of biotechnology proposed. A modifiedDelphi technique was followed in this study. The study included a literature survey, design of aquestionnaire, first distribution of the questionnaire, preliminary modification of questionnaire content,distribution of the modified version, face-to face expert panel discussions, and finally completion ofthe final version of the taxonomy. The expert panel included 11 experts at the first stage (questionnaire)and 16 experts at the second stage (expert panel discussion). The conclusions are as follows: (1)Biotechnology requires a wide range of knowledge from at least 16 disciplines, including microbiology,chemistry, and molecular biology, etc. (2) Biotechnology should cover 15 main concepts includingits definition, historical development, basic knowledge, basic techniques, manipulation of prokaryoticcell genes, manipulation of eukaryotic cell genes, protein engineering, application of microbes to producecommercial products, diagnostic reagents, vaccines and therapeutic drugs, microbial degradation,downstream processes and social aspects of biotechnology. The taxonomic structure of biotechnologyderived through expert consensus should allow for further development of curriculum for biotechnologyprograms.

Key Words: taxonomy, biotechnology, Delphi technique

I. Introduction

The term "Biotechnology" was proposed by KarlEreky in 1919 (Bud, 1993). Since then, biotechnologyhas become one of the major disciplines linking basicresearch with industrial production. It is widelyapplied to areas including bioprocessing (fermenta-tion), genetic manipulation, agriculture, medicine, en-vironmental protection and others. The definition ofbiotechnology suggested by the Organization forEconomic Co-operation and Development (OECD) in1993 was: "Biotechnology is the application, of sci-entific and engineering principles to the processingof materials by biological agents to provide goods andservices" (Bud, 1993).

Greater awareness of the profound effects ofbiotechnology on human life has resulted in a signifi-cant increase in investment in biotechnology researchand development. In order to meet the great demand

for manpower in this fast-growing industry, develop-ment of a biotechnology curriculum has been recom-mended by experts (Hudson, 1988). However, basedon the specific aim of the particular biotechnologicalprogram, the content of the curriculum could be verydiverse.

A study conducted by Rinard (1986) collecteddata from 242 biotechnology companies. Based onthe task listings and suggestions from survey respon-dents, he developed a model curriculum for two-yearbiotechnician training program. His model curriculumsought to provide sufficient background so that,with a reasonable amount of work experience, thegraduates could advance to positions with increasingresponsibilities. Therefore, his curriculum includedbasic knowledge (e.g. chemistry, molecular and cellbiology, microbiology), applied science (e.g. funda-mentals of instrumentation and control, industrial mi-crobiology) and communication skills (e.g. technical

86

Taxonomy of Biotechnology

communications).Gayford (1987) suggested an in-service training

program for teachers. The specific aim of the programwas to produce a course covering aspects of biotech-nology which would increase understanding of variousareas of the subject, as well as encouraging the inte-gration of biotechnology into science education for thewhole age and ability range of 13-18 year-old pupils.Therefore, the content of the curriculum was distinctfrom that developed by Rinard. Topics included anintroduction to biotechnology, safety and managementaspects of biotechnology, genetic manipulation, fer-mentation, tissue culture and cell culture, enzymology,downstream processes, and enhancement of the effectsof biotechnology.

Paolella (1991) suggested biotechnologyoutlines for both high school and college classes. Thecurriculum was designed for students who hadcompleted a molecular genetics unit which includedthe structure of nucleic acids, replication, transcrip-tion and translation. The content of the curriculumincluded definition of biotechnology, genetic engi-neering, history of biotechnology, and basic techni-ques of biotechnology (e.g. isolation of DNA, utiliza-tion of restriction enzymes, choice of a vector inwhich to insert a gene of interest). The authoralso included a concept and V map in a modelsection.

Another curriculum was suggested by Glick &Pasternak (1994), this time focussing on advancedbiotechnology. Based on a biotechnology courseoffered for over 12 years to advanced undergraduateand graduate students at the University of Waterlooin Canada, Glick & Pasternak (1994) wrote a bookentitled Molecular biotechnology: principles andapplications of recombinant DNA. The book wasdesigned to serve as a text for courses in biotechnol-ogy, recombinant DNA technology and genetic engi-neering, or for any course introducing both theprinciples and the applications of contemporary mo-lecular biology. The authors suggested that studentswho took this course should have an understanding ofbasic ideas from biochemistry, molecular genetics andmicrobiology. The curriculum included Fundamen-tals of molecular biotechnology (e.g. recombinant DNAtechnology, manipulation of gene expression inprokaryotes, directed mutagenesis and protein engi-neering), Microbial systems (e.g. microbial synthesisof commercial products, vaccines and therapeuticagents, large-scale production of proteins from recom-binant microorganisms), Eukaryotic systems (e.g.genetic engineering of plants, development and use oftransgenic animals, isolation of human genes) and

Regulating and patenting molecular biotechnology (e.g.regulating the use of biotechnology, patenting bio-technology inventions).

The purpose of this study was to develop aconceptual model of biotechnology's taxonomic struc-ture. Such conceptual model, derived through expertconsensus, would allow for future development ofbiotechnology curricula and would be applicable toboth technology education programs and biologicalscience programs. To achieve this goal, the followingresearch questions were addressed to a panel ofexperts:

(1) What are the main knowledge areas of biotech-nology?

(2) What are the proper subdivision titles for eachknowledge area?

(3) What are the major concepts and subconceptsunder each subdivision title?

H. MethodsThe research method used in this study was a

modified Delphi technique (Cyphert & Gant, 1971).The purpose of using the Delphi technique was toobtain a consensus from a panel of experts within thediscipline of biotechnology. The expert panel mem-bers were selected from seven research institutes.Among them were microbiologists, molecular biolo-gists, virologists, plant physiologists, immunologistsand downstream process experts.

The Delphi technique was developed during the1950s by workers at the RAND Corporation as aprocedure to "obtain the most reliable consensus ofopinion of a group of experts by a series of intensivequestionnaires interspersed with controlled opinionfeedback." The technique was originally applied todeal with complex defense problems (Uhl, 1990). Inlater years it has come to be used in a wide varietyof fields, including business, government, industry,medicine, regional planning and education, for awide variety of purposes, including future forecasting,goal assessment, and curriculum planning, etc. (Uhl,1990).

Four necessary features characterize the Delphiprocedure: anonymity, iteration, controlled feed-back, and statistical aggregation of group response. Inthis study, anonymity was achieved using question-naires and small group panel discussions. Interactionwas proceeded by two rounds of questionnaires.Controlled feedback was performed between the twoquestionnaires and the statistical analysis was donesimultaneously. Each member of the expert panel wasinformed of theresults of the statistical analysis and

87

9 3

J. Tseng

the comments of other experts. After the panel dis-cussion, a consensus from the panel of experts wasachieved. However, the greatest difficulty of usingthe Delphi technique in this situation was that thedisciplines within biotechnology are so diverse thatno one expert could be familiar with all the conceptson the questionnaire. Therefore, one or more partici-pants', views predominated with regard to severalcoricepts.

1. The First Questionnaire

After an extensive literature review, a prelimi-nary understanding of the main knowledge areas,subdivisions and major concepts was established. Thefirst questionnaire was subsequently developed on thebasis of this preliminary understanding. The question-naire was divided into two parts: The first part askedthe expert panel members for their responses andcomments in thirteen knowledge areas possibly relat-ing to the field of biotechnology. The second partasked the expert panel members for their judgmentsand written comments pertaining to the sixteen sub-divisions and concepts relating to those subdivisions.The experts were asked to rate each knowledge area,subdivision title and major concept on a five-pointscale as to their degree of agreement. In addition, theywere asked to give a brief written comment at the endof each subdivision.

The results of the first questionnaire were sta-tistically analyzed and the subjects that received anegative score were adjusted or deleted on the basisof written comments. The revised questionnaire wasresubmitted to the same expert panel. The experts

were asked again to make judgments on the subjectsbefore the panel discussion was held.

2. Panel Discussions

Panel discussions were held separately at seveninstitutes and the total number of participants wasexpanded to fifteen. The content and structure of theknowledge areas and taxonomy were further revisedbased on the results of the panel discussions. If apanelist with a response outside the consensus rangefailed to change his/her suggestion, he/she was askedto provide an additional statement. A final versionof the knowledge areas and taxonomy was then sub-mitted to the participants in order to confirm theconsensus.

Ill. Results1. Knowledge Areas of Biotechnology

Based on the degree of consensus, the expertswere asked to rate each of the knowledge areas on afive-point scale. The frequency of each level agree-ment and total points were recorded (Table 1). Thethree knowledge areas Industrial engineering, Elec-tronic engineering and Mechanical engineeringobtained a score below 0. Therefore these threeknowledge areas . were excluded from the secondDelphi questionnaire. The experts also suggestedthat Zoology, Plant biology, Physiology, Immunol-ogy, Animal breeding and Medical engineeringshould be included as knowledge areas. These sug-gestions were also included in the second Delphi

Table I. Degree of Consensus on the Knowledge Areas of Biotechnology

Knowledge Area Strongly Agree Agree Marginal Disagree Strongly Disagree Score

Microbiology 10' 0 21bChemistry 5 5 0 15Cell biology 10 0 21Molecular biology 11 0 0 22Food science 4 5 2 0 13Chemical engineering 3 6 2 0 12Industrial engineering 5 4 -1Electronic engineering 2 4 3 1 -1Genetics 8 3 0 19Food processing 3 7 1 13Biochemical engineering 5 5 1 IsBiochemistry 9 2 1 20Mechanical engineering 2 4 3 -1

°Indicates the number of specialists who selected the category.bScore: 2, strongly agree; 1, agree; 0, marginal; -1, disagree; -2, strongly disagree.

88

ut_


questionnaire.After the results of the second questionnaire and

panel discussion were sorted and analyzed, all the sixnewly added knowledge areas obtained scores above0. Only two out of fifteen experts disagreed withincluding Zoology, Plant biology and Physiology onthe list of biotechnology knowledge areas. The finalversion which reflects the consensus of the expertpanel is shown in Table 2.

2. Subdivision Titles and Major Concepts

The preliminary understanding of the conceptualframework was derived from Molecular Biotechnol-ogy by Glick & Pasternak (1994) and MolecularBiotechnology by Primrose (1991). The sixteen sub-divisions of Nature of biotechnology, History ofbiotechnology, Basic knowledge of molecular biotech-nology, Major techniques of molecular biotechnology,Manipulation of gene expression in prokaryotes,Manipulation of eukaryotic cells, Protein engineering,Microbial synthesis of commercial products, Molecu-lar diagnostics, Vaccines and therapeutic agents,Bioremediation, Large-scale production of proteinsfrom microorganisms, Transgenic animals, Genetherapy, Plant biotechnology and Social aspects ofbiotechnology were listed on the first questionnaire.All sixteen subdivisions obtained a score above zero(Table 3). However, during the panel discussion, 10out of 15 experts .suggested integrating the majorconcepts of Plant biotechnology into other subdivi-sions, such as Manipulation of eukaryotic cells.Therefore, Plant biotechnology was excluded from thefinal version (Appendix).

The major concepts under each subdivision allobtained a positive score. However, several conceptsdid obtain a relatively low score (below 10). In addition,the following issues were subjected to intensive debate

Table 2. The Knowledge Areas of Biotechnology-Order of Signifi-cance

Order Area Order Area

1 Molecular biology 9 Physiology2 Microbiology 10 Chemistry3 Cell biology 11 Biochemical engineering4 Biochemistry 12 Food science5 Genetics 13 Chemical engineering6 Immunology 14 Food processing7 Zoology 15 Animal breeding8 Plant biology 16 Medical engineering

Note: The order for individual areas was based on their scores fromthe questionnaire.

during the panel discussion:(1) Should biotechnology be divided into classical

biotechnology and modern biotechnology (ormolecular biotechnology)? Eight out of fifteenexperts agreed that there was a distinctionbetween classical and modern biotechnology.

(2) Should "site-directed mutagenesis", "expres-sion of recombinant genes", "polymerase chainreaction" and "transgenic animals" be includedin Basic knowledge of molecular biotechnol-ogy? Thirteen out of fifteen experts finallyagreed to consider these four concepts as "basicknowledge," instead of "basic techniques" ofbiotechnology.

(3) Nine out of fifteen experts regarded "majortechniques of protein engineering" as Majortechniques of molecUlar biotechnology. Inaddition, it was thought that "identification ofrecombinant genes" (i.e. Southern blotting,Northern blotting and Western blotting) shouldbe included among basic techniques.

(4) Ten out of fifteen experts agreed that "integra-tion of foreign DNA" need not be a major conceptunder the Manipulation of gene expression inprokaryotes subdivision.

(5) Eleven out of fifteen experts suggested mergingSaccharomyces and Pichia pastoris system inthe concept of "yeast expression system". Inaddition, it was felt that the concept of "plantcell expression system" should be added toManipulation of gene expression in eukaryoticcells since the "plant biotechnology" subdivi-sion had been deleted.

(6) Under the Microbial synthesis of commercialproducts subdivision, thirteen out of fifteenexperts suggested merging "restriction endonu-clease" in the major concept "protein pharma-ceuticals".

(7) Both "small biological molecules" and "biopoly-mers" obtained relatively low scores (4 and 5,respectively) since five out of fifteen expertsconsidered these concepts were not specificenough to make a judgment.

(8) The term "molecular diagnostics" could notcover all newly developed diagnostic systems.Therefore, the title of the subdivision was re-placed by Diagnostic products.

(9) In light of written comments and conclusionsof panel discussions, Large-scale productionof proteins from microorganisms was replacedby The major techniques of industrial micro-biology. The major concepts under this sub-division consisted of "downstream process-

89

loo4,e Ju

J. Tseng

Table 3. Degree of Consensus on the Subdivision and Major Concepts of Biotechnology

Subdivision StronglyConcepts Agree

Agree Marginal Disagree StronglyDisagree

Score

Nature of Biotechnology 0a 8 3 0 0 8b

Definition 2 8 1 0 0 12

Features of biotechnology 3 6 2 0 0 12

Social aspects of biotechnology 2 8 1 0 0 12History of Biotechnology 1 7 2 1 0 8

Early development (B.C. 900 to A.D. 1800) 1 8 2 0 0 10

Zymotechnology (A.D. 1800 to 1900) 2 6 3 0 0 10Organic compound production 2 7 1 1 0 10Antibiotics production 2 7 1 1 0 10Moleculai biotechnology 4 6 1 0 0 14Future development 2 8 0 1 0 11

Basic Knowledge of Molecular Biotechnology 3 8 0 0 0 14Impacts of molecular biology 1 10 0 0 0 12Basic knowledge 1 10 0 0 0 12

Major Techniques of Molecular Biotechnology 2 8 0 1 0 11

Synthesis of oligonucleotides 2 6 3 0 0 10DNA sequencing 2 9 0 0 0 13Polymerase chain reaction 1 8 0 2 0 8

Cell culture technique 3 8 0 0 0 14Manipulation of Gene Expressionin Prokaryotes

3 7 1 0 0 13

Prokaryotic cell gene structure andexpression

3 7 1 0 0 13

Isolation of a functional promoter 1 8 2 0 0 10Fusion proteins 2 7 2 0 0 11

Copy number 2 7 2 0 0 11

Protein synthesis in prokaryotic cells 2 8 1 0 0 12Integration of foreign DNA 2 8 1 0 0 12Pathways of protein secretion in prokaryoticcells

2 8 1 0 0 12

Manipulation of Eukaryotic Cells 2 8 1 0 0 12Saccharomyces expression system 2 7 2 0 0 11

Pichia pastoris expression system 1 9 1 0 0 11Insect cell expression system 1 6 3 1 0 7Mammalial cell expression system 2 7 2 0 0 11

Protein Engineering 2 9 0 0 0 13The purpose of protein engineering 2 9 0 0 0 13Site-directed mutagenesis 1 10 0 0 0 12Random mutagenesis 0 11 0 0 0 11

Major techniques of protein engineering 2 8 0 1 0 11

Microbial Synthesis of Commercial Products 2 7 0 2 0 9Protein pharmaceuticals 1 9 1 0 0 11

Restriction endonuclease 2 7 1 1 0 10Small biological molecules 1 4 4 2 0 4Biopolymers 1 4 5 1 0 5

Molecular Diagnostics 3 7 0 1 0 12Immunological diagnostic products 2 7 1 1 0 10DNA diagnostic system 2 7 2 0 0 11

Vaccines and Therapeutic Agents 0 .11 0 0 0 11

History of vaccines 2 7 2 0 0 11

Subunit vaccines 1 6 3 1 1 5Attenuated vaccines 1 5 4 1 1 4Monoclonal antibodies as therapeutic agents 1 5 4 1 1 4

(to be continued)

90

1 0 1


Table 3. (continued)

Subdivision Strongly Agree Marginal Disagree Strongly ScoreConcepts Agree Disagree

Biodegradation 3 6 1 0 11

Biodegradation of toxic waste 1 6 2 2 0 6Genetic engineering of biodegradativepathways

6 4 0 0 8

Utilization of starch 8 3 10Utilization of cellulose 8 2 0 10

Large-scale Production of Proteins from 9 0 1 I

MicroorganismsCharacteristics of microbial growth 8 1 0 9Large-scale cultivation 8 2 0 0 10Downstream processing 9 0 0 11

Immobilization 2 8 0 0 12

Transgenic Animals 9 1 0 10Transgenic mice: methodology 9 1 0 10Transgenic mice: applications 8 2 0 0 10Transgenic strains of livestock 8 2 o 8

Transgenic birds 10 o o 12Transgenic fish 8 2 0 0 10

Gene Therapy 2 7 1 0 10Target diseases for human gene therapy 1 7 3 0 0 9Gene mapping 1 6 4 0 0 8

Techniques of somatic cell gene therapy 1 7 3 0 0 9Plant Biotechnology 2 6 3 0 0 10

Breeding by recombinant DNA techniques 1 7 3 0 0 9Application of plant biotechnology 1 8 2 0 0 10

Social Aspects of Biotechnology 3 8 0 0 14Patent applications 1 7 2 0 7

Regulating the usage of biotechnology 2 7 2 0 0 11

DNA fingerprints 9 1 0 0 11

Social aspects of gene therapy 2 9 0 0 0 13

Prospects of biotechnology education 1 8 2 0 0 10

°Indicates the number of specialists who selected the categorybScore: 2, strongly agree; 1, agree; 0, marginal; -1, disagree; -2, strongly disagree.

ing", "characteristics of microbial growth","large-scale cultivation", "product purificationand waste disposal" and "immobilization tech-niques".

(10) Major changes were made to the content of thesubdivision Social aspects of biotechnologyafter the panel discussion. Thirteen out offifteen experts suggested that "DNA fingerprint","regulating the usage of biotechnology" and"social aspects of gene therapy" should bemerged in the single concept "the impact ofbiotechnology on the society". In addition, itwas suggested that the major concept "restric-tions on biotechnology research and develop-ment", which covered regulations and guidelineconcerning R&D, should be added under thissubdivision.

The final version of a taxonomic structure for thefield of biotechnology is shown in Appendix.

;

IV. Discussion

The potential of biotechnology has been realizedby most educators in our country. It is predicted thatindividuals who will be or are in biotechnology in-dustries will be requested to have specific educationand training in biotechnology. .

This study identified the basic knowledge areasof biotechnology and classified the content of biotech-nology. Table 2 presents the final version of theknowledge areas. Individuals in the field of technol-ogy education or biological science education mayhave different viewpoints regarding how to use thislist for curriculum development. Those who are de-veloping a curriculum for technology education mayemphasize chemical engineering, food processing,biochemical engineering, animal breeding and medi-cal engineering. For those who are developing acurriculum for biological science education, microbi-

102

J. Tseng

ology, chemistry, cell biology, molecular biology,zoology, plant biology, physiology, genetics, biochem-istry and immunology would be major disciplines ofinterest. Glick & Pasternak (1994), for example, intheir book "Molecular Biotechnology: Principles andApplications of Recombinant DNA" suggested that manyscientific disciplines contribute to molecular biotech-nology, including molecular biology, microbiology,biochemistry, genetics, chemical engineering and cellbiology. Those disciplines emphasize theory, discov-ery and experimentation. In this study, no technologyeducation or biotechnology industry experts wereinvited to join the expert panel. Therefore, the out-come inevitably emphasized basic knowledge andskills needed for research and development. Furtherinvestigation will be necessary to define the corecompetence for students who seek a career in thebiotechnology industry, and the emphasis should beon designing, creating, applying and ultimately lead-ing to a commercial product.

A complete taxonomic structure for biotechnol-ogy is listed in Appendix. A completed taxonomyshould be very useful in developing a blueprint forcurriculum development. Although the contents ofthe list ran into the same problem as the "knowledgeareas," which heavily, emphasized biological scienceeducation, an effort was made to include severalsubdivisions that were closely related to technologicaleducation, such as "Major techniques of industrialmicrobiology" and "Microbial synthesis of commer-cial products". De Landsheere (1991) addressed thefact that the content validity of a taxonomy will notbe considered perfect by any author, but it allowsnearly all the cognitive objectives of education to beclassified. In addition, the subcategorization used ina taxonomy is not always based on the same classi-fication principle. It should be based on whether thetaxonomy is to be used for comprehension or evalu-ation. The taxonomy presented here was not intendedto make a judgment about the value of any specifictheory, idea, process or methodology. Instead, it wasa comprehensive taxonomy. Therefore, the additionof several subdivisions relating to technological edu-cation did not go against our principles.

When developing a particular curriculum, thecontent of the taxonomy often only presents a theme,and there is no indication given about what instruc-tional objectives should be developed under the theme.Based on the taxonomic structure, instructional objec-tives developed for teaching biotechnology courses intechnology education and biological science programswould differ in approach, yet remain centered arounda consistent content (Wells, 1994). Therefore, an ideal

model curriculum should be tailored to meet the needsof various applications, and wodld require carefulscrutiny by both industry and academic institutions(Rinard, 1986).

A curriculum should be integrated instead ofsimply adding new materials into a preexisting courseas an appendage. Therefore, a conceptual frameworkfor biotechnology will be essential for developing thecurriculum (Hudson, 1988). DeVore (1966) indicatedthat a complete taxonomic structure would eliminateconfusion and simplify the task of curriculum planningby providing a perspective of the relationships be-tween the elements and the structure. It would orderknowledge areas into specific categories, and identifythe difficulty level of content areas in order to estab-lish instructional sequences at different learning levels.The taxonomic structure shown in Appendix suggesteda hierarchical relationship between a basic knowledgelevel and a higher technical level. Therefore, it allowsthe principles of individual techniques to be groupedinto the subdivision "Basic knowledge of molecularbiotechnology". The principle of creating a transgenicanimal, for example, was classified into the third levelof "Basic knowledge of molecular biotechnology" in-stead of being classified into the subdivision"Transgenic animals". The technical aspects oftransgenic animals, on the other hand, remained in thesubdivision "Transgenic animal". In other words,the order of learning biotechnology suggested by thistaxonomic structure was to understand the basic prin-ciple of individual techniques first and then to intro-duce specific applications in individual fields.

In conclusion, an ideal curriculum for biotech-nology programs should meet the objectives of theeducational program and cover both principles andapplications in a logical sequence. In addition, thestudents entering the program should have backgroundknowledge covering disciplines including molecularbiology, cell biology, genetics, biochemistry andmicrobiology.

Acknowledgment

I thank all the participants in this project. The help of Dr.Y. J. Cheng in analyzing data from the questionnaires is gratefullyacknowledged. This work was supported under NSC project NSC84-2511-S-003-046TG.

References

Bud, R. (1993). The use of life: A history of biotechnology. NewYork, NY: Cambridge University press.

Cyphert, F. R., & Gant, W. L. (1971). The Delphi technique: Acase study. Phi Delta Kappan, 52, 272-273.

92

103


De Landsheere, V. (1991). Taxonomies of educational objectives.In: A. Lewy (Ed.), The international encyclopedia of curriculum(pp. 317-327). New York, NY: Pergamon Press.

DeVore, P. W. (1966). Structure and content foundations forcurriculum development. Washington, D.C.: U.S. Departmentof Health, Education & Welfare, office of Education.

Gayford, C. (1987). Biotechnology 13-18: In-service training forteachers. Journal of Biological Education, 21, 281-287.

Glick, B. R., & Pasternak, J. J. (1994). Molecular biotechnology:principles and applications of recombinant DNA. Washington,D.C.: ASM Press.

Hudson, R. A. (1988). Biotechnology at the university of Toledo:Development and implementation of an integrated curriculum.American Journal of Pharmaceutical Education, 52, 355-357.

Paolella, M. J. (1991). Biotechnology outlines for classroom use.American Biology Teacher, 53, 98-101.

Primrose, S. B. (1991). Molecular biotechnology. London: BlackwellScientific Publications.

Rinard, B. F. (1986). Education for biotechnology. Waco, TX:Center for Occupational Research and Development, Inc.

Uhl, N. P. (1990). Delphi technique. In: H.J. Walberg & G.D.Haertel (Eds.), The international encyclopedia of educationalevaluation (pp. 81-82). New York, NY: Pergamon Press.

Wells, J. G. (1994). Establishing a taxonomic structure for the studyof biotechnology in secondary school technology education.Journal of Technology Education, 6, 58-75.

Appendix

Taxonomic Structure of Biotechnology

1. Nature of biotechnology1.1 Definition of biotechnology1.2 Nature of biotechnology1.3 Social aspects of biotechnology

2. History of biotechnology2.1 Prehistoric development (B.C.9000-A.D.I800)2.2 Zymotechnology (A.D.1800-A.D.1900)2.3 Industrial microbiology (A.D.1900-A.D.1960)

2.3.1 Production of organic compounds2.3.2 Production of antibiotics2.3.3 Production of special chemicals

2.4 Molecular biotechnology (A.D.1960 to present)2.4.1 Development of molecular biology2.4.2 Genetic engineering2.4.3 Transgenic animals2.4.4 Gene therapy

3. Basic knowledge of molecular biotechnology3.1 Impacts of molecular biology3.2 Basic knowledge

3.2.1 DNA3.2.2 Restriction endonuclease3.2.3 Plasmids3.2.4 Gene library3.2.5 Expression of eukaryotic genes3.2.6 Vectors3.2.7 Transformation of prokaryotic cells3.2.8 Site-directed mutagenesis3.2.9 Expression of recombinant genes3.2.10 Polymerase chain reaction3 .2.11. Transgenic animals

0

4. Major techniques of molecular biotechnology4.1 Synthesis of oligonucleotides4.2 DNA sequencing4.3 Identification of specific genes

4.3.1 Southern blotting4.3.2 Northern blotting4.3.3 Western blotting

4.4 Cell culture techniques4.4.1 Microbial cell culture4.4.2 Animal cell culture4.4.3 Plant cell culture

4.5 Major techniques of protein engineering

5. Manipulation of gene expression in prokaryotes5.1 Structure and expression of prokaryotic cell genes5.2 Promoters

5.2.1 Isolation of a functional promoter5.2.2 Criteria for a good promoter

5.3 Fusion proteins5.4 Control of the gene copy number5.5 Protein synthesis in prokaryotic cells5.6 Control of protein secretion from prokaryotic cells

6. Manipulation of gene expression in eukaryotes6.1 Yeast gene expression system

. 6.1.1 Saccharomyces cerevisiae6.1.2 Pichia pastoris

6.2 Insect cell gene expression system6.2.1 Baculoviruses

6.3 Mammalial cell gene expression system6.3.1 Vectors for mammalian cells6.3.2 Applications

6.4 Plant cell gene expression system

7. Protein engineering7.1 The specific aims of protein engineering7.2 Site-directed (oligonucleotide-directed) mutagenesis

7.2.1 M13 system7.2.2 PCR-amplified system

7.3 Random mutagenesis

8. Microbial synthesis of commercial products8.1 Production of protein pharmaceuticals using recombinant

DNA techniques8.2 Production of small biological molecules

8.2.1 Antibiotics8.2.2 Vitamins8.2.3 Starch-derived products

8.3 Production of biopolymers8.3.1 Principle and advantage of using microorganisms to

synthesize biopolymers8.3.2 Polymers produced by microorganisms

9. Diagnostic products9.1 Immunological diagnostic systems

9.1.1 Applications of monoclonal antibodies9.1.2 Antigen diagnostic systems9.1.3 Major techniques in immunological diagnostic proce-

dures.9.2 DNA diagnostic system

9.2.1 DNA hyhridization techniques9.2.2 Non-radioactive DNA diagnostic procedures9.2.3 Application of DNA diagnostic systems

93

10 4

J. Tseng

10.Vaccine5 and therapeutic antibodies10.1 History of vaccines10.2 Attenuated vaccines10.3 Subunits vaccines10.4 Monoclonal antibodies as therapeutic agents10.5 Application of genetic engineering to the production of

vaccines and antibodies

11.Bioremediation11.1 Biodegradation of organic wastes11.2 Manipulation of degradative pathways using recombinant

DNA techniques11.3 Biodegradation of cellulose

12.Major techniques of industrial microbiology12.1 Downstream processes12.2 Characterization of microbial growth

12.2.1 Factors that regulate microbial growth12.2.2 Measurement of microbial growth12.2.3 The principle of continuous culture

12.3 Large-scale cultivation12.3.1 Parameters of large-scale culture12.3.2 Basic design of a bioreactor12.3.3 Control system of a bioreactor

12.4 Purification and waste disposal12.5 Immobilization

12.5.1 Advantages of immobilization12.5.2 Techniques of immobilization12.5.3 Criteria for the choosing types of immobilization

13. Transgenic animals13.1 Methodology for generating transgenic mice

13.1.1 Procedures for generating transgenic mice13.1.2 Characterization of transgenic mice

13.2 Applications of transgenic mice13.2.1 Gene repair13.2.2 Gene knockout

13.3 Transgenic cattle13.4 Transgenic pigs13.5 Transgenic birds13.6 Transgenic fish

14. Gene therapy14.1 Genetic disorders14.2 Establishment of the human genome map

14.2.1 Genetic linkage14.2.2 Restriction fragment length polymorphism (RFLP)14.2.3 Positional cloning

14.3 Somatic cell gene therapy14.3.1 Ex vivo gene therapy14.3.2 In vivo gene therapy14.3.3 Antisense RNA gene therapy

15. Social aspects of biotechnology15.1 Patent applications15.2 Impact of biotechnology on society

15.2.1 Regulating the usage of biotechnology15.2.2 DNA fingerprints15.2.3 Social Impact of gene therapy

15.3 Restrictions on biotechnology research and development15.4 Prospects of biotechnology education

94

U 105


tt 450-i *4* '4" 341- f4 Vif ffi

MAfOMNAJQ*lt*

tr4 N

ttR*4*ARIAWARAOAtINA*MZ,*ff91011Untt4tti-44*ZUE1W994,ARRffl V9?-4-3)-ORM+X o r tVlq (Delphi technique) , r301Nii-1_, r -±-141

SJ rig's.4*J r 1-,11:10.1RMJ r4*MAJ -

1-VAJ 4VrT AtfoRM4*ntglighl-N4±. 34444*JAiIIARAPA ggolsim*,4*A- 11 ,

ghWANdl'AZ4* ilffY-EZ.21KA.A 16 L M;i1.34 ; (1) tttfitMf4*MN 16 1440-71 'gMttt-g-- ft* gigtfR4-' 3)--ttfif- t-Af=1-* 4,E*IN* Mt* J S t IP* 214* -ft-PnWpgifil* titifff tfok* tA* ; (2) ttlAg4f4V-INiAlfgRIV44-4-114J*

It4014f31-*ANA 3)-tfoRiigf=1-VAMAKOilt 33--TtfoRM-11-14-'04JIRA ratlfiA,N M4111ZA19, P:PAIlfg043N,* XOWIURV atM-AAIJ artgiNcifiRYNtl ttitftitaffifi 'Matt&rViliZtatt* *ItIOMOItt MIllifi*SctfoRW4*-±tr#24 15 ViI4IXtSzAtitflIJUIt3-MAlibiOMMit 21-MV+XMLI-1- tklItM14f4f4A11-ZiAXS tfottM4*-3V-atilf ttfi15MIM45tTiVA VAttfigtMfraitiiZOitA"fifit

-95 106:k,


A Study of the Concept "Living Organism" and LivingOrganism Classification in Aboriginal Children

SHIH-HIJEI CHEN AND CHIH-HSIUNG KU

Department of Mathematics and Science EducationNational Hualien Teachers College

Hualien, Taiwan, R.O.C.

(Received August 16, 1996; Accepted April 10, 1997)

ABSTRACT

This study investigated the understanding and application of the concept "living organism" inaboriginal children. Thirty-six 2nd, 4th,and 6th graders were selected from an aboriginal elementaryschool in Hualien. A structured, clinical interview with two different approaches was designed to assessthe subjects' ability' to classify familiar and unfamiliar objects according to life status. The resultsindicated four forms of classification: (1) living; (2) non-living; (3) animal; (4) botanical. Children inall grades used at least two of them. Movement and growth were the most commonly given attributesof life and "living organisms". Young children relied relatively heavily on attributes true for animalsbut not for plants, whereas older children did the reverse. It was also found that children's conceptsof "living organisms" and "living things" were different. They used individual experience more to attributelife and less to classify as "living organisms". "Human" was seen as living but not considered a "livingorganism-. The biological subsets of plants (eg. trees, grasses, flowers) and animals(eg. birds, fish,tigers) were classified as comparable to the set "plant" and "animal", not as subsets of them. Sophisticationof conceptual development was not found to improve with age. Finally, possible sources of misconceptionand the relationship between life attribution, cognitive development and linguistic factors were discussed.

Key Words: aborigines, living organism, life concepts, classification

Introduction

1. Research Background

Numerous research on the concept of life hasbeen carried out in this country and worldwide. Thishas mostly centered on animism or children's viewsof life attributes (Russell & Dennis, 1939; Russell,1940; Huang & Lee, 1945; 0-saki & Samiroden, 1990;Su, 1968; Wang, 1993). Though there has been someresearch done on the subject of the class concept of"living organism" (Ryman, I 974a, 1974b), in Englishthis term is closed related to the.terms "living", "alive"or "life" which makes the foci's of this research therealm of "life". The Chinese term for "living organ-ism" (Sheng-Wu) does not have such close associationwith the term for "living" (Hwo). "Living organism"(Sheng-Wu) is a little closer to "life" (Ming or Shen-Ming) (which like the term for "living"(Hwo) is incolloquial use), but is a biological term used mainlyin the classroom. Do aboriginal children also differ-

entiate these terms? This is what we wanted to findout. Moreover, although the aboriginal children re-ceive a standard Chinese education, they live mainlyin a natural environment with wildlife as one of theirfood sources. It is conceivable, therefore, that theirexperience and knowledge of living organisms shouldbe different from that of non-aboriginal children. Itfollows that it would be beneficial for future elemen-tary school course design and teaching to investigateaboriginal children's ability to recognize and classifyliving organisms.

2. Purpose of the Research

96

The purpose of our research is to:(1) Understand the differences, if any, in aborigi-

nal children's conceptual comprehension of"living", "life" and "living organism".

(2) Understand aboriginal children's definition of"living organism".

(3) Understand how aboriginal children classify the

1107

Conceptions of "Living Organism"

familiar living and non-living objects in theirsurroundings.

(4) Study how aboriginal children decide whetheran unfamiliar living thing is a living organism.How do they apply their concepts? What issignificant in their recognition of life attributes?

3. Limitations of the Research

Our criterion of measurement was the children' sclassifying behavior. That is, we studied the children'scognitive perspectives based on the generalizationsand discriminations of their classification scheme.Under these limitations, the scope of our research wasas follows:

(1) Our study dealt only with the familiarity ofconcepts, not how they are formed.

(2) "Aboriginal children" was limited to childrenof the Taya tribe only. They did not includechildren of other aboriginal tribes.

(3) The content of the children's concept of livingorganisms was limited to what they remem-bered during interviews, which may not berepresentative of their full knowledge.

Based on the purposes of this study, we choseMandarin as the language of interview.

II. Methods

1. Subject

The subjects of this study were 36 children inthe 2nd, 4th and 6th grades of a Taya elementaryschool. The school is situated in a traditional Tayavillage in a mountainous region of Hualien. As in mostaboriginal schools of eastern Taiwan, there is only oneclassroom for each grade with 20 or fewer studentsin it. Selection of students was done randomly by theresearcher. Three groups of 12 students each with anequal number of boys and girls were selected from the2nd, 4th and 6th grade classrooms.

2. Instruments

The study collected data through clinical inter-views divided into two phases. In the first phase thechildren' s understanding of "living things"(Hwo Wu),"things of life"(Shen-Ming Wu) and "living organ-ism" (Sheng-Wu), and their classification of livingorganisms was investigated. The procedure was asfollows:

Their scope of understanding of the concepts "living","life" and

"living organism""Name five living things (Hwo-Wu)Name five things of life (Sheng-Ming Wu)Name.five living organisms (Sheng-Wu)"

Their definition of "living organism""What is a living organism(Sheng-Wu)? Tell what youmean."

Their classification of "living organisms"Presented with 24 cards depicting instances(plants andanimals) and non-instances(inanimate objects, man-madeobjects and natural things) children were asked: "Is thisa living organism (Sheng-Wu)? Why do you think so? Give

as many reasons as you can."

In second phase, five living objects never taughtto the children were used to examine their conceptof "living organism". The children were asked threemain questions: "Is this a living organism(Sheng-Wu)?"; "Is this a living thing(Hwo-Wu)?" ; "Why doyou think so?"

Based on Berzonsky's and Tamir's studies(Berzonsky, Ondarado, & Williams, 1977; Tamir, Gal-Choppin, & Nuuinovitz, 1981), the illustrations on thecards depicted the following: tiger, bird, fish, girl,snail, snake, tree, grass, rose, vegetable, peach, mush-room, chair, car, ship, bicycle, clock, television, cloud,fire, mountain, river, lightning and sun. Objects oflife taken from around the children's own environmentwere: crustaceous lichen, liverworts, cushiony mosses,filamentous green algae and slugs.

The testing instrument was submitted to a panelof evaluators made up of two university biologyteachers, a science educator, an elementary schoolteacher and graduate student, and an aboriginal schoolteacher. Each item was reviewed for consistency withthe classification task, and for clarity of expression.The instrument was finalized through field testing ona group of elementary school children.

3. Analysis

Children were interviewed at school in a roomnear their home classroom. Each interview lasted 20-25 minutes and was videotaped for later analysis. Thefollowing scoring procedures were employed: Ex-amples elicited for each concept were recorded. If achild provided more than 5, only the first 5 exampleswere scored. All traits mentioned in their workingdefinitions were listed and tabulated. The justifica-tions for their classification were divided into thefollowing categories:

97

108

S.H. Chen and C.H. Ku

(1) Attributes of living organisms: (i) growth, (ii)breathing, (iii) reproduction, (iv) living, (v) life,(vi) death.

(2) Attributes of plants: (i) Photosynthesis, (ii)absorbing water, (iii) growing in soil, (iv) havingstems, flowers, seeds, (v) living in a specificlocation, (vi) being green.

(3) Attributes of animals: (i) movement, (ii) eating,(iii) excretion, (iv) making sound, (v) attacking.(vi) having specific behavior, (vii) having aspecific location, (viii) having a heart, (ix) beingan animal.

(4) Attributes of non-living organisms: (i) edibil-ity, (ii) absorbing nutrients(misused in plant),(iii) having the pith of a tree, (iv) usability, (v)showing intuitive responses.

Based on the above categories, four types ofthinking were identified in aboriginal children.

III. Results

1. Aboriginal Children's Understanding of theConcept of "Living Organism".

A. Examples of Living Organisms

The average number and scope of examplesof "living things", "things of life" and "livingorganisms" given by the children are shown asfollows (Table 1):

The aboriginal children had a different under-standing of "living things", "things of life" and "livingorganisms". Their performance was best with "livingthings", next with "things of life", and worst with"living organisms". "Living things" and "things oflife", though not different in colloquial Chinese usage,obviously held a different degree of familiarity foraboriginal children. As for "living organism", mostchildren did not know what it meant. This indicatedthat concepts from classroom teaching were not beingbrought home to daily life. Aboriginal children still

Table 1. Average 'and Range of Number of Examples for "LivingThings", "Things of Life" and "Living Organism"

2nd grade 4th grade 6th grade(N=12) (N=12) (N=12)

average range average range average range

had a problem with understanding the term "livingorganism".

B. Content of Living Organisms

An analysis and categorization of examples of"living things", "things of life", and "living organ-isms" given by the aboriginal children yielded thefollowing:

(1) The children' s knowledge was richest withrespect to "living things", with a total of 36animals, plants, and prokaryotes. Next was"things of life" with a total of 34 plants, animals,and living organisms (Sheng-Wu). "Living or-ganisms" had only 24 animals and plants given.Of all the examples, animals comprised thegreatest number. Next was plants. Livingorganisms (Sheng-Wu) and prokaryotes had thefewest. "Living organism (Sheng-Wu)" and bac-teria were used by only one child each.

(2) The Children's examples of "living things" and"things of life" were mainly animals. For"living organism", more total animals weregiven, but more children used plants as ex-amples, (Table 2). This indicated that mostchildren thought of animals as "living things"or "things of life", and plants as "living organ-isms" .

(3) When giving animals as "living things" and"things of life", the most common exampleswere "human" and "animal". The next were zooanimals such as "lion" and "tiger". Next cameanimals that are found in the children' s envi-ronment, such as "bird", "cat" and "dog". In thechildren' s thinking, the label "animal" is a classon the same level as human, lion, tiger, bird,Cat and dog.

(4) The children gave a very limited number ofplants as examples. Most used "tree", "flower","grass", or "plant". Very few used specificplant names. This showed that in children' sthinking the label "plant" is a class on the samelevel as tree, flower and grass.

(5) Generally speaking, the content of "livingthings" and "things of life" increased in com-plexity in the case more mature children.However, the mistake of classifying plant itemsand animal items as comparable to the sets"plant" and "animal" themselves also increased(seven 6th graders, five 4th graders and one 2ndgrader made this mistake).

Definition of Living Organisms

When asked what they meant by the term "living

*living things 1.92 0-5 4.25 0-5 3.25 0-5*things of life 1.25 0-5 2.58 0-5 2.92 1-5*living organisms 0.42 0-5 0.58 0-5 1.83 0-5 C.

Note: Each categery contains 5 examples.

9809


Table 2. Most Frequently Named "Living Things", "Things ofLife" and "Living Organism"

Provided as:

living things(N=36)

things of life(N=36)

living organisms(N=36)

Animals:1. animal 13 10 02. human 8 13

3. tiger 8 5

4. bird 6 3

5. elephant 5 1

6. lion 5 3 2

7. dog 5 3 08. monkey 4 0 1

9. rabbit 4 0 010. cat 4 3 011. snake 4 1 012. goat 3 1

13. fish 3 1 2

Plants:1. tree 9 11 5

2. flower 5 4 4

3. plant 3 2 2

4. grass 1 2 0

organism" (Table 3), none of the 2nd and 4th graderscould give a definition'. Among the 6th graders, onlyone child could use the biological definition "growth"to describe living organisms. One child used an at-tribute of plants ("breathes in carbon dioxide, breathesout oxygen") and an attribute of animals ("canmove") to describe living organisms. Another childused the colloquial term "alive" to describe livingorganisms.

2. Aboriginal Children's Classification ofLiving Organisms

Table 3. Number of Definitions of Living Organisms and LivingOrganism Attributes

2nd grade(N=12)

4th grade 6th grade(N=12) (N=12)

total(N=36)

No. of definitions 0of living organisms

0 3 3

Attributes:1. living 1 1

2. moving 1 1

3. breath in carbon dioxydebreath out oxygen

1 1

4. possess life 1 1

5. will grow 1 1

A. Classification of Living Organisms

As seen in Table 4 and Fig. 1, most childrencould identify over 4 animals or plants as livingorganisms. In general, the more senior the grade, thebetter the performance. But the 4th graders did notperform significantly better than the 2nd graders,Comparing the performance in each grade, we foundthat 2nd graders tended to classify animals as livingorganisms but not plants. The 6th graders tended toclassify plants as living organisms but not animals.

For each of the plant item's, Fig. 2 shows that:(1) The percentage of children who classified

grass, peach and mushroom as living organismsincreased with age.

(2) The 4th graders did no better than the 2nd gradersof classifying tree, rose and vegetable. In par-ticular for rose, only 67% of the 4th gradersclassified it as a "living organism", which wasless than for the 2nd graders (75%)

(3) In the classification of plants, the 2nd and 4thgraders showed a wide discrepancy, with cor-rect responses ranging from 50% to 75%. The6th graders were more consistent, with correctresponses ranging from 83% to 100%. Thisindicated that the older children were morepositive about plants being living organisms.

(4) Mushrooms are a member of the fungi and mostchildren, regardless of their grade, could iden-tify them as a living organism. Yet the veg-etables we eat on a daily basis had the lowestpercentage of being identified correctly as aliving organism by the 2nd and 4th graders(about 50%).

For the animal items, the children's performancedid not necessarily improve with age. Fig. 3 demon-strates that:

(1) On the whole, 2nd graders had the highestpercentage of classifying animals as living

Table 4. Average and Range of Number of Correct Living Organ-ism Classifications

Plant° animal

average range average range

2nd grade 3.83 0-6 4.75 1-6(N=12)4th grade 4.00 0-6 4.17 0-6(N=12)6th grade 5.58 3-6 5.08 0-6(N=12)

involve mushroom

99

no

100

80

60

40

20

0 plantEl animal

80%

64% 67%70%

93%85%


2nd grade 4th grade 6th grade

Fig. 1. Percentage of correct response for "animal and plant items"by aboriginal children of each grade.

organisms. Next were the 6th graders. Theworst were the 4th graders.

(2) In the classification of living organisms,"human"(girl) was regarded differently fromother animals. The lowest percentage of chil-dren in each grade classified human as a livingorganism,barely 50%.

(3) For non-human animals, the 2nd and 4th gradersshowed a large discrepancy in their viewpoint.The percentages for snake and snail beingclassified as living organisms were lower thanthat for tiger, fish and bird. The 6th graders,however, were Jnore consistent. Snake, snail,tiger, fish and bird were classified as livingorganisms by about the same percentage of chil-dren.

B. Attributes of Living Organisms

i. Living Organism Attributes of Plants

After children classified tree, grass, rose, veg-etable, peach and mushroom as living organisms, theywere a'sked to justify their choice. Table 5 shows theresults as follows:

100

80

60

40


Fig. 2. Distribution of correct response for "plant items" by aborigialchildren of each grade.

* treegrassrose

* vegetable

peach

mushroom

100

e- 100

a0. 80

400.7

60

d.'' 200.) 2nd grade 4th grade 6th grade

* tigerbird

fish

* girlsnail

snake

Fig. 3. Distribution of correct response for "animal items" by ab-original children of each grade.

(1) There were 3 types of reasoning children usedwhen classifying plants as living organisms.One was the presence of living organism at-tributes; one was the presence of plant attributesand one was the presence of non-living organ-ism attributes. As for frequency of use, thegreatest number of children used plant attributes,followed by living organism attributes. Veryfew used non-living organism attributes.

(2) Among living organism attributes, "growth","breathing" and "reproduction" are scientifi-cally acceptable, while "life" and "death" arecolloquial terms, not mentioned in biologytextbooks. "Growth" was the attribute mostoften used by all children in classifying plantsas living organisms. "Breathing" was usedmostly by 6th graders. "Reproduction" was notmentioned much. The colloquial terms "life"and "death" were not used much by the 2nd and4th graders, but were used quite often (almost50%) by the 6th graders.

(3) Plant attributes are not the basis for classifyingliving organisms. However, most childrenused them, and moreover the use increased withage. Attributes used by 6th graders included"photosynthesis", "absorbing water", "grow-ing in soil", and "specific location". Theseattributes were rarely given by the youngerchildren.

(4) The non-living organism attributes were usedmostly by the 6th graders. The terms they usedmost often were "edibility" and "absorbingnutrients". The former came from the children'spersonal life experience. The latter might havecome from a misunderstanding of what they hadbeen taught. This showed that though childrencould classify living organisms correctly, theymight not be using the correct attributes.

lii


Table 5. Analysis of Attributes Used in Classifying Plants as Living Organisms by Each Grade of Aboriginal Children

Attribute 2nd grade 4th grade 6th grade total(N=12) (N=12) (N=12) (N=36)

Attributes of living organisms:1. growth 5 4 8 17

2. breathing 3 4

3. reproduction 0 2

4. life 2 3 5

5. death 5 6

Attributes of plants:I. photosynthesis .0 5 5

2. absorbing water 3 6 9

3. grows in soil 3 6 10

4. has stems, leaves, flowers, seeds 3 5

5. lives in specific location 4 6 10

6. green 0

Attributes of non-living organisms:1. edibility 3 3 3 9

2. absorbing nutrients' 1 3 4

3. has pith of tree 1 0 1

4. usability o 0 2 2

5. intuitive responses 1 1 3

Misused criterion

ii. Living Organism Attributes of Animals

Table 6 shows the children' s use of attributes inclassifying animals as living organisms. It shows that:

(1) Children used three types of attributes to decideif animals were living organisms: living organ-ism attributes, animal attributes, and non-livingorganism attributes. The most frequently usedwere the animal attributes. Next were the livingorganism attributes. The least used were thenon-living organism attributes.

(2) Among the living organism attributes, only"growth", "breathing", and "reproduction" arecriteria used by biologists . Althought the rest-"living", "life" and "death"- are colloquial terms,the children used them more frequently. Thisindicates that even if viability is not used bybiologists, children tend to mostly rely on it toclassify animals as living organisms.

(3) Regardless of grade, children used animal at-tributes with the highest frequency, indicatingthat most children use them to classify livingorganisms. Among animal attributes, "move-ment" was the most frequently given, (over halfof each grade used it), followed by "attacking","eating", "making sound", and "specific behav-ior". This shows that the children rely on ananimal's behavior to decide if it is a livingorganism. One thing worth mentioning is that

1cl

some children related animals to their habitats,thinking that living organisms live in a "specificlocation". This attribute is very similar to thatof plants mentioned above.

(4) Non-living organism attributes were used veryrarely. The only two such attributes were "us-ability" and "edibility". This was similar to theusage for plant items, except that the percentagewas far lower.

C. Misclassification of Living Organisms

Figures 1-3 also show these facts:(1) The 2nd and 4th graders had a similar error rate,

far higher than that of the 6th graders.(2) The 2nd and 4th graders made more mistakes

on the animal and plant items. Except on "girl,"the 6th graders made very few misclassifica-tions.

The children' s reasons for misclassification ofanimal items result from two different types ofmisconception (Table 7): (1) Animals are living or-ganisms but humans are not animals, so humans arenot living organisms. This is why children did notthink of "girl" as a living organism. (2) Animals arenot living organisms, and only plants are. Thus ani-mals or things with animal attributes are not livingorganisms. For plants there were also two types ofmisconceptions: (1) Only animals are living organ-isms. Plants do not have animal characteristics and


Table 6. Analysis of Attributes Used in Classifying Animals as Living Organisms by Each Grade of Aboriginal Children

Attribute 2nd grade 4th grade 6th grade total(N=12) (N=12) (N=12) (N=36)

Attributes of living organisms:1. growth 4 1 3 8

2. breathing 0 0 2 2

3. reproduction 1 1 1 3

4. living 1 1 2 4

5. life 0 4 4 8

6. death 1 1 3 5

Attributes of plants:1.movement 10 6 9 25

2. eating 2 2 7 11

3. excretion 0 0 1 1

4. making sound 1 2 4 7

5. attacking 6 3 4 13

6. specific behavior 0 0 6 6

7. specific location 0 3 3 6

8. has heart 0 1 0 1

9. an animal 0 1 2 3

Attributes of non-living organisms:1. usability 1 0 0 1

2. edibility 1 1 1 3

'.Misused criterion

Table 7. Number of Children by Grade Citing Reasons Why Animals and Plants Are Not Living Organisms

Reasons for not beinga living organism

grade

2nd(N=12)

4th(N=12)

6th(N=12)

total(N=36)

Animal:1. is human not animal 3 3 3 9

2. is a snail 1 0 0 1

3. bites people 1 2 0 3

4. eats animals 1 1 0 2

5. clings on people 1 0 0 1

6. eats 0 1 0 1

7. comes up from water 0 1 0 1

8. is an animal 0 2 1 3

9. is not a plant 0 0 1 1

10. is not a tree 0 1 0 1

11. does not have stems 0 1 0 1

Plant:

1. does not move 3 3 0 6

2. eten by people 2 3 1 6

3. planted by people 0 1 0 1

4. can be crushed 1 0 0 1

5. will wither 1 0 0 1

6. is a vegetable 1 0 0 i

7. is not an animal 0 1 0 1

8. does not grow 1 0 0 1

9. has no life 0 1 0 1

10. has ants on it 1 0 0 1

102

lin


therefore are not living organisms. (2) Plants do nothave life characteristics and therefore are not livingorganisms.

3. Aboriginal Children's Classificatory Think-ing Towards Living Organisms

Children' s reasons for classifying animals orplants as living organisms reflect 4 types of thinking:

(1) "Living organism" type: classification was basedon the presence of viability or biological at-tributes.

(2) "Animal" type: classification was based onwhether it was a kind of animal or if it hadanimal attributes.

(3) "Plant" type: classification was based on whetherit was a kind of plant, or if it had plant attributes.

(4) "Non-living organism" type: classification wasbased on non-living, non-biological attributes,or attributes that were unrelated to animals orplants.

Figure 4 presents the percentages of the abovefour types of thinking by grade level. It shows that:

(1) The living organism type of thinking increasedwith grade level.

(2) The animal type of thinking was most commonamong all grades, especially among 2nd grad-ers. All children had a tendency toward using

.animal class and attributes to classify livingorganisms.

(3) The plant type of thinking was least commonamong the 2nd and 4th graders, only 17% and25% respectively, but reached 92% for the 6thgraders.

(4) The non-living organism type of thinking was

100

80

60

40

20

animal

living organism

LA

non-living organism

plant


Fig. 4. Distribution of the four types of classification thinking byaboriginal children of each grade.

least common among the 6th graders, only 17%,but 33% among the lower graders.

(5) The preferred types of thinking tendency foreach grade were as follows: (i) The 2nd graderstended toward the animal type, with livingorganism and non-living organism types next.The plant type was the least common. (ii) The4th graders did not show a strong tendencytoward any one type, but living organism andanimal types were used slightly more than plantand non-living organism. (iii) The 6th gradersused plant and animal types the most with theliving organism type next and the non-livingorganism type the least common.

To understand the patterns of thinking the chil-dren used to classify plants and animals, combinationsof the above four types were analyzed. The resultsare shown in Table 8 and summarized as follow:

(1) Not many children used a single type of think-ing. Most used a combination of two or moretypes.

(2) Of those who used a single type of thinking, theliving organism type was the most common,appearing among roughly 1/4 of the 4th graders.Next was the animal type, appearing amongroughly 1/6 of the 2nd graders. Only a few usedof the non-living organism type exclusively.These were among the 4th and 6th graders.There was no instance of plant type thinkingbeing exclusively used in any grade.

(3) Most children used a combination of types ofthinking. (i) The 2nd graders tended to use"living organism type + animal type" and "ani-mal type + non -living organism type". (ii) The

Table 8. Living Organism Thinking Patterns by Aboriginal Chil-dren of Each Grade

Patterns of thinkinggrade

2nd(N=12)

4th(N=12)

6th(N=12)

total(N=36)

0 3 0 3

II 2 0 0 2

IV 0 1 1 2

I + II 4 3 0 7

II + III 0 2 2 4

II + IV 3 1 0 4III + IV 0 1 o 1

I + II + III 1 o 8 9

I + II + IV I 1 o 2

II + III + IV 1 0 0 1

I + II + III + IV 0 0 1 1

103


4th graders used "living organism type + animaltype", followed by "animal type + plant type".(iii) The 6th graders were more consistent;2/3 of them used "living organism type + animaltype + plant type".

(4) When solving the problem of whether an animalor plant was a living organism, the children useda total of 11 forms of thinking: 3 were singletypes, which appeared among the 2nd, 4th and6th graders. 4 were combinations of 2 typesand 3 were combinations of 3 types, each ofwhich appeared in all the grades. The combi-nation of all 4 types was used only among the6th graders.

4. Aboriginal Children's Application of theConcept of Living Organisms

A. Identification of Living Organisms and Living Things

Children were asked to discriminate between"living organisms" and "living things" by comparing5 objects of life unknown to them. The results areas follows (Table 9):

(1) Children did best in the case of cushiony mosses,liverworts and slugs. Next was filamentousgreen algae. Crustaceous lichen was the mostdifficult.

(2) As for "living things" versus "living organ-isms" , regardless of grade or the object, childrendid better for "living things" than for "livingorganisms".

(3) According to grade, when it came to identifying"living things", the more senior the grade, thebetter the performance, generally. But in iden-tifying "living organisms" this was not the case.The 4th graders were not better than the 2ndgraders.

(4) Most children knew that cushiony mosses,liverworts, slugs, and filamentous green algaewere "living things", but they did not necessar-ily know that these were "living organisms". Asfor crustaceous lichen, most did not know it wasliving, nor knew it was a living organism.

B. Attributes of Living Organisms and Living Things

The attributes and reasons used by the childrenin identifying living organisms and living things canbe grouped into the following 7 categories:

(1) Life: Children thought something was a livingorganism or living thing because it possessedlife. For example, "It will die", "It lives in thisworld", etc.

(2) Animal or plant characteristics: Children clas-sified something as a living thing or livingorganism because it had animal or plant shape,function or class. For example, "It has leaves","It drinks water",etc.

(3) Growth and development: For example, "Itgrows", "It will sprout", etc.

(4) Habitat: Children classified something as aliving organism or living thing because it wasfound in a life supporting environment. Forexample, "It grows on rock", "It lives in water",

Table 9. Number of Aboriginal Children in Each Grade Who Identified Objects of Life as "Living Thing" or "Living Organism"

objects of life

grade

2nd(N=12)

4th(N=12)

6th(N=12)

total(N=36)

cushiony moss:living thing 12 11 11 34living organism 10 9 10 29

liverwort:living thing 9 11 12 32living organism 9 9 11 29

filamentous green algae:living thing 6 8 10 24living organism 4 8 7 19

crustaceous lichen:living thing 4 4 3 11living organism 4 3 3 10

slug:living thing 9 12 12 33living organism 7 6 10 23

104

L 115


etc.(5) Personal experience: For example, "I' ve seen

this", "I've eaten it".(6) Reproduction: For example, "It's born", etc.(7) Others: Children used other reasons or personal

misconceptions to classify something as a. liv-ing organism or living thing. For example, "It'sgreen", "It floats on water", etc.

An analysis of the percentages of each of theabove categories and distribution among each gradeis presented in Fig. 5. It shows that:

(1)Children cited more reasons for "living things"than they did for "living organisms".

(2) Children mainly relied on "animal and plantcharacteristics", "growth and development","personal experience" and "habitat".

(3) In the "life" category, children performed dif-ferently than in the other categories because ofthe questions asked: "Is it living?" and "Is ita living organism?" For the former, childrentended to think in terms of the attributes ofliving. For the latter, they tended to consider" If it is living, it has life". This differencemade children use the "life" category more for"living organisms" than they did for "livingthings".

(4) Other than the "life" category, all the other 6categories were used more frequently for "liv-ing things" than for "living organism" .

(5)In the "animal and plant characteristics" cat-egory, regardless of their grades childrenused it to classify "living things" with a highpercentage (94%). However, in classifying"living organisms" the percentage was only58%.

(6) "Personal experience" saw the greatest differ-ence in use in classifying "living organisms"and "living things". The percentage for "living

100

80

60

40

20

94% 0 living organisms111 living things

69% 67%

life animal o growth personalplant and experience

characteristics developement

Fig. 5. Percentage of reasons used by aboriginal children to identifyliving organisms and living things.

others habitat reproduction

4, I

things" was 69%, with at least half of eachgrade relying on it. However, only 14% usedpersonal experience to classify "living organ-isms".

IV. Discussion

Based on the above results,this study made thefollowing findings:

1. Comprehension Of "Living Organisms", "Liv-ing", and "Life"

Living organisms show the characteristics ofliving or having life. The attributes for "livingorganism", "living", "life" are not different but "life"and "living" are everyday colloquial terms. Theformation of the concept comes from interactionbetween individuals dnd their environment and gradu-ally matures with age. "Living organism" is mainlyused in the educational environment. The acquisitionof this concept comes during the learning process ofeducation. Since the three terms have a commonconcept, children can use them interchangeably. Inresearch on non-aboriginal children's concept of life,Huang (1993) used "living organisms", "non-livingthings", and "things of life" interchangeably in hisinterviews, and children had no problem understand-ing that the terms mean the same thing. Yet our studyfound that aboriginal children have a different com-prehension of the three terms. The degree of famil-iarity was: "living" > "life" > "living organisms". Thisdifference in familiarity is probably the result of: (1)Inferior comprehension of "life" caused by the influ-ence of aboriginal language and culture. "Living" and"life" are both colloquial terms in Chinese culture sonon-aboriginal children are accustomed to using themin describing an object. In the Taya language, how-ever, there is only the term "living" ('mulus) and nosuch term as "life", which made the aboriginal childrenless likely to use it in expressing concepts. Hence thecomprehension of "life" was not as good as that of"living". (2) The reason for poor comprehension of"living organism" probably originated in the learningprocess where this concept was handed down withoutdirect experience. The children of this study fre-quently used personal experience such as "seen" or"eaten" to explain "living". Yet for "living organ-isms", children rarely cited such reasons, indicatingunfamiliarity based on a lack of firsthand ex-perience.

This study also found that aboriginal childrenused more animal items and fewer plant items in

105


illustrating their concept of "living organism". Thisphehomenon of knowing animals better than plants isconsistent with previous researches work (Lin &Yamasaki, 1986; Huang, 1987; Chen, 1989; Ryman,1974a; Stavy & Wax, 1989) and shows that children'sgreater familiarity with animals than with plants is acommon widespread phenomenon. Even the aborigi-nal children who live in a more natural environmentfollow this pattern.

2. Attributes of Living Organisms

In general, children of all grades frequently used"movement" to classify animals, and "growth" toclassify plants as living organisms. Even the olderchildren did not use biological attributes much. Thiswas consistent with the findings of Huang (1993),Tamir et al. (1981), Stepans (1985), and Stavy & Wax(1989) that comprehending living organisms is asdifficult for children everywhere as it is for non-aboriginal and aboriginal Taiwanese children. Besidesthe above two attributes, a high percentage of aborigi-nal children used animal and plant characteristics andhabitats as living organism attributes, especially ani-mal behavior and plant habitat. Behavior such asaggressiveness, movement and eating were found quitefrequently in previous researches on animism, andwere included among activity attributes (Wang, 1993;Su,1968; Russel & Dennis, 1939; Bruce, 1941).However, habitat and living environment were littlementioned. The percentage, if it occured at all, wasvery low (0-Saki & Samirdoen, 1990). The use ofenvironment as a living organism attribute by aborigi-nal children (for example, "grass grows in soil","vegetable grows in garden", "snakes hide in grass")indicated that their thinking concerning living organ-isms was different from that of the non-aboriginalchildren.

3. Classification of Living Organisms

As for the class concept of living organisms, thisstudy found that the children were ignorant of theattributes and scope of classes. Children knewthat human, animal, bird, plant, tree, grass and flowerwere living (that they possessed life) but they did notknow that these living things all belonged to theclass of "living organisms". Neither were they awarethat there is a hierarchical relationship between hu-man and animal, or between tree and plant. Theybelieved that all living things are under the class ofnatural things. Their classification system is as fol-lows:

Human Natural things

I i i I 1 I

bird tiger fish grass tree plant animal

This single-leveled classification structure is verysimilar to that of New Zealand children as discoveredby Bell (1981): However, there are some differences.Our abori2inal children did not think of "human" asa living organism. They believed humans are humans.Humans are living but are not living organisms. Thisdifference might have stemmed from the belief ofaboriginal culture that a human is a unique being apartfrom other living organisms. It might also havestemmed from language differences. In the Chineselanguage, living and living organism are two dif-ferent terms, whereas in the English language, livingand living thing share a common linguistic symbol.This difference in linguistic expression might bewhy aboriginal children could not realized that ahuman is living and therefore can be classified as aliving organism. Another difference from the NewZealand children is that the younger aboriginal chil-dren tended to think of animals as living organisms,while the older children tended to think of plants asliving organisms. This discrepancy might stem fromtheir learning process. The biology portion of naturalscience courses for the lower grades covers animalsmore, while the higher grade material covers more onplants. Also, the fact. that the concept of classrelationship between "living organism", "plants" and"animals" is not often properly clarified may be afactor.

4. Identification of Living Organisms

When it came to identifying something as a "livingorganism", the aboriginal children' s process andreasons for giving a particular response followed anorder of difficulty as follows:

(1) They can tell that something "Is not a living organism" easily.

(2) They.can tell that something "is a living organism" somewhat

easily.

1.

(3) They can list "examples of living organisms" less easily.

(4) They can recite a "biological definition" with some difficulty.

(5) They can give a "biological definition" with great difficulty.

It was easier for the children to identify thatsomething was not a living organism, than that it was

106.1/


to identify a living organism.

5. Characteristics of Living Organism Think-ing

From the aboriginal children's judgment of andjustifications for something being a living organism,we could see that their thinking might have the fol-lowing characteristics:

(1 )Lacking a concept of class-inclusionThe children generally thought of animal andplant items such as bird, fish, animal, tree, grass,plant, as all being on the same level as "animal"and "plant" and equal to "living organism".They were not aware that there is a hierarchicalrelationship. This was particularly so amongthe younger children. When asked "Is it a livingorganism?", they intuitively interpreted thequestion as "Is it equivalent to living organ-ism?", rather than (as an adult would) "Does itbelong to the class of living organisms?. Thislack of the concept of class-inclusion mighthave caused the children's difficulty in compre-hending "living organism".

(2) Self-centeredFrom the interviews, we found that most of theyounger children tended to be strongly self-centered. Russell & Dennis (1939) found thatduring the early development stage, childrentend to think of things that are useful or im-peccable as living. Our aboriginal childrenwere even more subjective. They were human-centered, dichotomizing things into "useful"and "useless", "good" and "bad". Those theythought were useful to humans, or consideredin a positive way, were living organisms. Thosethat were bad, or thought of in a negative way,were not living organisms.

(3) Relying on experienceThe children's concept of living organisms wasvery vague. They did not know'what the criticalattributes were. The living organism attributesthey used were often all the possible life or non-life characteristics. This is a probabilistic per-spective on life.

(4) Environmental attachmentCompared with non-aboriginal children, aborigi-nal children's life attributes possessed moresuperficial qualities, such as animal and plantproperties and physical structures. What waseven more different from non-aboriginal chil-dren was that the aboriginals were stronglybiased toward the environment of living things,

believing that the environment of living thingswas, if not a living thing itself, at least thefoundation of life. The children's thinking fol-lowed two patterns:(i) Association of life:

Thinking that there is an association be-tween the environment and living things,believing living things and the growingenvironment are both living organisms.Hence, "The sun gives light to trees, so treescan grow", therefore sun is a living organ-ism. "A mountain grows trees, a river hasfish and plants", therefore mountains andrivers are also living organisms.

(ii)Foundation of life:Thinking that the environment is an impor-tant attribute of life and believing that theenvironment where a living organism livesis an important attribute of the organism

-itself (because if the living organism "leavesit will wither, it will die"). Therefore anenvironmental attachment is an importantbasis for identifying living organisms.

V. Conclusions and Suggestions

1. Conclusions

(1) The aboriginal children had a different compre-hension of the concepts "living organism","living", and "life". Language, culture, andlearning experience might have been the mostimportant factors influencing this.

(2) In classifying living organisms, the aboriginalchildren tended toward using a single-leveledclassification structure.

(3) The living organism attributes used by theaboriginal children were not consistent withthose used in the natural science courses forelementary schools. The living organism at-tributes they used most were "movement" and"growth". They also emphasized the relation-ship between "living organisms and environ-ment".

(4) The children had four types of thinking whenidentifying living organisms: (i) living organ-ism type, (ii) animal type, (iii) plant type,and (iv) non-living organism type. The children'spattern of thinking was usually a combinationof two or more of the above four types buta different combination for each grade. Themajor patterns were:(i) 2nd grade: (a) living organism type + animal

107

11 8


type; (b) animal type + non-living organ-ism type.

(ii) 4th grade: (a) living organism type + ani-mal type; (b) living organism type.

(iii) 6th grade:living organism type + animaltype + plant type

(5) When identifying living and non-living organ-isms, the aboriginal children showed the fol-lowing degrees of familiarity: They can tell thatsomething "is not a living organism" > Can tellthat something "is a living organism" > Can list"examples of living organisms" > Can recite a"biological definition" > Can give a "biologicaldefinition"

(6) The most often cited reasons for identifying"living organisms" and "living things" were: (i)animal and plant characteristics, (ii) growth anddevelopment, (iii) habitat, and (iv) personalexperience. There was a great discrepancy inusing "personal experience" for "living organ-isms" and for "living things". The childrenrarely relied on personal experience in identi-fying "living organisms", but often in identi-fying "living things".

2. Suggestions

Based on our findings, we suggest the followingfor future teaching methodology:

(1) Improve children' s language ability.In this research, we foulid that aboriginal chil-dren used short sentences, lacked vocabulary,and were imprecise in expression of meaning.Therefore, improving the children' s languageability would be helpful to the acquisition ofthe living organism concept.

(2)S trengthen teachers' understanding of aborigi-nal culture.Teachers in the aboriginal elementary schoolsare mostly non-aboriginals, with a non-aborigi-nal way of thinking. Teachers should have anunderstanding of aboriginal culture so that theycan work from a pro-cultural basis and properlyconvey the information contained in their teach-ing materials.

(3) It would be helpful to the children' s acquisitionof knowledge if teaChers do the following whenteaching:(i) Provide examples and non-examples of

concepts.The acquisition of concepts is a process ofdiscrimination and generalization. If teach-ers do not provide 'enough examples and

non-examples for children to discriminateand generalize, misconception is likely tooccur.

(ii) Establish a hierarchical relationship of con-cepts.

(iii) Provide children with the opportunity for"hands-on" experience

Acknowledgment

The authors wish to thank the National Science Council forfunding this research(NSC 84-2511-S-026-002N).

Reference

Bell, B. F. (1981). Animal, plant, living; notes for teachers. Learn-ing in science project (Working Paper No.30). Hamilton: Uni-versity of Waikato.

Berzonsky, M. D., Ondarako, N. A., & Williams, G. T. (1977).Modification of the life concept on reflective and impulsivechildren. The Journal of Genetic Psychology, 130, 11-17.

Bruce, M. (1941). Animism vs. evolution of the concept "alive".Journal of Psychology, 12, 81-90.

Chen, S. H. (1989). A study of children's acquisition of selectedclassificatory concepts in life science relationship of masteryof concept and misconception with urban/rural, grade, sex, andcognitive factors. Hualien, Taiwan: Chen-Yi PublishingComapny.

Huang, D. S. (1993). A study of the development and the alternativeof life concepts in elementary school students. In: Proceedingsof the National Pingtung Teachers College Conference onElementary School Natural Science Education (pp. 84-111).Pingtung, Taiwan: National Pingtung Teachers College.

Huang, I., & Lee, W. H. (1945). Experimental analysis of childanimism. The Journal of Genetic Psychology, 66, 69-74.

Huang, T. C. (1987). A study of the difficulty of and relationshipbetween teaching and the learning of the concept of livingorganisms by first year high school students: Final report forNSC-74-0111 -S-003 -13. Taipei: National Science Council,R.O.C.

Lin, C. L., & Yamasaki, T. (1986). A study by associative analysison students' cognitive process of nature. Paper collections ofthe 1985 Conference on Science Education, pp. 61-87. Taipei.

0-saki, K. M., & Samiroden, W. D. (1990). Children's conceptionsof "living" and "dead". Journal of Biological Education, 24(3),199-207.

Russell, R. W., & Dennis, W. (1939). Studies in animism: I. Astandardized procedure for the investigation of animism. TheJournal of Genetic Psychology, 55, 389-400.

Russell, R. W. (1940). Studies in animism: II The development ofanimism. The Journal of Genetic Psychology, 56, 353-366.

Ryman, D. (1974a). Children's understanding of the classificationof living organisms. Journal of Biological Education, 8, 140-144.

Ryman, D. (1974b). The relative effectiveness of teaching methodson pupils' understanding of the classification of living organ-isms at two levels of intelligence. Journal of Biological Edu-cation, 8(4), 129-223.

Stavy, R., & Wax, N. (1989). Children's conceptions of plants asliving things. Human Development, 32, 88-94.

Stepans, J. (1985). Biology in elementary schools: children's con-

108

119


ceptions of "life". The American Biology Teacher, 47(4), 222-225.

Su, C. W. (1968). A study of the development of children's lifeconcepts. Examination Annual., 15, 522-529.

Tamir, P., Gal-Choppin, R., & Nuuinovitz, R. (1981). How dointermediate and junior high school students conceptualize livingand nonliving? Journal of Research in Science Teaching, 18(3),

t_111-Ak*.t.

241-248.Wang, M. F. (1993). Glossary of terms used by young children

in explaining the phenomena of life. Paper collections of theROC's 9th Conference on Science Education Coordinated byNational Changhua Normal University, pp. 382-410. Changhua,Taiwan: National Changhua Normal University.

RA&Szt

-f-Rat

N

3-i'fA2,h3FA

-'1**

*W91.1)..11taglillth /1N___1 IN 7:Vii.*IA 36 #11z. "It" Ryt..1. fiffAIZAMf141=13MIRIPJOECA 3-)-thAttgbITIAttOttt Akfift44k AlhITIM:52,2*Eltr1413i-MEA.±.4R1VOA (1) It (2) Utt (3) Wit (4) fAtt , Zr*AR5E1AUTAILLWIHrEstAttig , "tt" fo-

rth rt.-MJ MffilliVR.92,11%.4)it "IVO" RAVARIIIIIVAAt hiltattt c' "tt" tit.±.*MIRIAgla0 , ffi "an" i',1141fig. Elltrhht10VIIPANI.,5EI'Mtd V A 4C44E3M.4 "Mt" "Mt" Aft , ig1.1.:s2ATON-

tafalk*A. 1.1y *W-FAffig-R ri411N,AZ thIM-hAtitZNEig 0-Vgillta MIN% , *fF9-t(q:itaffina

109

120


An International Investigation of Preservice ScienceTeachers' Pedagogical and Subject Matter Knowledge

Structures

NORMAN G. LEDERMAN* AND HUEY-POR CHANG**

*Department of Science and Mathematics EducationOregon State University'Corvallis, OR, U.S.A.

Physics DepartmentNational Changhua University of Education

Changhua, Taiwan, R.O.C.

(Received November 11, 1996; Accepted April 10, 1997)

ABSTRACT

The nature and development of an international sample of preservice science teachers' subjectmatter and pedagogical knowledge structures were assessed as they proceeded through student teaching.Twelve American and 14 Taiwanese presef,vice science teachers were asked to create representationsof their subject matter and pedagogical knowledge structures before and after their student teachingexperience. They also participated in a videotaped interview concerning knowledge structure represen-tations immediately following student teaching. Qualitative analyses of knowledge structure represen-tations and transcribed interviews with and between subjects were performed. Initial knowledge structureswere typically linear and not coherent. Subject matter representations were stable, while pedagogystructures were susceptible to change, in the American sample, as a consequence of teaching. TheAmerican preservice teachers perceived pedagogy and subject matter as distinct and exerting separateinfluences on classroom practice, while the Taiwanese sample consistently exhibited difficulty in sepa-rating subject matter from pedagogy. Differences between the nature and development of the knowledgestructure representations of the two groups:were related to both cultural differences and differences inapproaches to teacher preparation between the two countries. Taken within the context of prior research,the results support the assertion that a clear relationship exists between the complexity of teachers'knowledge structures and subsequent translation into classroom practice. Given that only one teacherpreparation program from each country constituted the sample for the investigation, the reader is cautionedagainst over-generalizing the results to Taiwan and the U.S. in general. However, this initial cross-cultural investigation does raise some questions concerning the differences in approach to teachereducation in Taiwan and the U.S.

Key Words: teachers' knowledge structures, teacher education

I. Introduction

Recent concerns about the quality of teachereducation programs (Carnegie Forum, 1986; ;Hdlmesgroup, 1986; Kennedy, 1990) and the evaluation ofteaching (Shulman, 1986, 1987) have focused, atten-tion on the subject matter and pedagogical knowledgeof teachers. As a consequence, many states haveincreased subject matter requirements for admission

f

to teacher education programs. These increased re-quirements have taken the form of mandatory degreesin subject matter and/or subject matter competencyexaminations (e.g., National Teacher Examination)for prospective teachers. Such changes in policyhave been made in spite of the fact that prior attemptsto relate quantitative-oriented measures of whatteachers know (e.g., GPAs, college credit hours, degreesattained) with measures of effective teaching have

110

1.21

Science Teachers ' Knowledge Structures

rarely produced relationships of strong, practical sig-nificance (Brophy & Good, 1986).

The results of previous research, however, havenot caused educators and policy makers to abandonthe rather intuitive notion that a teacher' s subjectmatter knowledge necessarily influences classroompractice. Rather, it has been recognized that the "older"process-product oriented research paradigms (most ofwhich were quantitatively-oriented) are not sufficientto answer questions concerning teachers' subject matterknowledge, its formation, and its potential impact onclassroom practice. Prior research, whether quanti-tative or qualitative, has yielded little more thanrelationships among superficial attributes of teachers'thinking and classroom behaviors. Consequently, olderresearch paradigms have yielded to more in-depthqualitative measures of teachers' subject matter con-ceptual frameworks.

Recent attempts to explore teachers' conceptualunderstanding of subject matter have used a widevariety of approaches, notably semantic networks, wordassoeiations, concept maps, and various versions ofcard sorting tasks (Baxter, Richert, & Saylor, 1985;Hashweh, 1986; Hauslein & Good, 1989; Hauslein,Good, & Cummins, 1992; West, Fensham, & Garrard,1985; West & Pines, 1985; White & Tisher, 1985;Wilson, 1989; among others). Although such ap-proaches are often used in concert with interviewprotocols, respondents are typically asked to organizeand/or categorize topics or themes provided by theresearcher in order to elucidate underlying subjectmatter structures. Although the data yielded by theaforementioned techniques are qualitative in nature,the structure imposed on data collection arguablycompromises the benefits and purpose of using aqualitative research design. To date, relatively fewstudies have avoided the pitfalls of limiting subjects'representations of content knowledge to an a priori listof topics when assessing development over time.

Although the growth and role of subject matterknowledge within teachers' professional developmentis presently the source of much research and contro-versy, the parallel development and role of pedagogi-cal knowledge, with few exceptions (Hoz, Tomer, &Tamir, 1990; Lederman, Gess-Newsome, & Latz, 1994;Morine-Dershimer, 1989), and the interaction of-these two domains of knowledge has yet to be sys-tematically analyzed. Furthermore, the nature anddevelopment of pedagogy and subject matter knowl-edge structures among preservice science teachers indifferent countries is a totally uncharted area of in-vestigation.

The purpose of this international investigation

was to assess the nature, development and changes ofpreservice secondary scienee teachers' subject matterand pedagogical conceptions/knowledge structures asthey proceeded through their student teaching expe-rience. In particular, this investigation attemptedto answer the following questions: (1) What is thenature/form of preservice science teachers' subjectmatter and pedagogical knowledge structures? (2)What is the source(s) of these knowledge structures?(3) Are these knowledge structures stable duringstudent teaching?, and (4) What is the relationshipbetween these knowledge structures and how do theyrelate to the act of teaching? In addition to combineddata analyses, each of the research questions was usedto examine the similarities and differences between thetwo cultural groups (and distinct approaches to teacherpreparation) represented in the sample.

For the purposes of this investigation, "knowl-edge structure" refers to the knowledge an individualpossesses and the manner in which this knowledge isorganized. Our research definition is intentionallybroad and it is recognized that we might be moreaccurate in describing our teachers' knowledge as"conceptions" (and at times we use the terms synony-mously) of subject matter and pedagogy as opposedto formal knowledge structures. Whether the label"knowledge structure" or "conception" is preferred,such referents should not distract the reader from theprimary focus of this investigation: the nature, devel-opment, and changes of preservice science teachers'knowledge of subject matter and pedagogy as theyproceed through the student teaching experience.

II. Design

1. Sample

Twel ve preservice secondary school scienceteachers (seven biology, three general science, onechemistry, and one physics; seven males, five females)from the U.S. and 14 preservice teachers (seven physics,four biology, and two chemistry; eight males, fourfemales) from Taiwan were studied as they proceededthrough their student teaching. The American stu-dents were completing a one-year Master of Arts inTeaching (MAT) program and possessed at least a B.S.degree in their subject matter field (two had a M.S.degree and one a Ph.D.). The Taiwanese students werein their final year of a four-year teacher preparationprogram at a normal university. The Taiwanesepreservice teachers possessed the U.S. equivalent ofa B.S. plus 15 credit hours in their subject matterspecialty. Consequently, all of the preservice teachers

111

122

N.G. Lederman and H.P. Chang

(American and Taiwanese) possessed a level of subjectmatter knowledge well above that of most preserviceteachers. In particular, preservice teachers in the U.S.commonly possess only a B.S. or less within theirsubject matter specialty. Even with the proliferationof MAT programs, preservice teachers commonly donot possess much more than a B.S. degree within theirscience specialty. Each of the preservice teachers inthis investigation was seeking initial certification.

There were significant differences between theteacher preparation programs undergone by the U.S.and Taiwanese preservice teachers prior to their stu-dent teaching experience, which may have contributedsignificantly to several of the noted differences be-tween the two samples. The duration of the teacherpreparation program for the American preservice teach-ers was one year and consisted primarily of subject-specific pedagogy courses, with subject matterbackground required as a prerequisite for admission.Student teaching was performed during the thirdquarterof a four-quarter program. In contrast, the Taiwanesestudents proceeded through a four-year teacher prepa-ration program in which both subject matter andpedagogy were addressed throughout the four years.Science pedagogy was emphasized in subject mattercourses as well as within specific teaching methodcourses. As a consequence of Changhua Universityof Education's emphasis on the preparation of teach-ers, subject matter courses were consistently taughtin a context focusing on the ultimate teaching of subjectmatter to secondary students. Therefore, pedagogywas both implicit and explicit within subject mattercourses. Neither the subject matter courses nor peda-gogy courses in Taiwan or the U.S. explicitly empha-sized or discussed the "structure" of science disci-

,

plines or science pedagogy. Student teaching wascompleted at the end of the fourth year. Aif the timeof student teaching, however, each of the groups ofpreservice teachers had received instruction in learn-ing theories, teaching methods and strategies,microteaching, and had participated in field-basedpractica.

The student teaching experiences also differedsignificantly between the two programs. The Ameri-can preservice teachers worked full time in a schoolsetting and assumed full instructional responsibilityfor 3-4 classes (two preparations). Full instructionalresponsibility was assumed for a period of 10 weeks.The Taiwanese students worked full time in a schoolsetting for only one month, during which time theyassumed full instructional responsibility.

The two researchers (one from Taiwan and onefrom the U.S.) were well acquainted with the preservice

teachers as a consequence of their significant instruc-tional responsibilities within their respective programs.We believe that the rapport between researchers andsubjects served to facilitate the gathering of in-depth,accurate data and did not act as a hindrance.

2. Data Collection and Analysis

The case study design specified by Bogdan andBiklen (1992) was considered most appropriate forthis investigation. In this particular instance, the casestudy focused on two culturally different groups ofindividuals who were proceeding through two dis-tinctly different teacher preparation programs andstudent teaching experiences. Data was collected andanalyzed in two phases. Of initial interest was whetherpreservice science teachers possess coherent concep-tions and/or structures for their subject matter spe-cialty and pedagogical knowledge. This question wasaddressed primarily in Phase I. The additional ques-tions proposed by this study were addressed in Phase

A. Phase I.

Two weeks prior to the beginning of studentteaching, each subject was given approximately 30minutes to answer the following questions:

(1) What topics make up your primary teachingcontent area? If you were to use these topicsto diagram your content area, what would it looklike?

(2) Have you ever thought about your content areain the way you have been just asked to do so?

One week later, each subject was asked to answer thesame questions, but with "important elements/con-cerns of teaching" substituted for the phrase relatedto primary teaching content area. The preserviceteachers were asked to answer Question #1 againimmediately following the completion of student teach-ing. For the second administration of the question-naires, Question #2 was replaced with: "Have yourviews changed? If so, how and why?" Naturally,questionnaires were written and filled out in the nativelanguages of the preservice teachers.

It should also be noted that no specific methodsOf formatting or organizing the subject matter and.pedagogy "diagrams" were suggested to the preserviceteachers. In addition, the preservice teachers were toldthat their descriptions of subject matter and pedagogycould focus on topics, themes, processes, strands, etc.and could be "represented" by use of a diagram, conceptmap, picture, description, or in any manner with whichthey felt comfortable.

112

S11 2 3

Science Teachers' Knowledge Structures

Overall, it was felt that this methodology wassuperior to past attempts to assess subject matter andpedagogical knowledge structures because it gave re-spondents the freedom to select their own topics,themes, processes, or strands, etc. (as opposed to cardsorts) and to organize these elements of knowledgein any manner with which they felt comfortable (asopposed to artificially forcing representations intocategories, hierarchies, dimensions, or particularformats). It was hoped that this approach wouldprovide a clearer portrait of the preservice teachers'conceptions/structures of subject matter and peda-gogy. All representations and written text producedby the Taiwanese preservice teachers were trans-lated into English prior to data analysis by the firstauthor.

Qualitative analysis of the data collected duringthis phase (two administrations each of the subjectmatter and pedagogy questionnaires) attempted toderive any evident patterns between and within bothgroups of preservice teachers' stated subject matterand pedagogical structures. This initial analysis (con-ducted by the first author) served as a guide foradditional data collection during a follow-up interviewwhich took place one to two weeks after the comple-tion of student teaching.

B. Phase II.

Within two weeks following the completion ofstudent teaching an attempt was made to assess changesin the preservice teachers' knowledge structures andclarify any patterns elucidated in Phase I. EachAmerican subject was asked to participate in a 45-60minute videotaped interview conducted by the firstauthor, while the Taiwanese preservice teachers wereinterviewed by both researchers with the second authorserving as translator. The interviews consisted ofquestions that asked the subjects to describe theircurrent knowledge structures, discuss changes whichhad occurred and any reasons for these changes, dis-cuss any relationships between the knowledge struc-tures or between either knowledge structure and theirteaching, and their feelings about completing thequestionnaire. During the interview, the previouslycompleted knowledge structure diagrams/representa-tions were displayed and discussed individually andas a group. Finally, all subjects were given an op-portunity to revise the second diagrams/representa-tions produced for subject matter and pedagogy toconform to any changes which might have occurredsince its completion.

Importantly, the interview was also viewed as ameans to compensate for any confusion created by the

pencil-and-paper questionnaire (either with respect tothe respondents' reactions or the researchers' inter-pretations of responses). The problems associatedwith researchers' attempts to infer individuals' con-ceptions, knowledge, and beliefs solely from pencil-and-paper materials have been clearly recognized(Lederman, 1992). All interviews were transcribed(and translated when necessary) for analysis. Datawere compared within and between individuals to deriveany evident patterns for this particular group ofpreservice teachers.

Both phases of data analysis were conducted bythe first author with the second author independentlyanalyzing both the Taiwanese and American preserviceteachers' knowledge structure representations andvideotaped interviews. The second author is fluentin both Taiwanese and English. The independentfindings of the two researchers were compared, con-trasted and discussed. Given the cultural differencesbetween the two groups of preservice teachers and theresearchers, this was a critical step in the analysis ofthe data. There was no attempt to achieve totalagreement between the perceptions of the two re-searchers. Such an attempt would only have servedto eradicate the richer understanding which wasgained from the different perspectives brought to thedata analysis by the use of multiple researchers(Bogdan & Biklen, 1992; Eisner, 1991; Lincoln &Guba, 1985). The result was a clearly more compre-hensive and deeper understanding of the preserviceteachers' conceptions, while at the same time pro-tecting interpretations from being overly influencedby the perspective of an individual researcher(Lederman & Gess-Newsome, 1991; Miles &Huberman, 1984).

Ill. Results and Discussion

The reported results represent the culmination ofseveral rounds of .data analysis, by each of the tworesearchers, and have been organized in terms of theinitial questions guiding the investigation.

1. What is the Nature/Form of PreserviceScience Teachers' Subject Matter and Peda-gogical Knowledge Structures?

Interview responses indicated that the preserviceteachers were quite hesitant while completing the first(and sometimes the second) subject matter question-naire. Many felt tentative or uncertain about what towrite. They indicated that they had no problem un-derstanding the question or task at hand, but rather

113

1 2 4t


were hesitant about the content (and quality) of theirresponses, as indicated by the following representativecomments:

I knew what I was supposed to do. Still, you don't want tolook like you don't know what you're talking about. I knowmy subject matter well, but I worry about communicating myknowledge to others. (American Preservice Teacher)

Knowing your subject matter is important. It would be verybad if a teacher did not know his subject. I did not want youto think I did not know my subject. (Taiwanese PreserviceTeacher)

In short, both groups of preservice teachers wereconcerned that the questionnaire was a test of theirsubject matter understanding. No similar hesitancyor concern was expressed with respect to the pedagogyquestionnaire, but the Taiwanese preservice teachersexpressed much difficulty in conceptualizing or dis-cussing pedagogy apart from their subject matter.

I am to become a physics teacher. So, I do not know how tothink about teaching separately from physics. (TaiwanesePreservice Teacher)

This was a lot easier than the other questionnaire. I've seenso many teachers in my life, it's pretty easy to figure out whatit's about. (American Preservice Teacher)

Initial subject matter representations were prima-rily listings of discrete topics or science courses takenat the university. The Taiwanese group, however,consistently included various pedagogical concerns(e.g., teaching approach, level of students) within theirsubject matter representations. The inclusion ofpedagogical concerns within the subject matter rep-resentations of the Taiwanese preservice teacherswas a consistent pattern throughout the investi-gation, serving to further reinforce the inability ofthe Taiwanese group to separate conceptions ofsubject matter from the teaching of the subjectmatter.

The pedagogical structures were primarily list-ings of the teacher-oriented components of instruc-tion. Student-oriented components of instruction (suchas motivation, prior knowledge) were given little orperipheral attention by the American group while theTaiwanese preservice teachers consistently took stu-dents as a focal point. The presence of integrativecurriculum themes (e.g., nature of science, S-T-S) orconnections between or within the components of eithersubject matter or pedagogical structures were not

commonly noted by either group of preservice teach-ers. Again, it is important to note that the oral in-structions provided with the questionnaires explicitlyemphasized that the word "topics" need not be takenliterally and that respondents should feel free to in-clude topics, themes, processes, or strands, etc. Inaddition, it was also emphasized that representationsneed not be "diagrams," and could take whatever formmost accurately portrayed each individual' s concep-tions.

Organizational patterns were quite traditional withrespect to subject matter. In general, subject matterstructures were presented in three general formats:discrete (Fig. 1), simple hierarchy (Figs. 2 and 3), andweb-like (Fig. 4). The Taiwanese group overwhelm-ingly presented subject matter in the form of simplehierarchies (again, with pedagogical factors included)while their American counterparts could be primarilycharacterized as striking a balance between discreteformats and simple hierarchies. The web-like formatwas clearly not common within either group ofpreservice teachers. Naturally, the labels used to

Biochemistry

11\

PopulationGenetics

Ecology

CellBiology

Genetics AnatomyTaxonomy

Fig. 1. Discrete topic/course format for subject matter structure.(American Preservice Teacher)

Physiology

EVOLUTION

Plant Animal

Ecology

1/7

Micro

Science Methods

History EnvironmentalStudies

Fig. 2. Simpie hieracrchy format for subject matter structure (Ameri-can Presservice Teacher)

114

1125


PHYSICS

Mechanics Electricity

Magneficism

Optics Heat Modem

PhysicsLab

Motion

Rotation

Collision

FaradayAmpere

Maxwell

Optics

Wave

Geometry

I str2nOrdLaw

Einstein

Planck

Rutherford

Labs of

All Kinds

Common Components

PHYSICS

Fig. 3. Simple hierarchy format for subject matter structure (Tai-wanese Preservice Teacher)

EARTH SCIENCE, ENVIRONMENTAL SCIENCE,ECOLOGY, BIOLOGY

FundamentalPhysica. Chemistry

ChemicalPOlution

TopogaphyStructure of Earth

Surface Water

Rock CycleSediment, Igneous,Meta./

Oceans

Physics - Ocean,Atmospheric Circulation

HydrologicCycle

PopulationsCommunitiesEcosystems

PhysicalEnvironment

Ground Water

Surface Processes

IAccumulated/ Produced Materials

Plate Tectonics

Geological History

Fig. 4. Web-like/interrelated format for subject matter structure.(American Preservice Teacher)

describe the appearance of subject matter representa-tions were a matter of convenience. As opposed to thedescriptive labels, of more significance are the cleardistinctions among the representations. It is im-portant to again note that the preservice teacherswere allowed to represent subject matter in any waythat they felt best depicted their understanding.Although all of the teachers were familiar with con-cept mapping, none chose to follow the conceptmapping approach of describing the meaning ofconnecting arrows and/or connecting lines. Whenasked for the meaning of these connections during

interviews, the preservice teachers (both Taiwaneseand American) consistently described the meaning asa "connection or relationship" with no further elabo-ration. Rather, the responses tended to focus on thegeneral organizational pattern of the representation,with the nature of the pattern having implications forthe type of relationship depicted by connecting arrowsand/or lines (e.g., in a hierarchical depiction, arrowsor lines extending from a superordinate idea wereintended to denote inclusion of those subordinateconcepts/ideas listed below).

Pedagogical structures tended to be organized asweb-like/interrelated representations of concerns,knowledge, and/or activities performed (Fig. 5), withstudents conspicuously absent as a primary focus inthe representations of the American group (but as aclear focal point in the representations of the Taiwan-ese group) or as discrete "listings" of teacher-focusedresponsibilities and activities (Fig. 6). Again, descrip-tive labels were for convenience and should not dis-tract from the clear visual and substantive distinctionsamong representations. As with the subject matterrepresentations, connecting lines and arrows weredescribed as simply denoting relationships, with theoverall organizational pattern of the representationproviding further clarification of the nature of the

CLASSROOM

AllExperience

CONTENT

Persc nality

Philosophy

Motivation

apaoility

Respect

Connections of Rapport

Fig. 5. Web-like/interrelated format for pedagogical structure. (Tai-wanese Preservice Teacher)

115

126

OutsideCurriculum

Evaluation

Methodsof Teaching

Funding


Personal

_.4___Number ofStudents

Facilities

Administration

Fig. 6. Discrete responsibilities/activities format for pedagogicalstructure. (American Preservice Teacher)

relationship.

2. What is the Source(s) of these KnowledgeStructures?

When asked about the source of their subjectmatter structures, many student teachers admitted, asmight be expected, that the portrayed elements andorganizational scheme came from college courses andthat the representations were only tentatively delin-eated without any conscious rationale. For example,comments along the lines of the following were com-mon:

I just put down the things we learned and did in my physicsclasses. Everything I know about physics comes from myteachers and books I have read. (Taiwanese Preservice Teacher)

How I view earth science is what I learned in school. Probably

from all of my schooling, but mostly from what I had in college.(American Preservice Teacher)

These findings suggest that preservice science teach-ers (regardless of nationality) are not being presentedwith an overt or covert structure (or global conceptualframework) of subject matter (or at least one that isrecognized) as part of their content preparation. Thereader is also reminded that both groups of preserviceteachers possessed subject matter knowledge back-grounds exceeding that included as part of an under-graduate degree in the U.S. Consequently, the lack ofany recognizable subject matter structure does notappear to be unique to those with only undergraduatelevel preparation in subject matter. The reader isreminded that neither subject matter nor pedago-

gical structures were stressed within the teacherpreparation programs of the Taiwanese or Americanpreservice teachers. Consequently, the representa-tions given indicate the individual knowledge struc-tures formulated by the preservice teachers as theyproceeded through subject matter and pedagogycourses.

When asked about the source of their pedagogyknowledge structures, the preservice teachers uniformlyreferred to introductory education courses and per-sonal experiences as a student:

I have been a student and a student in education courses. Where

else could I better learn about teaching? (American PreserviceTeacher)

My science education courses at the university have taught mewhat I need to know to be a good teacher. (Taiwanese Preservice

Teacher)

When asked if they had ever thought about theirsubject matter specialty, or pedagogy in the mannerrequested by the questionnaire, only one of theAmerican preservice teachers, and none of the Tai-wanese, admitted having previously thought about hissubject matter in this manner. No individuals of eithergroup admitted having done so for their knowledgeof pedagogy. This finding is quite consistent with thelack of explicit attention to knowledge structures inboth the Taiwanese and American teacher preparationprograms. Contrary to the findings of previous re-search which has relied on card sorting tasks and otherpossibly restrictive assessment procedures (Baxter etal., 1985; Hashweh, 1986; Hauslein et al., 1992; Hozet al., 1990; Wilson, 1989), but consistent with re-search using more open-ended assessments (Ledermanet al., 1994), the preservice teachers appeared to possessno coherent, as typically defined by curriculum reformmovements (Kennedy, 1990), or carefully consideredstructure for their subject matter. Furthermore, thetopics, themes, and processes, etc. used in the repre-sentations of this group of preservice teachers exhib-ited little resemblance to the a priori elements/topicsused in previous investigations. Perhaps the moredirected approaches (e.g., concept maps, card sortingtasks, semantic maps) used in previous investigationsof subject matter structures served to create the result-ing structures (with respect to both content and orga-nization) and did riot necessarily provide an objectiveassessment. With respect to pedagogy, the results ofthis investigation were consistent with those obtainedin previous investigations (Lederman et al., 1994;Morine-Dershimer, 1989).

116

C)-r7;


3. Are these Knowledge Structures Stable Fund ngDuring Student Teaching?

Overall, virtually no changes were noted in thesubject matter representations of either group. Al-though changes were clearly noted in the pedagogicalknowledge structures of the American group, the rep-resentations of the Taiwanese group remained quitestable. The lack of change in subject matter concep-tions of either group, despite the planning and imple-mentation of lessons during student teaching, is afinding that contradicts an emerging and consistentbody of literature (e.g., Hauslein et al., 1992; Gess-Newsome & Lederman, 1993; Lederman et al., 1994).When asked to discuss their conceptions of subjectmatter during the interview, typical responses clearlyreinforced the impressions provided by the writtenrepresentations.

How I view biology really hasn't changed much. I pretty muchthink of things the same as I said before. (American PreserviceTeacher)

I am probably a bit more frustrated than before I taught. ButI still view the teaching of physics the same. (TaiwanesePreservice Teacher)

The interviews definitively indicated that thesepreservice teachers had not altered their views towardthe subject matter in response to the use of the subjectmatter in the context of teaching. Of particular interesthere is the Taiwanese preservice teacher's referenceto "the teaching of physics." The representation andrelated discussion was intended to be limited to thesubject matter. However, as mentioned before, theTaiwanese preservice teachers consistently exhibiteda subconscious (and often conscious) difficulty orunwillingness to consider subject matter as separatefrom the teaching of the subject matter. The signifi-cance of this clear difference from the Americanpreservice teachers will be addressed in the Implica-tions section of this paper.

Pedagogical representations became increasinglymore complex for the American preservice teachers.A proliferation of student-focused components (e.g.,motivation, learning styles, relevance, etc.) as well asadditional teacher roles (e.g., friend, counselor) andresponsibilities were clearly evident. Of most signifi-cance was a general shift away from linear represen-tations of pedagogical knowledge to more web-likeframeworks which placed the students and their con-cerns at the center (Figs. 7 and 8). For example, theindividual who created Fig. 7 had initially created

OutsideCurriculum

Personal Number ofStudents

..41 Administration

Evaluatioh

Facilities

Feedback

Methods ofTeaching

Learning Styles

Personal -44DevelopmentInterests

ScheduleParental Support

or Lack of Peers

Fig. 7. Web-like/interrelated format for pedagogical structure(American Preservice Teacher)

Teacher's Personali

Teach ng Techniques

Teacher's Pedagogy

vironment

Student's Motivation/1nterest

Subject---- Curriculum

Student's Abilities/Interest

Student's Age/Maturity

Fig. 8. Web-like/interrelated format for pedagogical structure.(American Preservice Teacher)

Fig. 6.In general, the pedagogical representations of the

Taiwanese preservice teachers remained the same as

117

.228


before student teaching. The representations of theAmerican preservice teachers became more similar tothe initial and stable representations of their Taiwan-ese counterparts. The changes in representations ofpedagogy by the American group appeared to be in-fluenced by the planning and implementation of actuallessons. A common explanation for the change in theAmerican preservice teachers' pedagogical structuresis illustrated by the following comments:

The students demand your attention. You couldn't ignore themif you wanted to. (American Preservice Teacher)

You can talk about the importance of the students all you want.You may even believe you have a focus on students' needs.But it's all abstract until you're actually face to face with 30of them. (American Preservice Teacher)

In short, the American preservice teachers reinforcedone of the commonly voiced shortcomings of campus-based teacher preparation courses. It is interesting tonote, however, that the Taiwanese preservice teachersinitially placed students as a focal point in theirpedagogical structures and continued to do so through-out the duration of the investigation.

4. What Is the Relationship between theseKnowledge Structures and How Do TheyRelate to the Act of Teaching?

During the interview the preservice teachers wereasked to discuss and relate to each other the set of fourquestionnaires (two subject matter and two pedagogyquestionnaries). Whenever overlaps or similaritiesbetween the two types of structures were noted, thesubjects were asked if they could be combined intoone diagram or whether a combined depiction wouldbe more accurate. The American preservice teachersuniformly responded negatively:

It makes more sense to me to keep the two separate. After all,knowledge of subject matter and how you teach subject matterare two different things. (American Preservice Teacher)

They're different things. When I teach I need to know mysubject matter, but my knowledge of teaching tells me how topresent what I know. (American Preservice Teacher)

On the other hand, the Taiwanese students were clearlyless willing to distinguish between subject matter andpedagogy. As noted previously, they continued tointegrate pedagogy into their conceptions"of subjectmatter:

There is much overlap, of course. I have learned sciencebecause I wanted to become a science teacher. I learned science

always with a view of teaching it. (Taiwanese Preservice

Teacher)

The American preservice teachers clearly perceivedpedagogy and subject matter knowledge as separateentities which were applied in an integrated mannerduring teaching, While the Taiwanese preservice teach-ers perceived the two in a much more integrated manner.During the interview, individuals were providedwith a hypothetical teaching situation in which stu-dents are unable to understand a particular aspect ofsubject matter. When asked about what their responsewould be, the two groups of preservice teachers de-scribed their decision making process quite differ-ently:

If students do not understand, I must find where the confusionis. To do this involves my knowledge of physics teaching. Itis not a problem of subject matter or pedagogy. (TaiwanesePreservice Teacher)

High school students do not really know much biology. If theydo not understand something, I must rely primarily on myknowledge of teaching. My knowledge of subject matter isimportant, because it gives me alternative examples to use, butit is my knowledge of teaching that lets me choose the correctsolution to the problem. (American Preservice Teacher)

In short, even when presented with a hypotheticalclassroom situation/problem, the American grouptended to conceptualize the influence of subject matterknowledge and pedagogy separately, while their Tai-wanese counterparts exhibited a more integrated ap-proach to the two knowledge domains.

As previously mentioned, neither group ofpreservice teachers altered their conceptualizations ofsubject matter knowledge in response to their exposureto public school students and the planning and imple-mentation of science lessons. This finding does notsupport prior suggestions (Hauslein & Good, 1989;HaUslein et al., 1992) that it may be impossible to viewsubject matter as separate from the manner in whichit 'is, or will be, used. The act of teaching and/orthinking about how one will teach subject matter didnot appear to have a significant influence on the waysubject matter was conceptualized among these twogroups of preservice teachers.

The pedagogical structures of the American groupwere seen to shift toward a focus on student concernsfollowing the student teaching experience. This find-ing is consistent with assertions made by Lederman

118

1,29


& Gess-Newsome (1991) concerning the shift inconcerns of preservice science teachers toward stu-dents as soon as they begin to teach lessons in actualfield settings.

When specifically asked if their stated subjectmatter and pedagogical knowledge structures wereevidenced in their teaching, both groups of preserviceteachers were confident that each type of knowledge(i.e., subject matter and pedagogy) was reflected inhow and what they taught:

Of course, at least I hope, my. teaching is based on what I know

and think. (Taiwanese Preservice Teacher)

I teach chemistry the way I view chemistry and I teach in away that reflects my philosophy of teaching. I believe modeling

to be very important in chemistry and it is continually stressed.I believe students learn best if they are actively involved andso I organize my class in that way. (American Preservice

Teacher)

These results are consistent with a large body of lit-erature on the relation of subject matter structures andteaching (e.g., Baxter et al., 1985; Hashweh, 1986)and contradicts recent research (Gess-Newsome &Lederman, 1993; Hollingsworth, 1989) which indi-cates that preservice teachers are too overwhelmed byday-to-day instructional responsibilities to adequatelyand consciously incorporate integrated subject matterstructures into daily instruction. The present resultsconcerning the translation of subject matter and peda-gogy knowledge structures into classroom practicemust, however, be interpreted with extreme caution.The discrepancies between teachers' self-reports andactual classroom practices have been well documented.Additional research of this nature that includes actualclassroom observations should be pursued.

IV. Implications for ScienceEducation

Prior to any elaboration of the implications ofthis investigation, the reader is reminded that only oneteacher preparation program from the U.S. and Taiwanwere investigated. Consequently, it would be inap-propriate to generalize differences between the twosamples to obtain definitive differences between Tai-wanese and American teacher education. Neverthe-less, the two teacher preparation programs investi-gated were significantly different in approach andseveral of the noted differences in findings are seem-ingly related to programmatic differences.

It does not appear that these preservice science

teachers, regardless of nationality, possess "well-formed" or highly integrated subject matter or peda-gogical knowledge structures. Consistent with pre-vious research (Gess-Newsome & Lederman, 1993;Hauslein et al., 1992; Lederman et al., 1994), thesubject matter knowledge structures that do exist arelargely the result of college course work and are oftenfragmented and disjointed with little evidence ofcoherent themes. Consequently, the currently popularpolicy of requiring stronger subject matter backgroundsfor preservice and inservice teachers, as a means ofresolving the myriad of concerns about the quality ofscience instruction, may not be an effective approach.Such an approach, as seen with this group of preserviceteachers; would most likely not lead to the develop-ment of the highly prized integrated subject matterconceptions advocated by prominent science educa-tion reform movements (A.A.A.S., 1989; NRC, 1996;NSTA, 1993). Furthermore, the preservice teachersinvestigated in similar studies (Gess-Newsome &Lederman, 1993; Lederman et al., 1994) possessed farless extensive backgrounds in science but developedmore integrated subject matter structures in responseto the planning and implementation of instruction. Inaddition, it is important to note that the Americanpreservice teachers studied in the present investigationcompleted the same professional teacher educationcoursework (e.g., methods, microteaching, practicum,etc.) as those in the studies by Gess-Newsome &Lederman (1993) and Lederman et al. (1994), with theonly difference being the extent of subject matterbackground (i.e., degrees attained and course credithours).

It is possible that the more extensive academicbackgrounds of both the Taiwanese and Americanpreservice teachers (which is consistent with currentteacher preparation reforms) may result in the devel-opment of more firmly entrenched and inflexibleconceptions of the subject matter. Consequently,although few would argue with the desirability ofscience teachers with extensive academic backgrounds,it might be that present approaches to college-levelscience instruction promote the development of rela-tively inflexible cognitive structures which are at oddswith the integrated framework required for the imple-mentation of currently advocated curriculum reforms.Although acquiring a relatively static view of one'ssubject matter as a consequence of a more extensiveacademic background is a problem in need of solution,the situation is further exacerbated if the nature of thestructure is less tban desirable. Since any significantreform in the instructional approach that currentlytypifies college science teaching seems unlikely, re-

119

130


sponsibility for stimulating students for reflecting ontheir subject matter (in an effort to promote the de-velopment of more integrated and flexible knowledgestructures) seems to be most appropriately placed withinthe domain of the science educator. It is possible thatrepeated opportunities for reflecting on one' s subjectmatter, as it is being learned, may be sufficient toprovide preservice teachers with a coherent schemafor their subject matter and allow them to integratemore of the information presented in their sciencecourses. Certainly, the possible benefits to be derivedfrom increased reflection upon subject matter withinscience education courses is an area needing furtherresearch.

The inability of the American preservice teachersto present a coherent conceptualization of pedagogyprior to student teaching is not surprising. As priorresearch has indicated (Lederman & Gess-Newsome,1991), a well formed pedagogical knowledge structureshould not be expected without actual experience with"real" secondary students. Other than simply increas-ing the length of field experiences (as many teachereducation programs are already doing), it may benecessary to provide increased opportunities forpreservice teachers to conduct systematic classroomobservations (Good & Brophy, 1991) and reflect uponinstructional sequences.

The American and Taiwanese preservice teach-ers clearly conceptualized pedagogy, and the relation-ship of pedagogy and subject matter, differently.Initially, the Ameican group gave students only pe-ripheral attention, while the Taiwanese group tookstudents as a focal point (a view the American groupadopted following student teaching). Furthermore, theTaiwanese group consistently exhibited difficulty inconceptualizing subject matter as separate from theteaching of the subject matter, while the Americangroup clearly preferred to keep subject matter knowl-edge and knowledge of pedagogy distinct. It appearsthat distinct differences between the professionalteacher education programs undergone by the twogroups of preservice teachers may be responsible forthe noted differences in conceptions of pedagogy andsubject matter. In particular, the American group wascompleting a MAT program. This program requiresa B.S. degree (or beyond) in one' s subject matterspecialty for admission. The program does includean additional nine graduate hours in the subject matter,but is primarily focused on science pedagogy withsubject matter knowledge assumed to have been ac-quired prior to entrance. Virtually all of the Americanpreservice teachers had decided to become secondaryschool science teachers during their senior year in

college or following graduation. The Taiwanese group,on the other hand, had been educated in a system muchlike the historic American Normal School. Thesestudents had decided to become teachers when theywere completing their high school education, and spentmuch effort preparing for entrance examinations thatwould enable them to attend a college dedicated to thepreservice education of teachers. Consequently, theTaiwanese preservice teachers received their subjectmatter background concurrently with their teacherpreparation, and the subject matter was typicallypresented from the perspective of eventually havingto teach it to others. Although this was not an ex-perimental investigation, it does not take much to seewhy the Taiwanese students initially focused on stu-dents when conceptualizing pedagogy (while theAmerican group did not) and experienced much dif-ficulty in representing subject matter apart from itsteaching (while the American group did not). Perhapsthe U.S. may want to reconsider its current trendtoward graduate level teacher preparation, with sub-ject matter knowledge "front loaded," and focus spe-cific attention on the conceptions of pedagogy andsubject matter which are apparently needed to imple-ment the currently popular reforms.

Keeping in mind the caution necessitated by thefact that classroom observations of these preserviceteachers were not performed, the self-reported influ-ence of preservice teachers' subject matter structureson classroom practice is consistent with much of theresearch on pedagogical content knowledge(Gudmunsdottir & Shulman, 1987; Hashweh, 1986;Shulman, 1987). However, the conclusions of Ameri-can preservice teachers concerning the separate appli-cation of subject matter knowledge and pedagogicalknowledge to instructional decisions are at odds withcurrent thinking related to pedagogical content knowl-edge. Again, the Taiwanese preservice teachers weremore integrated in their application of pedagogy andsubject matter to classroom decisions. It may be thatteacher preparation in a manner similar to the NormalUniversity is more consistent with promoting peda-gogical content knowledge than current approaches toachieving this end. Again, however, it must be notedthat the Normal University approach to teacher edu-cation experienced by the Taiwanese group did notappear to alleviate the problem of discrete and frag-mented conceptions of subject matter. This findingserves as a clear reminder that significant reform inscience teacher education will necessarily involve fullcooperation between subject matter specialists andscience teacher educators.

The apparent contradiction with those of Gess-

120


Newsome & Lederman (1993) of the findings relatedto subject matter structures is particularly intriguing.The subjects in that study included global, integrative(and arguably abstract) curriculum themes such as thenature of science and science-technology-society in-teractions in their subject matter structures. Suchthemes were virtually absent from the representationsof both the American and Taiwanese preservice teach-ers, making their knowledge structures relatively simplein comparison. Consequently, it is quite possible thatthe ease with which a subject matter structure affectsclassroom practice (if at all) is as much a function ofthe relative complexity of the knowledge structure asit is of curriculum constraints, administrative policies,management concerns, etc. Indeed, the findings of thisinvestigation concerning the complexity of a teacher' ssubject matter structure are supported by the recentlyreported investigation of Lederman et al. (1994).However, given that the data concerning translationof subject matter conceptions/structures into class-room practice was self-reported in nature, additionalresearch that includes direct classroom observationsshould be performed in order to focus on the relation-ship between knowledge structure complexity andclassroom practice. The possible importance of thecomplexity of one's knowledge structure is especiallyproblematic since many of the new reforms in scienceeducation seem to depend on the incorporation ofhighly integrative themes such as the nature of scienceand science-technology-society interactions. If suchhighly complex and integrated subject matter struc-tures are so difficult to translate into classroom prac-tice, our expectations with respect to the ability ofbeginning and novice teachers to implement curricu-lum reform may have to be drastically re,considered.

References

American Association for the Advancement of SOience (AAAS)(1989), Science for all Americans. Project 2061. Washington,D.C.: AAAS.

Baxter, J. A., Richert A., & Saylor, C. (1985). Learning to teachbiology: A consideration of content and process. Knowledgegrowth in a profession publication series. Stanford, CA:Stanford University, School of Education.

Bogdan, R. C., & Biklen, S. K. (1992). Qualitative research foreducation: An introduction to theory and practice. Boston, MA:Allyn and Bacon, Inc.

Brophy, J., & Good, T. L. (1986). Teacher behavior and studentachievement. In: M.C. Wittrock (Ed.), Handbook of researchon teaching (3rd ed., pp. 328-375). New York, NY: Macmillan.

Carnegie Forum on Education and the Economy, Task Force onTeaching as a Profession (1986). A nation prepared: Teachersfor the 21st century. New York, NY: Carnegie Forum onEducation and the Economy, Task Force on Teaching as aProfession.

Eisner, E. W. (1991). The enlightened eye. New York, NY:Macmillan.

Gess-Newsome, J., & Lederman, N. G. (1993). Preservice biologyteachers' knowledge structures as a function of professionalteacher education: A year-long assessment. Science Education,77(1), 25-45.

Good, T. L., & Brophy, J. E. (1991). Looking in classrooms. NewYork, NY: Harper-Collins Publishers, Inc.

Gudmunsdottir, S., & Shulman, L. S. (1987). Pedagogical contentknowledge in social studies. Scandinavian Journal of Educa-tional Research, 31, 59-70.

Hashweh, M. Z. (1986). Effects of subject-matter knowledge inthe teaching of biology and physics. Paper presented at theannual meeting of the American Educational Research Asso-ciation, San Francisco, CA. (ERIC Document ReproductionService No. ED 275 502).

Hollingsworth, S. (1989). Prior beliefs and cognitive change inlearning to teach. American Educational Research Journal, 26,160-189.

Hauslein, P. L. & Good, R. G. (1989). Biology content cognitivestructures of biology majors, biology teachers, and scientists.Paper presented at the annual meeting of the National Asso-ciation for Research in Science Teaching, San Francisco, CA.

Hauslein, P. L., Good, R. G., & Cummins, C. L. (1992). Biologycontent cognitive structure: From science student to scienceteacher. Journal of Research in Science Teaching, 29(9), 939-964.

Holmes Group (1986). Tomorrow's teachers: A report of theHolmes Group. East Lansing, MI: Holmes Group.

Hoz, R., Tomer, Y., & Tamir, P. (1990). The relations betweendisciplinary and pedagogical knowledge and the length ofteaching experience of biology and geography teachers. Jour-nal of Research in Science Teaching, 27(10), 973-985.

Kennedy, M. M. (1990). Trends a'nd issues in: Teachers' subjectmatter knowledge. Washington, D.C.: ERIC Clearinghouse onTeacher Education and the American Association of Collegesfor Teacher Education.

Lederman, N. G. (1992). Students' and teachers' conceptions ofthe nature of science: A review of the research. Journal ofResearch in Science Teaching, 29(4), 331-359.

Lederman, N. G., & Gess-Newsome, J. (1991). Metamorphosis,adaptation, or evolution?: Preservice science teachers' con-cerns, and perceptions of teaching and planning. Science Edu-cation, 75(4), 443-456.

Lederman, N. G., Gess-Newsome, J., & Latz, M. (1994). The natureand development of preservice science teachers' conceptionsof subject matter and pedagogy. Journal of Research in ScienceTeaching, 31(2), 129-146.

Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. BeverlyHills, CA: Sage Publications.

Miles, M. B., & Huberman, A. M. (1984). Qualitative data analysis:A sourcebook of new methods. Beverly Hills, CA: Sage Pub-lications.

Morine-Dershimer, M. (1989). Preservice teachers' conceptions ofcontent and pedagogy: Measuring growth in reflective, peda-gogical decision-making. Journal of Teacher Education, 40(5),46-52.

National Research Council (1996). National science educationstandards. Washington, D.C.: National Academy press.

National Science Teachers Association (1993). Scope, sequence,and coordination of secondary school science: The contentcore. Washington, D.C.: National Science Teachers Associa-tion.

121

132


Shulman, L. S. (1986). Those who understand: Knowledge growthin teaching. Educational Researcher, 15(2), 4-14.

Shulman, L. S. (1987). Knowledge and teaching: Foundations ofthe new reform. Harvard Educational Review, 57(1), 1-22.

West, L. H. T., Fensham, P. J., & Garrard, J. E. (1985). Describingthe cognitive structures of learners following instruction inchemistry. In: L.H.T. West & A.L. Pines (Eds.), Cognitivestructure and conceptual change (pp. 29-49). Orlando, FL:Academic Press.

West, L. H. T., & Pines, A. L. (1985). Cognitive structure andconceptual change. Orlando, FL: Academic Press.

White, R. T: (1985). Interview protocols and dimensions of cog-nitive structure. In: L.H.T. West & A.L. Pines (Eds.), Cognitivestructure and conceptual change (pp. 51-59). Orlando, FL:Academic Press.

White, R. T., & Tisher, R. P. (1986). Research on natural sciences.In: M.C. Wittrock (Ed.), Handbook of research on teaching (3rded., pp. 874-905). New York, NY: Macmillan.

Wilson, S. M. (1989). Understanding historical understanding: Ananalysis of the subject matter knowledge of teachers. Paperpresented at the annual meeting of the American EducationalResearch Association, San Francisco, CA.

IIPP4---4- tic RO tic Mi* PI lit Vflz

Norman G. Lederman* kti***

*Department of Science and Mathmatics EducationOregon State University

Corvallis, OR, U.S.A.

"MAZMWT.t*Itt

N

f-*

am**ftwam4omlamorammz*wwom,Aftv=-F-AllAnganrao,AAMZWV(*Zr13$ NA°A.-F=t1MMUMVO-Ofn-FnetOjAMMVOOX*MitMISI*EV*IIMIk'it4itill0**40MRVt*UMUASOUXNUtONW*.k*'11TVIOWV0-00

vAvs/alwia,mitommuftwomtk,gwiftworimmiozvmtamamftozmfmaciamol*4Jilmwmmutmo)-*.himmxmit,Imntkoomnomme,MAKI'MTnt't,019-X°31W8q141Z4Nent040MVCIRMIRS'Affi,ft*OMMORtoffilt,MIMMVOCUAIA,W*ROMANY:TAWAOOM,ffiRVNIft*OMW-ftk-hArtgoVE*40,1fAtk*OMMUMM'RPIAAIMtWIMYAAUAA.MSVanIER°M:REMMAOMMSARIZft*UMW4OMMM(eAKRWRWZAA'SigN$MWM130t042KA,RO-Mig*MAZAW5VA.-214UMR-AMMIVAItt-SAN,AKR9VeJMXIMM40TOI3R:raTOMM*2.**AWAWEAMft*OAIM*M40M*°

2133

a *41-11iFtEgMitt-I-P: CD) :

Proceedings of the National Science Council (Part D):

Mathematics, Science, and Technology Education

ik-Wq*A7E--1-44-t- fftcri Pi 14=7.1(-1-4*,,t. °

4-1--A _EA )014).1*i.) 253iffl PX) St7614 Y-j12904ffl t-14 'f`A' -*44R, AAA_

, ERIC -W-4,1-,-4c4.*

44-*Disciplines: science education, mathematics education, environmental education,

technology education, special education, health education, medical edu-

cation, and information education, etc.

Domain: learning and learner, curriculum and materials, instruction (including

computer assisted instruction), assessment and evaluation, history and

philosophy of science, teacher preparation and professional development,

etc.

tf34-t-034c--P14-4, #fs4-*-Pl#,A4-441.

4t44-z-444+1-tql*Akitik--=i--1-14,-IUOIA-4-1z

RT- -PJ 1- Ai &Ai

Ag-J 4*-

: *41,71-5t4-77,--alowl74-4171311- 4444+t : 02-7377594 Email: [email protected]

4-#4 : 02-7377248 [email protected]

13 4

Call for Paper

Proceedings of the National Science Council (Part D):


invites you to contribute Original or Review papers in the following:

Disciplines: science education, mathematics education, environmental education,



Domain: learning ahd learner, curriculum and materials, instruction (including

computer aSsisted instruction), assessment and evaluation, history and


etc.

Manuscripts should be written in English and sent in triplicate to:

Proceedings of the NSC, R.O.C., Part D

Room 1701, 17F, 106 Ho-Ping East Road, Sec. 2

Taipei 10636, Taiwan, Republic of China

TEL: (02) 737-7594

FAX: (02) 737-7248

Email: [email protected]

[email protected]

135

?ft .ft- A: ttNit.121

*:ffIE*Tlf4t A V= VC 11 A V /JNPI

14* tit* A A: James Bensen

Brmidji State University,Bemidji, MN, U.S.A.






A+ *M

Alan J. BishopMonash UniversityVictoria, Australia


AA.Vti





Alk NA110636RTAQM=g0061-4t17f1(02) 737-7594 737-7595 Email : [email protected]

14 : (02) 737-7248 [email protected]

tiPi&EPS11Jth

g RAJA

TrlItAKM.VE0*

161014**IR

(

: (02)732-1234

: (02)361-7511

AIMYXILM3843A :.(02)361-7511

Itilt:01MAMMMM 'CA : (02)3821394

WIL:APWM.23* : (04)2260330

(04)725-2792

Att Ara$1,1*M1413A cit : (07)332-4910

N:MIVA05En

AA,N*MkA'44AM-1-41MK:4AR-P-titEgIBIMN*M.frr*:i.

'PlA IIPM*0110611JIRMORP0101222-5at J=3 Fien0 -ft °2. R4M-17=1M1).aMilt -.HRICV-21K-Cr

riptimit : (02)737-7764

Pc. Jta *$1-1fal AN Zts =1' El

136

tar

a

et2ty024068860025

cMA 511. A+*#Mg

1E1

W*iii404M**tAM(OMQM)g 67

***MMME:t4MACRIWgAigAt 77

±WOMMX4P9VftOMMZAM86

Wit52.2±001tR±Oft/ROFAWg4 *V* 96

anNelanamA*vinkommxmomNorman G. Lederman Mist4 116

P.7r,(Si

IRJAKI VOV/4/DMI

a

at makes, Science, er0 lrechno off natio RAT9 aDorASeptem

CONTENTS

Empirical Review of Unidimensionality Measures for the Item Response Theory(Invited Review Paper)

Miao-Hsiang Ling

Development of A Grade Eight Taiwanese Physical Science Teacher's Pedagogical ContentKnowledge Development

Hsiao-Lin Tuan and Rong-Chen Kaou

Identification of the Essential Elements and Development of a Related Graphic Representationof Basic Concepts in Environmental Education in Taiwan

Ju Chou

123

135

155

A Beginning Biology Teacher's Professional DevelopmentA Case Study-Jong-Hsiang Yang and 1-Lin Wu 164

PERMISSION TO REPRODUCE AND

DISSEMINATE THIS MATERIAL HAS

BEEN GRANTED BY

L Pave-TO THE EDUCATIONAL RESOURCES

INFORMATION CENTER (ERIC)

1

U.S. DEPARTMENT OF EDUCATIONOffice of Educational Research and Improvement

UCATIONAL RESOURCES INFORMATIONCENTER (ERIC)

This document has been reproduced aseceived from the person or organizationoriginating it.



11111111

HERDDLOMED

CXD,59 ftti@H@GMOPE, MOWN

NEVEM OF OHM

Co

(..);

EDITING: Editorial Board of the National Science Council,Republic of China

EDITOR IN CHIEF: Chorng-Jee GuoEDITORS:Cheng-Hsia Wang Ovid J.L. Tzeng Tien-Ying LeeFou-Lai Lin Tze-Li Kang Jon-Chao HongYung-Hwa Cheng Yeong-Jing Cheng Tak-Wai Chan















EDITORIAL OFFICE: Room 1701, 106 Ho-Ping E. Rd., Sec. 2, Taipei 10636, Taiwan, R.O.C.Tel. +886 2 737-7595; Fax. +886 2 737-7248; E-mail: [email protected], [email protected]

Send subscription inquiries to: Room 1701, 106 Ho-Ping E. Rd., Sec. 2, Taipei 10636, Taiwan, R.O.C.Tel. +886 2 737-7595; Fax. +886 2 737-7248; E-mail: [email protected], [email protected]

SUBSCRIPTION RATES: NT$ 240.00 for one year in R.O.C.; US$12.0 for one year outside theR.O.C.







The list of contents is available at: http://www,4c.gov.tw/sciedu


Humanities and Social Sciences, and (D) Mathematics, Science, and Technology Education. Part D is devoted to the publication ofresearch papers in Mathematics, Science, and Technology Education, covering domain/content areas such as learning and learner,curriculum and materials, instruction (including computer assisted instruction), assessment and evaluation, history and philosophy ofscience, teacher Preparation and professional development, etc. Research papers of similar interests in other disciplines including:environmental, special, health, medical, and information education, etc. may also be submitted. Complete papers describing the resultsof original research that have not been and will not be published elsewhere will be considered for acceptance. Review papers providingsystematic evaluative reviews or interpretations of major substantive or methodological issues in certain disciplines or fields may also besubmitted. Manuscripts are welcome from any country, but must be written in English.


Proceedings of the NSC, R.O. C., Part DRoom 1701, 17F, 106 Ho-Ping East Road, Sec. 2Taipei 10636, Taiwan, Republic of ChinaTEL: (02) 737-7594FAX: (02) 737-7248Email: [email protected]

[email protected]



1. Title page should contain the following information:(1) Title


The department, institution and country of each author should be furnished. If there are several authors with different affiliations,authors should be paired with their respective institutes by means of superscript symbols after the author's names in this order*, **, ***, etc. Unless indicated, a single star refers to the author to whom correspondence should be addressed.



3. UnitThe metric or international standard system should be used for recording length, area, volume and weight. When used in conjunctionwith numbers, units should be abbreviated and punctuated (e.g. 3 ml, 5 gm, 7 cm).

4. FootnotesFootnotes should be kept to a minimum and indicated in the text by superscript numbers.



7. Figures and Graphs(1) There should be only one figure per sheet, and its position should be indicated in the text by typing Fig. on a separate line.(2) Figures and graphs should be of the smallest possible printed size that does not hinder clarity.(3) High-contrast computer-generated artwork, drawings in black ink on white paper or tracing paper, or high quality photographic




1 4 0



be indicated by typing Table on a separate line. Explanatory matter should be in footnotes, not as a part of the legend.(3) Short or abbreviated column heads should be used, and nonstandard abbreviations should be defined in the legend if not defined



10. References(1) References should be cited in the text by the last name of the author and year. If there is more than one reference to the same

author(s) and the same year of publication, the references should be differentiated by a, b, etc. e.g., (Jones, 1996; Jones &Smith, 1996; Jones, Smith, Chang, & Lee, 1992; Reed, 1994a, 1994b); Jones (1996) suggested that ; Jones & Smith (1996)reported that ; reported by Smith, Jones, & Chen (1995) [first citation]; as described by Smith et al. (1995)[subsequent citations].

(2) .References should be listed in alphabetical order according to the last name of author(s), patentee(s), or editor(s). Give completeinformation with the title of the paper, report, patent, etc. Do not number references.



Johnson-Laird, P. N., Byrne, M. J., & Tabossi, P. (1989). Reasoning by model: The case of multiple quantification.Psychological Review, 96, 658-673.

(ii) For books*Kuhn, T. S. (1970). The structure of scientific revolutions (2nd ed.). Chicago, IL: University of Chicago Press.Gentner, D., & Gentner, D. R. (1983). Flowing waters or teeming crowds: Mental models of electricity. In: D. Gentner


Brush, S. (1993). A case study of learning chemistry in a college physical science course developed for prospectiveelementary teachers. Unpublished doctoral dissertation, Florida State University, Tallahassee, FL.


meeting of the Third International Conference on Fluidization, Henniker, NH.Edmondson, K. M., & Novak, J. D. (1992). Toward an authentic understanding of subject matter. In: S. Hills (Ed.),


(v) For technical reports*Tiller, F. M., & Leu, W. F. (1984). Solid-liquid separationfor liquefied coal industries: Final report for RP-1411-1 (EPRI

AP-3599). Palo Alto, CA: Electric Power Research Institute.(vi) For patents*


III. PROOFS AND REPRINTSI. Authors will receive galley-proofs of their papers to check for errors. Proofs should be checked against the manuscript and return

as soon as possible. Keep alternations to a minimum, and if any, mark them on both galley and manuscript.2. Fifty reprints of each paper will be supplied to contributors free of charge. Additional copies may be supplied at the author's own

expense.

IV. PAGE' CHARGEThere are no page charge for publication in the Proceedings of the NSC.

V. COPYRIGHT

Upon acceptance of a manuscript for publication in the Proceedings of the NSC, the author will be asked to sign an agreementtransferring copyright to the Publisher, who reserves the copyright. No published material may be reproduced or published elsewherewithout the written permission of the Publisher and the author.

Revised April, 1997



Empirical Review of Unidimensionality Measures for theItem Response Theory

Miao-Hsiang Lin

Institute of Statistical ScienceAcademia Sinica


(Received June 11, 1997; Accepted, September 5,1997)

Abstract

This study reviewed the performance of various measures of unidimensionality by using simulateddata sets which were in compliance with the two-parameter normal ogive model conditions. Theinvestigated measures were based on homegeneity, factor analysis, item fit statistic, and Stout's DIMTESTmethodology. The simulated data sets varied in terms of the item-ability correlation range and test length.

Based on the overall results of this empirical review, the following suggestions are offered regardinguse of different unidimensionality measures. To arrive at a unidimensional and homogenous set of items,an index based on average inter-item correlation is better than one based on the coefficient alpha. Indetermining whether a data set satisfies an assumption of unidimensionality in terms of a single commonfactor model, the principal axis method performs much better than do the other methods in terms ofexamining the extracted eigenvalue pattern based on tetrachoric correlation matrices. To ensure thattest items hold essential local independence, Stout's DIMTEST factor procedure is preferred not onlyto the prior knowledge procedure, but also to other unidimensionality measures. To assess whether anindividual item fits the two-parameter normal ogive model, a two-stage procedure consisting of goodness-of -fit by chi-square statistic and root mean square standardized residual may be used when r120.

Key Words: homogeneity, item fit statistic, maximum likelihood-SMC analysis, principal componentanalysis, principal axis analysis, Stout's DIMTEST statistic, two-parameter normal ogivemodel, unidimensionality measures

I. Introduction

In testing theory, the assumption of unidimen-sionality refers to a set of test items sharing a singlelatent trait; it is this unobservable quantity whichaccounts for item performance correlations. Theassumption of unidimensionality ensures a meaningfulscore interpretation for measurement situations in whichcomparisons are based on test scores which consist ofitem score composites (McNemar, 1946). In addition,the assumption is considered a crucial component ofunidimensional item response theory models (Lord &Novick, 1968). Several researchers have shown thatwhen data sets violate this assumption, larger mea-surement errors occur which jeopardize meaningfulinferences concerning the relative standings of exam-inees in terms of certain traits (Wang, 1988; Hattie,1984). Stout (1987) summarized three reasons fordetermining data set unidimensionality : 1) to avoid

test contamination by another trait; 2) to indicate ifa test measures two or more traits (requiring a splitinto subtests for analysis and interpretation); and 3)to ensure in advance the appropriateness of applyingunidimensional item response models.

According to Hattie (1985), over 80 indices havebeen proposed for the purpose of determining unidi-mensionality-- all of them were developed from meth-ods based on homogeneity, principal component andfactor analyses, and fit statistics under one- or two-parameter logistic models. Example of indices cor-responding to these methods include coefficient alphaor average inter-item correlation (Cronbach, 1951;Novick & Lewis, 1967), first- to second- eigenvalueratio (Lord, 1980; Reckase, 1979), a chi-square teston the number of common factors (Bock & Lieberman,1970), and the sum of mean-squared or absolute re-siduals (McDonald, 1981; Rogers, 1984; Wright &Stone, 1979).

123

1:tEST COPY MARLA LE

M.H. Lin

Hattie (1985) compiled a list of unidimension-ality indices and summarized their problems asfollows: 1) alpha is not a monotonic function of uni-dimensionality whereas mean inter-item correlation isinfluenced by communality; 2) the eigenvaluemagnitude is dependent on which correlation type isanalyzed while, at the same time, the chi-square testis sensitive to sample size; and 3) indices of the sumof mean-squared or absolute residuals are obtainedusing an ad hoc approach to assessing item behavior,and sampling distributions are un-known.

Hattie (1984) had previously conducted a studyin which he generated item responses via a multivari-ate, three-parameter logistic latent trait model to assessthe adequacy of selected indices in determining uni-dimensionality. His results showed that, relative toindices based on other methods, an index of the sumof absolute residuals based on the two-parameter latenttrait model was more efficient in discriminating be-tween cases with one and more than one latent trait.

Stout (1987) took into consideration the inherentnature of test item performance being multiply deter-mined to explore whether a set of test items fit an'essential unidimensionality' --that is, the existenceof one (and only one) 'dominant' factor exertinginfluence on item responses. According to the prin-ciple of essential local independence, Stout developeda T index to assess essential unidimensionality as usedin the DIMTEST procedure (Stout, et al., 1992;Nandakumar & Stout, 1993). DIMTEST is a nonpara-metric test procedure in the sense that Stout's statisticT is derived by making no assumption about the shapeof either the underlying ability distribution or the itemresponse function.

The performance of Stout's statistic T has beeninvestigated extensively in a variety of test settingsusing both Monte Carlo-simulated and real data(Nandakumar, 1991, 1993, 1996; Nandakumar & Stout,1993; Stout, 1987; De Champlain & Gessaroli 1991;Hattie, et al., 1996). The studies yielded consistentresults showing that Stout's statistic T 1) stronglymaintained a desired Type I error rate; 2) was capableof discriminating between one and more than onedimension underlying test data; 3) obtained robustperformance over a variety of ability distributionshapes. These results support the dominance ofStout's statistic T over the other methods for assessingthe lack of fit of test item unidimensionality. How-ever, Lin (1997) Described the following three caveatsregarding conceptual and technical considerations.

First, the DIMTEST procedure, based on theprinciple of essential local independence, takes themathematical form of:

1 Cov(UUj10.0n(n

where n is test length (1)

Equation (1) states that, on average, the covariancesof any two item responses (Ui, LI') approach zero whenconditioned on the dominant ability vector O. In termsof (1), the DIMTEST procedure seeks to determinewhat is called "essential dimensionality"--the mini-mum dimensionality accounting for item responses. Incontrast, under the two-parameter normal ogive model(Lord & Novick, 1968) the mathematical meaning ofunidimensionality for a test consisting of n dichoto-mously-scored items is defined as:

g(u). _g(ule)f(e)dO for all u

[F.g(op [1_ pg(9)11 -"g}f(0)dO (2)

g = 1

where

u (u 1, u 2, ..., un) denotes any pattern ofn binary responses, and

g(u) denotes the marginal distribution of u

P (Oa'(0 -6 )

g g 1

V2g,---e2dt ,

- 00(3)

bg Pg(b

g)='0.50 item difficulty,

where ag

... NriiP (bg

) -= item discrimination, andg

,Pg is the derivative of Pg .

Equation (2) states that the marginal distribution g(u)of U=(U1, U2 . . . , Lin ) for a randomly chosen examineecan be expressed as the conditional distribution ofg(uI0) and a prior distribution f(0) for the randomvariable of latent ability O.

Compared to the unidimensional IRT model, theDIMTEST procedure is weaker in that it only looksfor the minimum dimensionality necessary to satisfyan assumption of essential local independence. TheDIMTEST procedure and IRT model stem from some-what different conceptualizations in terms of factorswhich influence item performances. For example, thelocal independence assumption under the unidimen-sional IRT model demands that covariance of eachitem pair be zero when conditioned on a single latent

124

143

Empirical Review of Unidimensionality Measure for the IRT

ability.Second, most examinations of DIMTEST perfor-

mance have used simulation data sets generated ac-cording to either a one-dimensional, three-parameterlogistic model for a Type I error rate study or a multi-dimensional, three-parameter logistic model for a powerstudy. These simulation studies generate zero-oneitem responses by comparing the probability magni-tude of correct /incorrect responses with a randomnumber between 0 and 1 generated from a uniformdistribution. It should be pointed out that such aprocedure fails to characterize the hierarchical Baysianstructure underlying the ITR dichotomous data re-sponses.

Third, Stout's statistic T=TL- TBfor assessing

\/2essential unidimensionality requires the selection ofan initial set of M test items in order to form anassessment subtest for computing the statistics 71 inT. Specifically, T's goodness of performance is sub-ject to the extent to which the chosen M items aredimensionally homogeneous. Stout (1987) has sug-gested use of either factor analysis or prior knowledgefor the selection of the M item test set. Most of thestudies cited above used the DIMTEST factor proce-dure for this purpose and found that the desired TypeI or power rate of Stout's statistic was obtainable. Theextent to which Stout's statistic T obtains a similarlevel of performance when the M items are selectedby means of prior knowledge is unknown.

In view of these three concerns, the intent ofthe present study was to examine the performance ofStout's statistic T using data sets complying with thesufficient conditions of the two-parameter normal ogivemodel in Eq. (3). In other words, the data sets mustsatisfy the hierarchical Baysian structure from a normaldistribution of ability, and the dimensionality of testitems must equal one. As such, this examination wasanalogous to a Type I error study. A secondary goalwas to examine the performance of the unidimensionalindex in terms of methods based on the factor analysis,homogeneity, and item fit statistics; this performancehas not been examined empirically using a data struc-ture consistent with the sufficient conditions describedfor Eq. (3).

II. MethodologyThe two-parameter normal ogive model of Eq.

(3) may be inferred from the regression assumptionsby imposing a continuous /latent variable Y; and aconstant yg for item g; from this, it is possible todetermine whether examinee a will answer item g

correctly via such relations as Prob(Ug=116)=Prob(Y;>ygle) and Prob(Ug=01(9)=Prob( 11y010). Lord(1968, 1980) pointed out that given that 0 is normallydistributed in the group tested -- that is, f(0)=T(0)in (2) -- the observed multivariate distribution U1, U2,

U,, for a set of n items can be arisen from somemultivariate normal distribution of (Y; , Y, ..Y,c) bydichotomizing them by (yi, 72, ...,y). The observedmultivariate distribution (U 1, U2, ..., U ) for any setof data is determined, by yg, as well as by theintercorrelations p',j of Yi and Ili (ij; j=1, 2, ..., n);and the resultant data is consistent with the twoparameter normal ogive model and the assumption ofunidimensionality. In this particular case, unidimen-sionality may be interpreted from the Spearman one-common factor model, since 0 acts as the commonfactor underlying test item performance. In addition,a parameter such as item factor loading 10; or itemdiscrimnation ag can be inferred from a factor analysisof tetrachoric correlations between item U, and itemU.

The above theoretical relationships provide theguidelines for generating zero-one data sets. For thesake of clarity, they are summarized in the diagrambelow (Fig. 1):

Hypothetical Level Observable Binary Data Set

0'1, Y2, n)(Yl, Y2, rg, Yn) Ug-data

n items n items(1, 2, ..., g, n) (1, 2, ..., g, n)

e203

(1' 11, Y ", 1,n)

(Yp, Y;2, Y4, Y;n)(Y31, Y32, Y3g, Y3n)

Ya2, '", Yag, , Y an)

(YN1, YN2, 1' Ng, ..,

(U11, U12, U 1 g, U In)(U21, U22, ..., U2g, U2n)(U31, U32, ..., U3g, ..., U3,1)

(Ual, Ua2, Uag, Uan)

(UNI, U N2, U Ng, ( Nn)

Y;

YZ

1/1, 1/2, Yr

product- moment

corr.

between Yi and Y5

UU,

estimate .

U

i/g.p;61+\11-pg'2E;

item factorloadingon 0where P; = correlation between Y and

biserial -corr. between Ug

and E, ={the unique random variable

g the unique part of item g

(11 U2 U n

tetrachoric

correlations

between U, and

and 0

Fig. 1. Data structure consistent with the sufficient conditions forthe two-parameter normal ogive model.

125

1 4 4

M.H. Lin

Readers are referred to Lin (1997) for a detaileddescription of the procedures used to generate Ug-dataof zero-one responses that comply with the two-pa-rameter normal ogive model. The following is asummary of the four major steps:

(1)N examinees are generated from a standardnormal distribution of O. A vector such as0=I0i, 02, , 0,,,dc0(0,1) is produced.

(2) Random variates of Yclea, 1/g lea,11 ea are generated for examinee a's responsesto the n items via the conditional normal dis-tribution of 1/g 119a having a mean of KI0=P;O

and a standard deviation ofNote that this step is completed for all N ex-aminees.

(3)The constant vector -11.=[yi, y2, ..., yg, yn] isobtained for the set of n items. This is possibledue to the capability of determining yg from lrgaccording to the relation:

g Prob (U =1)= 1 e 2 dt =4:1)( 7g).

(4) When 1 'agng, the corresponding binary quan-tity Uag=1; when Kgyg, Uag=0. Comparisons'are conducted relating each examinee's re-sponses to' each individual item.

In the present study, these four steps were fol-lowed in order to simulate five Ug-data sets-- that is,five matrices containing the zero-one responses of Nexaminees (ranging from 1000 to 2000) to n items(20, 30 and 40). For a test length of 20, three matriceswere generated by varying the range of pg' used inStep 2. For example, three sets of 20-p; were gen-erated from a uniform distribution with limitsfalling within intervals of 0.05-0.95, 0.30-0.90,and 0.50-0.90. When n=30 or 40, only one set ofp; was generated from a uniform distribution withlimits falling within an interval of 0.05-0.95. Theintent of this design was to determine whetherthe performance of Stout's statistic T and relatedmeasures was dependent on the test length, range of

or both.

Ill. Data Analyses and Results1. Methods Based on Stout's Statistic T

The analytic procedure of the DIMTEST statistic(Stout, et al., 1992) for testing the unidimensionalityof a set of n dichotomously scored items (H0:d=1) canbe diagrammed as follows (Fig. 2.):

a set of n items

idivided sequentiallyinto three subtexts under conditions

1st set of 2nd set ofM items M items n-2M items

ATI AT2 PT

used to partitionN examinees into

Subtests

with the same PT score

k is the number of correct PT scores, and k ranges from1 to n-1 with # of examinees >=20.

Fig. 2. A flow chart of the analytic procedure of the DIMTESTstatistic.

In the diagram, AT1 and AT2 represent Assess-ment Subtests 1 and 2, respectively, and PT stands forthe Partitioning Subtest. The conditions imposed onthe selection of two sets of M items are such that 1)the M items selected for AT1 are as unidimensionalas possible and dimensionally distinct from the sametest's n-M items ; 2) the set of M items selected forAT2 is dimensionally similar to the n-2M items , withdifficulty levels as similar as possible to those of theAT1 items; and 3) the ideal M is 11/114, where /120.The statistic TB, which is used in T (the DIMTESTstatistic), compensates for biases in TL. It is to rejectH0:d=1 when T>Za, where Za is the upper 100(1a)percentile for a standard normal distribution, a beingthe desired level of significance.

In accordance with the conditions imposed on theDIMTEST statistic T, this study used sets of M=5,M=7, and M=10 for five sets of tests with lengths of20, 30, and 40, respectively. In addition, all M itemsselected for ATI were either chosen independently bythe factor routine of the DIMTEST software (FA), orwere based on a prior knowledge procedure whichcapitalized on the fact that the range of p; was simu-lated. With respect to the latter method, two sets ofM items for AT1 were selected based on the degreeof homogeneity. For example, one set contained itemshaving the highest possible common-factor loadings(p;), representing the most homogeneous set (H1); theother set contained items with common-factor load-ings spread over the range of pg', thus representing theless homogeneous set (H2). In this manner, all three

126

5


sets of M items for ATI were selected for each of thefive data sets.

Table 1 summarizes the results of Stout's uni-dimensionality test of the five test data sets based ona sample size of 2000, including such statistics as TL,Tg, T and P-value, plus statistics associated with twoAT1 and AT2 subtests obtained under the H1, H2, andFA procedures (including the item number, means ofitem loadings / difficulties, and mean differencesbetween the two subtests). As shown in Table 1 , theH1, H2, and FA procedures yielded different sets ofM items for AT1; consequently, M items for AT2differed according to the selection conditions describedabove. In terms of the degree of homogeneity asmeasured by the mean of the common-factor loadingsof AT1 items, it was found that H1 yielded signifi-cantly larger mean values compared to those associ-ated with AT2 items, regardless of the test length. Incontrast, mean differences of loadings between AT1and AT2 under the H2 and FA procedures were neg-ligible, again regardless of the test length. Examiningthe mean differences in item difficulty between AT1and AT2, none were found among the three proce-dures.

Table 1 also shows that the calculated values ofT associated with the five test data sets resulting fromthe H1 procedure ranged from 1.80 to 9.42 -- all largerthan 4.05=1.65; the calculated values of T obtainedfrom the H2 and FA procedures were all less thanZ0.05=1.65, with the exception of one whose T valuefollowing the H2 procedure was measured at 1.71 fora test length of 20, with p; generated from U(0.05-0.95). That the HI procedure yielded values differentfrom those resulting from the H2 and FA proceduresindicated that acceptance of H0:d=1 is dependent onhow the M items for AT1 are selected. In addition,since the primary differences among the H1, H2, andFA procedures center on the degree of homogeneityassociated with AT1, it can be argued that Stout'sunidimensionality test might be invalid in cases whereAT1 items hold a high degree of homogeneity relativeto the same test's n-M items.

2. Method Based on Item Fit Statistic

Since Eq. (3) is the model expectation for an itemresponse, that is, P g(19)=E(U gle), the difference (re-sidual) between expected and observed data canbe used to examine item goodness-of -fit. If oneset of items fits the two-parameter normal ogivemodel, including the unidimensionality assumption,then their residual values should be small. On theother hand , a large residual value indicates failure

of the model in terms of goodness-of- fit. In accor-dance with this traditional argument, it is to com-pute the chi-square statistic for the mean squarederror associated with any item. Since all of thesimulated test data used in the present studywere based on test lengths greater than 20, item fitstatistics were computed with formulae from version1.22 of the BILOG-3 program (Mislevy & Bock,-1990).

Table 2 lists the likelihood ratio chi-squaresfor individual items in the five sets of tests withlengths of 20, 30, and 40; asterisks mark poor-fitting items for which significant levels were lessthan 0.01. The table shows the poor-fitting itemnumbers to be 3, 2, 6, 3, and 6 for the five tests withlengths of 20, 30, and 40. Since the simulatedtest data were in compliance with the two-parameternormal ogive model, these results indicated that thelikelihood ratio chi-square statistic invoked errormisclassification.

Since the chi-square statistic is sensitive totests with n?.20, a further examination of item fitwas conducted for five subtests composed of itemsflagged as poor-fitting. As all of these subtests werevery short; the fit test was based on the root meansquare of the standardized posterior residual. As shownin Table 2, none of the residuals were greater than 2.0,indicating no problem with the goodness-of-fit of theseitems. This result was true for each of the four short-ened subtests (note the lack of analysis for the subtesthaving only two items). It should also be noted thatthe root mean square standardized residual was sen-sitive to tests with rt20 as well.

3. Methods Based on Factor Analysis

Since the simulation data sets were consistentwith Spearman's one-common factor model, theauthor compared the performance of the unidimen-sionality measure based on the principal component(PC), principal axis (PA), and maximum likelihood(ML) methods (Harman, 1976). The primary distinc-tion among the three methods comes from the amountof variance analyzed--that is, the number placed inthe diagonal of the correlation matrix. Within thecontext of n items, the three models were addressedas follows:

127

Pt

{PC U1= a i1F1 + a i2F2

PA: Ui = a ilF, + a i2F2 + ... + aML: U1 =ai1F1-1-a12F2 +... + a 1,,,F,n + ... +cliQ1.

1 4 6

Tab

le 1

. Sta

tistic

s A

ssoc

iate

d w

ith: 1

) St

out's

DIM

TE

ST P

roce

dure

for

the

Five

Tes

t Dat

a Se

ts, a

nd 2

) A

TI

and

AT

2 Su

btes

tsO

btai

ned

Usi

ng th

e H

I, H

2, a

nd F

A P

roce

dure

s

AT

IA

T2

AT

I-A

T2

TL

TB

T P

-Val

ueM

item

num

ber

load

ing

diff

. .M

item

num

ber

load

ing

diff

. di

ff-L

oad

dif-

diff

.

HI

1 2

3 4

50.

860

0.49

811

10

16 1

8 6

0.67

60.

564

0.18

4-0

.066

8.86

6.32

1.80

0.04

*n=

20p;

:=U

(0.5

0-0.

90)N

=20

00H

21

5 10

15

200.

716

0.52

912

4 1

6 18

80.

688

0.58

50.

028

-0.0

565.

485.

99-0

.36

0.64

M.5

mea

n=0.

710

FA13

3 1

1 9

60.

752

0.43

814

8 2

0 2

120.

696

0.45

00.

056

-0.0

124.

213.

040.

830.

20H

I1

2 3

4 5

0.84

00.

498

12 1

0 16

18

60.

558

0.56

60.

282

-0.0

6810

.80

4.83

4.23

0.00

*n=

204

=-U

(0.3

0-0.

90)N

=20

00H

215

10

15 2

00.

624

0.52

911

4 1

6 18

60.

600

0.55

50.

024

-0.0

264.

124.

59-0

.33

0.63

M=

5m

ean=

0.61

5FA

19 1

7 10

18

80.

498

0.49

213

12

7 1

90.

678

0.56

0-0

.180

-0.0

680.

883.

83-2

.09

0.98

H I

1 2

3 4

50.

860

0.49

812

10

16 1

8 9

0.41

00.

561

0.45

0-0

.063

13.8

72.

188.

270.

00*

n=20

p;=

.U(0

.50-

0.90

)N=

2000

H2

15 1

0 15

20

0.53

60.

529

11 2

16

18 6

0.51

80.

555

0.01

8-0

.026

5.51

3.09

1.71

0.04

*oo

M=

5m

ean=

0.52

3FA

14 8

6 7

16

0.53

60.

354

17 9

20

15 2

0.42

80.

366

0.10

8-0

.012

3.25

1.88

0.97

0.17

HI

1 2

3 4

5 6

70.

860

0.45

810

29

16 1

5 21

24

90.

449

0.47

40.

411

-0.0

1615

.52

5.95

6.77

0.00

*n=

304=

U(0

.50-

0.95

)N=

2000

H2

1 5

9 13

17

21 2

50.

590

0.37

926

3 6

20

18 1

6 12

0.54

70.

389

0.04

3-0

.010

5.64

3.71

1.37

0.09

M=

7m

ean=

0.51

5FA

13 3

25

19 1

4 15

12

0.54

70.

512

11 2

8 10

7 2

4 6

90.

573

0.48

9-0

.026

0.02

33.

993.

710.

200.

42

H1

I 2

3 4

5 6

7 8

9 10

0.84

70.

438

25 2

4 20

38

37 4

0 33

29

280.

311

0.45

70.

536

-0.0

1917

.20

3.88

9.42

0.00

*n=

404.

=U

(0.5

0-0.

95)N

=20

00H

2 5

9 13

17

21 2

5 29

33

370.

536

0.46

026

3 6

36

38 7

40

2 10

35

0.50

60.

458

0.03

00.

002

6.74

5.18

1.10

0.14

M=

10m

ean=

0.50

2FA

719

5 2

28

37 8

4 3

4 39

0.52

90.

587

3 32

29 4

0 22

18

1 30

26

0.46

20.

572

0.06

70.

015

5.03

4.00

0.73

0.23

Hl:

AT

I M

item

s ha

ving

the

high

est v

alue

s of

com

mon

-fac

tor

load

ings

4.

H2:

AT

I M

item

s sp

read

ing

over

the

rang

e of

P.

HA

: AT

I M

item

s se

lect

ed b

y m

eans

of

fact

or a

nlay

sis

of D

IMT

EST

.*:

sig

nifi

cant

at 5

% le

vel.

147

148


Table 2. Chi-square Test of Individual Item Fits for the Five Test Lengths (20, 30, and 40 items), Including Root Mean Square StandardizedItem Residuals

items

/Ye-(0.50-0.90)n=20, N=1000

P'81,(0.30-0.90)n=20, N=1000

Pg' =(0.05-0.90)n=20, N=1000

p'g =(0.05-0.95)n=30, N=1000

fiVg =(0.05-0.95)n=40, N=1000

Chisq Residual* Chisq Chisq Residual Chisq Residual Chisq Residual

1 19.4* 0.287 30.9* 49.8* 0.481 17.6* 0.343 28.8* 0.772

2 21.7* 0.196 15.2 20.9* 0.497 27.8* 0.383 11.7

3 8.5 13.6 26.7* 0.495 11.3 11

4 10.4 9.6 14.6 7.2 22.9* 0.958

5 9.1 10.1 13.9 11.6 17.5* 0.340

6 12.6 12.4 18.9* 0.310 8.8 10.4

7 6.8 15.2 14.6 17.9 10.7

8 14.0 10 22.2* 0.611 2.9 18.8

9 17 12.9 13.4 6.2 16.6

10 5.1 3.9 13.2 7.5 23.7* 0.438

11 12.1 9.4 5 13.2 7.9

12 5.6 7.7 6.4 21.5* 0.953 7.2

13 5.7 8.6 13.7 10.9 8.8

14 6.6 4.3 4 10 5.9

15 25.4* 0.632 28* 26.3* 1.042 14.1 6.4

16 15.6 14.5 20 16.7 5.6

17 4 1.3 8.5 18 16.9* 1.005

18 5.2 6.7 10.4 4.1 5.4

19 4.1 7.7 8.5 8.3 5.4

20 16.7 9 7.2 5.5 5.5

21 5 10.8

22 7.8 14.5

23 10.4 8.7

24 7 13.3

25 5.3 13.5

26 4.4 4.4

27 9.1 14.3

28 10.3 7.8

29 15.5 . 6.7

30 17.7 26* 1.550

31 6.1

32 5.4

33 10.1

34 6.1

35 6

36 4.8

37 8

38 8.2

39 17.4

40 4.9

*: significant at 0.01 level.*: residual obtained by analyzing the shortened subtests consisting of items with an asterisk.

However, within a context of a unidimensional set ofn items, the number of common factors extracted fromeach of the three methods should equal one if themethod is indeed valid.

Table 3 summarizes the results of analyzing fivetetra-corr matrices based on the five Ug-data sets,including associated eigenvalues and their patternsand the number of factors extracted under individualcriterion sets according to each method. It is importantto mention here that rows 3, 8, 13, 18, and 23 in Table

3 contain parametets which exhibit only a single,positive, non-zero characteristic root. The informa-tion in each row serves as a baseline reference forcomparing the relative performance among the threefactor analysis methods--PC, PA (Iterated principalaxis), and ML-SMC (maximum likelihood with squaredmultiple correlation as a communality estimate).

As shown in the last column of Table 3, in termsof the number of retained factors, each of the threemethods failed to display a one-common factor pattern

129

140

Tab

le 3

. Eig

enva

lues

and

Num

bers

of

Fact

ors

Ext

ract

ed f

rom

Fiv

e T

etra

chor

ic M

atri

ces

Usi

ng T

hree

Fac

tor

Ana

lysi

s M

etho

ds f

or T

ests

with

20,

30,

and

40

Item

s

Ran

ge o

f

P;R

ow 3

U(0

.50-

0.90

)Pa

ra.

10.3

50

00

00

00

00

.0

1

PC10

.05

1.01

30.

883

<0.

883

2n=

20P-

Axi

s9.

610.

457

<0.

457

Red

uced

-Cor

r2

N=

1000

ML

-SM

C U

W9.

59

W26

.85

0.75

1<

0.75

1R

educ

ed-C

orr

>1

x2=

1523

.8*

Row

8U

(0.3

0-0.

90)

Para

.8.

160

00

00

00

00

01

PC8.

101.

188

1.14

1<

1.00

3n=

20P-

Axi

s7.

650.

638

<0.

638

Red

uced

-Cor

r3

,N

=10

00M

L-S

MC

UW

7.60

'.._f

....

W21

.42

0.70

5 <

0.70

5R

educ

ed-C

orr

>1

x2=

1322

.8*

t.,..)

Row

13

U(0

.05-

0.95

)Pa

ra.

6.81

00

00

00

00

00

10 I

PC6.

881.

267

1.18

31.

171.

07<

1.00

5n=

20P-

Axi

s6.

480.

626

<0.

626

Red

uced

-Cor

r5

N=

1000

ML

-SM

C U

W6.

42

W24

.00

0.64

9 <

0.64

9R

educ

ed-C

orr

>1

x2=

1345

.5*

Row

18

U(0

.05-

0.95

)Pa

ra.

9.98

00

00

00

00

00

1

PC10

.23

1.39

1.35

1.29

1.17

1.11

1.11

1.03

<1.

008

n=30

P-A

xis

9.87

0.99

<0.

994

Red

uced

-Cor

r8

N=

1000

ML

-SM

C U

W9.

78.

W32

.03

1.06

<1.

00R

educ

ed-C

orr

>1

x2=

3720

.7*

Row

23

U(0

.05-

0.95

)Pa

ra.

12.8

80

00

00

00

00

01

PC12

.92

1.27

1.23

1.20

1.15

1.13

1.10

1.08

1.05

1.04

<1.

0010

n=40

P-A

xis

12.5

20.

78<

0.78

Red

uced

-Cor

r10

N=

1000

ML

-SM

C U

W12

.45

W37

.29

1.20

<1.

00R

educ

ed-C

orr

>I

f=64

38.3

*

Met

hod

ofE

igen

valu

e pa

ttern

# of

fac

tors

Fact

or1s

t2n

d3r

d4t

h5t

h6t

h7t

h8t

h9t

h10

th11

-20,

-30

, -40

*: s

igni

fica

nt a

t 5%

leve

l.


because more than one significant factor was retained- regardless of the test length or range of the pg'. In

addition, it was observed that the number of retainedfactors increased as either the test length increasedor as the range of the pg' expanded. In terms of sizeof the first eigenvalue (column 5), the three methodsperformed equally well in that they all yielded valuesclose to the parameter values of the baseline reference-

again, regardless of the test length or the range offrg.

Concerning the eigenvalue patterns associatedwith each of the five tetra-corr matrices, the PA methodperformed much better than did its PC and ML coun-terparts-- that is, the eigenvalue pattern extracted viathe PA method was much more similar to the baselinereference. The magnitudes of all the other eigenvalueswere less than 1.00, regardless of the test length orrange of pg. It should be noted that the resultingeigenvalue patterns were extracted from reducedcorrelation matrices.

4. Methods Based on Homogeneity

Homogeneity measures from domain samplingtheory (Nunnally, 1978) place primary emphasis oninternal structure-- the relationships among items whichcomprise a test. The basic formula for deriving an

n pi;index of homogeneity is: p a =

1 + (n, where pu

is the is the average of inter-item correlations. Thisformula states that when n items tend to measure asingle trait (common core), the average inter-itemcorrelation is high, and the test is, therefore, morehomogeneous. Coefficient alpha (a), one of the mostimportant deductions from domain sampling theory,is another homogeneity measure-- one taking the form

a. n [1E7 ],

n 0-2

wheretrF = item score variance

= test score variance

The present study used both average inter-item cor-relation and coefficient alpha to examine the degreeto which a unidimensional item set reflects a homo-geneous item set.

Table 4 lists population and sample averages ofinter-item correlations and rr, and et for the fivetests. tr was calculated based on a tetra-corr matrixof n by n, and et was obtained from a U-data matrixof N by n.

As shown in Table 4, the population averages ofthe inter-item correlations (pu) were 0.503, 0.377,0.269, 0.263, and 0.250 for the five test sets. Acomparison of 7,./ values reveals that the inter-itemcorrelation averages were dependent on the pg rangebut not on the length. The same pattern held true forthe sample statistic, rura. This result can be explainedby the fact that the two highest values of either Tyor r7tm were associated with a test length of 20, fromwhich p; fell on the 0.50-0.90 and 0.30-0.90 interval,while the three lowest values were associated with testlengths of 20, 30, and 40-- all with pg' falling on the0.05-0.95 interval.

As shown in Table 4, the observed at values forthe five tests were 0.829, 0.784, 0.727, 0.801, and0.833. The first three values indicate that when thetest length was fixed at 20, et was a function of therange of pg . The last three values indicate that whenthe pg' range was fixed at an interval of 0.05-0.95, etwas a function of test length. et was, therefore, afunction of both test length and p; . In addition, theseresults show that a et value higher than 0.80 coulddepend on where pg fell along the interval of 0.50-0.90 or 0.05-0.95. This means that a unidimensionaldata set could contain less homogeneous items sincetheir common factor loadings could be less than 0.30.

5. Conclusions and Suggestions

This study reviewed the performance of unidi-mensionality measures under a variety of simulateddata sets which were in compliance with the sufficientconditions for the two-parameter normal ogive model.The study design was similar to that of a Type I errorstudy. The unidimensionality measures investigatedwere based on homogeneity, factor analysis, the itemfit statistic, and Stout's DIMTEST methodology. Thesimulated data sets varied in terms of the range of item-ability correlations (pg' ) and of test length.

The two unidimensionality measures based onhomogeneity theory (domain sampling theory) werethe average inter-item correlation (rra) and coeffi-cient alpha (et). Results showed that the average inter-item correlation was a function of the range of p;(the item-ability correlation) but was independent ofthe test length (n); coefficient alpha was found to bea function of both the range of pg and test length.However, given a fixed test length, both measuresincreased as pg' increased. Therefore, the averageinter-item correlation may be regarded as a betterunidimensionality measure compared to coefficientalpha. In addition, it was observed that a unidimen-sional set of test items did not necessarily represent

5

M.H. Lin

Table 4. Population Averages of Inter-item Correlations and Corresponding Sample Statistics rr° , Along with Alpha Values for theFive Tests

n=20p'g =U(0.50-0.90)

n=20p'g =U(0.30-0.90)

n=20p'g =U(0.05-0.95)

n=30p'g =U(0.05-0.95)

n=40ffg=U(0.05-0.95)

Para. iT.=0.503 fT0 =0.377 /7=0.263 iT=0.250Statis. e" =0 463

Yrr =0.344 rr=0.247 rr =0.255 rr =0.242

6c=0.829 eit=0.784 '6=0.727 5t=0.801 6c=0.833N=1000 N=1000 N=1000 N=I000 N=2000

a homogeneous set since the items' first factorloadings could be less than the criterion value of0.30.

Three factor analysis methods principal com-ponent, principal axis, and maximum likelihoodwere used to examine unidimensionality. Differenceswere found in the relative performance associated withthe three methods according to the type of correlationanalyzed and the parameter examined. When analyz-ing the tetra-corr matrix, the principal axis methodperformed much better than did the other two withrespect to holding an eigenvalue pattern similar to thatexpected under H0:d=1, regardless of the test lengthor range of pg' . In terms of the first eigenvalue, thethree methods performed equally well, with each oneproducing a value close to its parameter counterpart.In terms of the number of extracted factors, allthree methods failed to display a one-common factorpattern because more than one significant factor wasretained.

The two unidimensionality measures based onitem fit statistic were the likelihood ratio chi-squarestatistic and root mean square standardized residual.The performance of the likelihood ratio chi-squarestatistic in examining item goodness-of-fit wasfound to be sensitive to test length. For example,when the test length n?_20, the statistic tended tomistakenly flag items as poor-fitting in proportion totest length. However, when those poor-fitting itemsunderwent a second- stage examination via theroot mean square standardized residual, none of theitems showed a significant residual value -- indicatingan absence of problems in terms of goodness-of-fit.

It was found that the validity of Stout's DIMTEST

statistic (T -= TL 7TB) with respect to the type I errorV 2

depended on the method used to select AT1 M itemsfor the TL, statistic. The present study found that whenATI M items were chosen via the DIMTEST factorprocedure, the calculated T statistic values supported1-10:d=1; however, when the M items were selected by

means of a prior knowledge procedure, the calculatedT statistic values rejected 1-10:d=1. These results heldtrue regardless of the test length or range of p; . Fur-thermore, it was found that both selection proceduresyielded M item sets which differed in the degree ofhomogeneity. For example, M items selected by theDIMTEST factor analytical procedure were, on aver-age, less homogeneous than their counterparts selectedby the prior knowledge procedure. The results of thisstudy suggest that Stout's DIMTEST statistic mightbe invalid in situations where AT1 M items are ex-tremely homogeneous relative to the remaining n-Mtest items.

Based on the overall results of this empiricalreview, the following suggestions are offered regard-ing the use of unidimensionality measures based ondifferent approaches. To arrive at an unidimensionaland homogenous item set, the average inter-item cor-relation index functions better than coefficientalpha does. To determine whether a data set satisfiesan assumption of undimensionality in terms of aone-common factor model, principal axis methodol-ogy is the clear choice for examining extracted eigen-value patterns based on a tetrachoric correlationmatrix. If the primary concern is whether or not testitems hold essential local independence, Stout'sDIMTEST factor procedure is preferred not only tothe prior knowledge procedure, but also to theother unidimensionality measures. For assessingwhether individual items fit the two-parameter normalogive model, a two-stage goodness-of-fit procedurewhich entails both the chi-square statistic and rootmean square standardized residual may be used whenn..20. Here, it should be noted that all of the abovesuggestions are valid when the group of examineestested has approximately normal distribution of abil-ity.

Acknowledgments

This paper was excerpted from a project supported by a grantfrom the National Science Council, R.O.C. (NSC-85-251I-S-001-001).

132

VIC,1153


References

Cronbach, L. J. (1951). Coefficient alpha and the internal structureof tests. Psychometrika, 16, 297-334.

Bock, R. D., & Lieberman, M. (1970) Fitting a response modelfor n dichotomously scored items. Psychometrika, 35, 179-197.

De Champlain, A. F., & Gessaroli, M. E. (1991.) Assessing testdimensionality using an index based on nonlinear factor analy-sis. Paper presented at the Annual Meeting of the AmericanEducational Research Association, Chicago, IL.

Harman, H. H. (1976). Moddern factor analysis (3rd ed.). Chicago,IL:University of Chicago Press.

Hattie, J. (1984). An empirical study of various indices for de-termining unidimensionality. Multivariate Behavioral Research,19, 49-78.

Hattie, J. (1985). Methodology review: assessing unidimensionalityof tests and items. Applied Psychological Measurement, 9(2),

139-164.Hattie, J., Krakowski, K., Roger, H. J., & Swaminathan, H. (1996).

An assessment of Stoutds Index of essential Unidimensionality.Applied Psychological Measurement, 20(1), 1-14.

Kendall, M. G., & Stuart, A. (1961). The advanced theory ofstatistics, Vol. 2: inference and relationship. NY: Hafner.

Lin, M. H. (1997). Comparison of validity of methods in deter-mining test items unidimensionality (NSC Report No. 85-00 1-S-001-001). Taipei, Taiwan: National Science Cuncil, R.O. C.

Lord, F. M., & Novick, M.R. (1968). Statistical theories of mentaltest scores. Reading, MA: Addison-Wesley.

Lord, F. M. (1980). Applications of item response theory to prac-tical testing problems. Hillsdale, NJ: Lawrence Erbaum As-sociates.

McDonald, R. P. (1981). The dimensinality of tests and items.British Journal of Mathematical and Statistical Psycchology,34, 100-117.

McNemar, Q. (1946). Opinion-attitude methodology. Psychologi-cal Bulletin, 43, 289-374.

Mislevy, R. J., & Bock, R. D. (1990). Item analysis and test scoring

with binary logistic mode/s---BILOG 3. Mooresville, IN: Sci-entific Software, Inc.

Novick, M. R., & Lewis, C. (1967). Coeffficient alpha and thereliability of composite measurements. Psychometrika, 32, 1-13.

Nandakumar, R. (1991). Traditional dimensionality versus essen-tial dimensionality. Journal of Educational Measurement, 28,

99-117.Nandakumar, R. (1993). Assessing dimensionality of real data.

Applied Psychological Measurement, 17, 29-38.Nandakumar, R., & Yu, F. (1996). Empirical validation of DIMTEST

on nonnormal ability distributions. Journal of EducationalMeasurement, 33, 355-368.

Nandakumar, R., & Stout, W. (1993). Refinements of Stoutdsprocedure for assessing latent trait unidimenisonality. Jounalof Educational Statistics, 18, 41-68.

Nunnally, J. C. (1978). Psychometric Theory (2nd ed.). NY: McGraw-Hill Series in Psychology.

Reckase, M. D. (1979). Unifactor latent trait models applied tomultifactor tests: Results and implications. Journal of Educa-tional Statistics, 4, 207-230.

Roger, H. J. (1984). Fit statistics for latent trait models. Unpub-lished masterds thesis, University of New England, Armidale,Austrlia.

Stout, W. (1987). A nonparametric approach for assessing latenttrait dimensionality. Psychometrika, 52, 589-618.

Stout, W., Nandakumar, R., Junker, B., Chang, H. H., & Steidinger,D. (1992) . DIMTEST: A Fortran program for assessing dimen-sionality of binary item responses. Applied Psychological Mea-surement, 16, 236.

Wang, M. M. (1988). Measurement bias in the application of aunidimensional model to multidimensional item-response data.Paper presented at the annual meeting of the American Edu-cational Research Association, New Orleans LA.

Wright, B. D., & Stone, M. H. (1979). Best test design. Chicago,IL: Mesa Press.

133

154

M.H. Lin

gr,g IRT VgiWZM

4#A-

,11;EUfARea4,44-4PMg

tr4

*3(5uffimAfflmmilmwon,m-ta5c A*AunmffiowmatactmmgwatenTmIto*3amaapTmunmmwom-vm*=s4Agamx awm*vmmlatawAnttft-*5(muliami4mmTnnumomtmmmuNRmaRft,ffiffmomatact*Artmft*M,WAM4wM,stoin421KWAM,URAMAIM.

WAiiifMRN,*)ZgAlstAYMOCAItMMOT:OaffiV0140AtC-WhAVOtORAMWd,RMEMMTAMVEY:mphaig ;2)UWANAV#.1-AW*42JAAMaffiW,I ftffiaiiMIA*ftthatiait,00**MWMdMEIBTStillAMAXA4*-W4Att-t , pi Stout Eri4j DIMTEST rn 111*MMA02)WARAIV***8-;tatiMMUMti*=IS *MAXmNIlitnUM,SE tte;a4li,:20 ,5uku2 eavRalmanmitb-zAatatal

134

155


Development of A Grade Eight Taiwanese PhysicalScience Teacher's Pedagogical Content .

Knowledge Development

HSIAO-LIN TUAN* AND RONG-CHEN KAOU**

*Graduate Institute of Science EducationNational Changhua University of Education

Changhua, Taiwan, R.O.C.**Hsing-Shen Junior High School,

Taidong, Taiwan, R.O.C.

(Received October 9, 1996; Accepted August 22, 1997)

Abstract

The purposes of this study were to examine a beginning Taiwanese female science teacher's physicalscience teaching, and how her science teaching developed from her senior year to her first semester ofclassroom teaching. Both microscopic and macroscopic pedagogical content knowledge (PCK) wereimplemented in the investigation. A qualitative research method was applied. The research was conductedfrom Shu-May's senior year to the first semester of her beginning year. The findings revealed that thenature of Shu-May's PCK consisted of instructional strategies and representation. Her instructionalstrategies were characterized by the verification method of introducing science concepts. Her instructionalrepresentation included linguistic expressions, calculation problems, demonstration, daily-life experi-ences and relevant examples via verbal and visual display. Shu-May's PCK development revealedintegration, relevance, and specificity features. Shu-May's knowledge of her students changed froma general view of student learning to a more specific understanding of their content learning ability.Her content knowledge also changed from a general view of the physical science discipline to specificknowledge of what students should know about each topic. Factors influencing Shu-May's PCKdevelopment were her perceptions of science teaching and students' science learning, her insensitivityin both judging students' level of understanding and deciding on appropriate goals for the lessons, herdeclining retention of pedagogical knowledge learned in teacher education programs, and the differentcultural norms she perceived in different teaching contexts.

Key Words: pedagogical content knowledge, beginning science teachers, teacher's professionaldevelopment, physical sciences, junior high schools

I. Introduction

Teachers play a key role in influencing students'science learning. Science teacher education also playsa crucial role in educating science teachers to becompetent and capable of integrating content andpedagogical knowledge in classroom teaching. Dur-ing the past decade, research on science teacher edu-cation has gained more attention than before. Reformsin science teacher education programs have progressedin many states in the U.S. (Brunkhorst, Brunkhorst,Yager, Apple, & Andrews, 1993; Gilbert, 1994; Tippins,Nichols, & Tobin, 1993).

In Taiwan, a decentralized teacher education

system was approved by legislators; teacher educationis now in a transition stage. Both normal universitiesand general universities can offer science teachereducation programs. In normal universities, scienceteacher education programs are located within sciencedepartments. When students pass the entrance exami-nations and enroll in the science departments, theyneed to take both content and education courses to-gether for four years; practice teaching in a junior highschool requires an additional year. Therefore, studentsenrolled in normal university program take five yearsto become a teacher. In other general universities,students enrolled in a chemistry department need totake additional education courses offered by a teacher

135

56

H.L. Tuan and R.C. Kaou

education program to become a teacher. Once theyreceive a diploma they need to find a high school inwhich to have a year-long practice teaching experi-ence.

The new legislation in teacher education raisesa lot of research issues, including developing anassessment system for evaluating science teachers anddeveloping appropriate educational programs for sci-ence teacher education. Although there are many waysto improve science teacher education, Finley, Lawrenz,& Heller (1992) indicated that carefully examining theimpact of science teacher education on beginningteachers' professional development was crucial indesigning future teacher education programs.

Presumably, a teacher needs to have contentknowledge and know how to teach in order to be acompetent teacher. In other words, to be competent,teachers need to both possess and integrate content andpedagogy knowledge. However, few researchers haveinvestigated the above assumption until the last de-cade. Science educators (Gess-Newsome & Lederman,1993; Hashweh, 1987; Hoz, Tomer, & Tamir, 1990;Lederman, Gess-Newsome, & Latz, 1994; Lederman& Latz, 1995) started to investigate the importance ofboth pre- and in-service science teachers' content andpedagogy knowledge and the way the knowledge wasintegrated in teachers' thinking and acting. Shulman's(1986, 1987) concept of PCK, which addressed theteachability of content knowledge, is another directionin investigating the integration of teachers' contentand pedagogy knowledge.

Besides revealing the nature of how teachers'integrated content and pedagogy knowledge in think-ing and acting influenced their teaching, teachereducators also expected to unravel the underlyingprocess of teachers' professional development fromthe university through their field experiences. Thereare several models of teacher professional develop-ment including those of Fuller & Brown (1975), Berliner(1988), and Kagan (1992).

Fuller & Brown (1975) identified the process ofa teacher's professional development from the sur-vival stage to the impact stage. In other words, ateacher's concern passes from concerns about one'sown adequacy to the teaching task and finally to his/her effect on students.

Applying both schema theory and cognitivescience, Berliner(1988) constructed a five stageprocess of a teacher's professional development. Itincludes: novice, advanced beginner, competent, pro-ficient, and expert stage. In the novice stage, a teacher'steaching performance is based on his/her rationale,which is inflexible and requires purposeful concentra-

tion. In the advanced beginning stage, a teacher learnsstrategic knowledge and becomes flexible in followingrules in the classroom. In the competent stage, ateacher is able to choose a course of action, prioritizethings and make plans. In the proficient stage, ateacher is able to pick up information from the class-room without conscious effort and is able to predictclassroom events. In the expert stage, a teacher'sperformance is fluid and seemingly effortless. He orshe uses standardized, automated routines in handlinginstructions and management. Berliner thought mostbeginning teachers were in the novice stage, and manysecond and third year teachers were in the advancedbeginner stage.

Kagan (1992) categorized a teacher's professionaldevelopment into five components: an increase inmetacognition, an increase in knowledge about stu-dents, a shift in attention from self to students, de-velopment of standard procedures, and, finally, growthin problem solving ability. In Kagan's model, therewas no specific indication that years of teachingexperience were related to the developmental stages.

These authors provided some general ideas abouthow teachers may develop their teaching competence.However, they did not address how beginning teachersintegrated what they had learned from content andpedagogy courses into their content teaching presen-tations, or how they develop their content teaching inthe field, that is their PCK development. Future researchneeds to investigate specifically how beginning sci-ence teachers' develop their content teaching in anatural setting, and try to trace back the impact or lackof impact of science teacher education programs ontheir content teaching development. Only by this kindof-examination can we know how to improve our futurescience teacher education in a more meaningful wayfor preservice science teachers.

Previous research on PCK includes: examiningthe nature and substance of teachers' PCK in math-ematics and science teaching (Marks, 1990; Tuan,1996a); investigating the knowledge beginning sci-ence teachers need to teach a particular science topic(Geddis, 1993; Geddis, Ons low, Beynon, & Oesch,1993); exploring the development of pre-service sci-ence teachers' subject-matter knowledge, pedagogicalknowledge, and PCK development (Lederman, Gess-Newsome, & Latz, 1994; Lederman & Latz, 1995;Tuan, 1996b); illustrating methods to facilitate scienceteachers' PCK (Barnette, 1991; Clermont, Krajcik, &Borko, 1993; Dana & Dana, 1996; Klienfeld, 1992;Ormrod & Cole, 1996); and comparing the differencesof the PCK between novice and experienced scienceteachers (Clermont, Borko, & Krajcik, 1994; De Jong,

136

157

Development of A Physical Science Teacher's PCK

1997). The above studies used different definitionsto describe PCK, thus influencing their research ap-proach.

Tuan(1996c) reviewed research on PCK and con-cluded that research done on PCK revealed two views.The microscopic view has been emphasized by re-searchers like Shulman (1986, 1987), Kennedy (1990),McDiarmind, Ball, & Anderson (1989). These re-searchers have emphasized how teachers representtheir understanding of a particular content topic to aparticular classroom students' understanding level.The macroscopic view of PCK has been emphasizedby research by Cochran, DeRuiter, & King (1993),who have treated PCK as the overlap of several do-mains of knowledge learned in the teacher educationprogram, such as curriculum knowledge, pedagogicalknowledge, subject matter knowledge, knowledge ofstudents, and knowledge of context, and how thesetranslated into thoughts and actions in classroomteaching. The more integration of the above domainsof knowledge into the teachers' thoughts and actionsin classroom teaching, the more their PCK developed.The difference between the microscopic and macro-scopic views of PCK is that the former addresses aparticular content topic while the latter addresses abroader base of content teaching across different topics.Tuan(1996c) also suggested that future studies doneon this area should define the concept of PCK.

Researchers have reviewed a few studies on howpre-service/ beginning science teachers developed theirPCK (Gess-Newsome & Lederman, 1993; Lederman,Gess-Newsome, & Latz, 1994; Lederman & Latz, 1995;Tuan, 1996b; Yang & Wu, 1996). These investigationseither addressed the macroscopic view of PCK devel-opment (Gess-Newsome & Lederman, 1993; Lederman& Latz, 1995; Lederman, Gess-Newsome & Latz, 1994),or used the microscopic view to look at the PCKdevelopment of preservice teachers (Tuan, 1996a) orbeginning biology teachers (Yang & Wu, 1996). Noneof the studies identified traced the development of aphysical science teacher's PCK from one's senior yearto the first year of teaching, or investigated the de-velopment of a teacher's macroscopic and microscopicview of PCK.

Based on the above reasons, the purposes of thisstudy were to examine the characteristics of a begin-ning Taiwanese female science teacher's junior highschool physical science teaching, how her scienceteaching developed naturally from her senior year toher first semester of classroom teaching, and examinethe factors influencing the development of her PCK.Both microscopic and macroscopic levels of PCK wereimplemented in the investigation. Findings from this

study should help science teacher educators not onlygain more understanding about the nature of beginningscience teachers' PCK, and how this knowledge changesduring the first semester of teaching, but also helpscience teacher educators develop better science teachereducation programs in the future.

II. Design and Procedure

1. Description of Science Teacher EducationProgram

This study was done in a normal university locatedin the middle of Taiwan. As mentioned before, pre-service teachers enrolled in the chemistry departmentof the normal university have to take chemistry andeducation courses in the same department with thesame classmates. Pedagogy courses offered in thechemistry department included foundations of educa-tion, secondary education, educational psychology,science education, instructional materials and chem-istry teaching methods, and a practicum course. Coursesthat integrated content and pedagogy were chemistryteaching methods courses and practicum courses. Afterfinishing their course work in the fourth year, preserviceteachers have a year-long internship experience in ajunior high school to practice teach, obtain a teachingcertificate and receive a graduation diploma. Begin-ning teachers in their fifth year are graded and super-vised by junior high school principals and collegeprofessors; however, in reality, they are doing thesame work as professional teachers and seldom asktheir supervisors for help.

2. Description of Instructional Materials andTeaching Methods Course and PracticumCourse

The first author offered two courses for thechemistry department, a chemistry teaching methodscourse and a practicum course. In the chemistry teach-ing methods course, the first author covered teachingstrategies, such as lab teaching, conceptual teaching,discrepant event teaching, the learning cycle andcreative science teaching. The aims of the course wereto increase the preservice teachers' appreciation andteaching repertoire of different science teaching strat-egies. Instructional methods used for this course in-cluded introducing new methods, demonstration, groupdiscussion, role playing, group presentation, and wholeclass discussion. Basically, the author was a rolemodel using the above strategies in the class andthen conducted group discussions to help preservice

137


teachers discuss what they thought of the teachingmethods.

Practicum courses were offered for two semes-ters. For the first semester, preservice teachers prac-ticed microteaching twice on the topic they hadchosen; in the second semester, they had a one-monthpracticum in a junior high school. Instructional meth-ods used in the practicum course mainly focused onreflective teaching (Schön, 1986). The author appliedmany methods such as journal writing, self analysisof one's own teaching and reflection on other class-mates' suggestion to increase the preservice teachers'reflective ability on their teaching. The author alsoapplied peer coaching to supplement their reflectiveteaching experience. Each preservice teacher wasasked to choose their own peer coach. They thenhelped one another with lesson plans and gave feed-back on each other's performance. After finishingeach microteaching lesson, each preservice teacherneeded to analyze one's own teaching, and also re-spond to the feedback from their classmates and peercoach. During their one month field experience, eachpreservice teacher needed to make a journal of theirclassroom teaching experience. After they came backfrom school, the author covered topics related toclassroom management, science fairs, and teachers'professional development.

3. Context of the Junior high School

The junior high school where Shu-May taughtwas located in a rural county in central Taiwan, withfarmers or blue-collar workers constituting the major-ity of the population. Thus, most of the students'parents did not have high expectations for theirchildren's achievement. The learning environment inthis school was more casual without the competitive-ness of city schools. The class observed was an eighthgrade physical science class of 45 "ordinary" students.An ordinary class in that school meant that the stu-dents' achievements were lower than the A classstudents. The school selected higher G.P.A. studentsfrom among the same grade level and put them intothe A class in order to prepare them for passing thesenior high school entrance examination. The otherstudents were put into ordinary classes where probablynone of the students would pass the senior high schoolentrance examination nor be able to enter publicprofessional schools after their graduation.

4. Data Collection

Based on the nature of this study, a qualitative

research method (Bogdan & Biklen, 1992) was ap-plied. A beginning Taiwanese female science teacher,Shu-May, was involved in the study for many reasons:(1) she took the first author's chemistry teachingmethods and practicum course in her senior year ofcollege; (2) the researchers had collected teachingperformance data in her senior year in the program;(3) the location of Shu-May's school was close to theuniversity, enabling intensive classroom observation;(4) among her classmates, her G.P.A. was in the averagerange; and (5) she was willing to participate in thestudy.

A. Data Collection in the First Year

Data collection in Shu-May's senior year in-cluded two micro-teaching video-tapes of about 20minutes each, one hour of real classroom teachingvideotape, her lesson plans for the microteachingexperience, and other assignments for chemistry teach-ing methods and practicum courses. These data weremainly collected by the first author, who was also herinstructor for both the chemistry teaching methods andpracticum courses. The second author had one yearof junior high school teaching experience and twoyears of science education graduate courses. He wasShu-May's friend in college, played the role of an-observer-as-participant (Gold, 1958), and collecteddata on Shu-May's first year of teaching.

B. Data Collection in the Second Year

In the second year, data collection included atwice weekly classroom observation for one semester,interviews with Shu-May before and after her class-room teaching, and a collection of Shu-May's writtendocuments such as her lesson plans, notes and stu-dents' tests. Formal interviews were conductedbefore, during, and at the end of the semester andaddressed Shu-May's perceptions of junior high schoolphysical science teaching, her chemistry knowledgeand pedagogical knowledge (see appendix I). Infor-mal interviews conducted before and after each class-room observation focused on Shu-May's plans for thetopic she was going to teach, her knowledge of thetopic she was teaching, the references she used in herteaching, and her reflection on her teaching. Research-ers also interviewed Shu-May on the reactions col-lected from students. Document collection includedher lesson plans, tests and students' test results duringher first year of teaching.

5. Data Analysis

138

Data analysis was based upon the open-coding

15.0


method (Bogdan & Biklen, 1992). The second authorfirst analyzed all the interview transcripts and class-room observation field notes, and then generated anopen coding system. The two researchers examinedthe raw data such as interview transcripts and video-tapes of Shu-May's classroom teaching together inorder to elaborate on the coding systems. The firstauthor played a critical role in analyzing the raw dataand reading the temporary findings written by thesecond author. She then evaluated these tentativeresults and suggested that the second author find moreevidence to support or reject the findings until theauthors reached a consensus. The peer debriefingprocedure (Lincoln & Guba, 1985) was used in thedata analysis to increase the trustworthiness of thefindings.

Miles & Huberman (1994) suggested using tablesto present qualitative findings, and a table was gen-erated listing the frequency of Shu-May's teachingactivities (Table 1). Together the researchers analyzedthe beginning of three video tapes of Shu-May's teach-ing to count the frequency of her teaching strategiesin order to gain consensus on the attribute and fre-quency of each teaching strategy. The second authoranalyzed the rest of the video-tapes of Shu-May'steaching performance. The definitions of each of herteaching activities are listed in Table 2.

The purpose of the study was to look at how thebeginning teacher integrated content, pedagogy, andstudents together in thought and performance on hercontent teaching in a natural setting. Thus, teachingactivities (Table 1, 2) were analyzed but the appro-priate level of Shu-May's teaching performance wasnot. Another part of the data analysis that needs tobe considered was that, based on the nature of thetopics, each topic could accommodate different teach-ing approaches based on the content. For instance,an element and a compound can be introduced by usingthe history of science approach; for the water and airtopic, the teacher could first use daily-life experiencesfor students to appreciate the importance of water andair, and then discuss the property of these matter.However, when we examined Shu-May's teaching, herteaching strategies were almost the same when cov-ering different topics. Thus, we coded her teachinginto different teaching activities (Table 2) and thenexamined how Shu-May organized these teachingactivities, which were her teachling strategies, acrosstopics. The third important issue that needs to beaddressed is that during her fifth year Shu-May reliedon her mental thinking when planning instead ofactually writing down everything she needed to coverin the class. In addition, in responding to the authors,

her thinking about teaching was to mention in general,some teaching tasks, such as lecture, demonstration,and problem solving. Thus, the authors could hardlyanalyze Shu-May's planning in terms of frequenciesand level appropriateness of the representations listedin Table 1 & 2. Therefore, our data analysis reliedon her verbal descriptions of teaching and actualobservation of her classroom performance.

Lincoln & Guba (1985) listed four criteria formaintaining the trustworthiness of qualitative researchfindings. These are credibility, transferability, de-pendability, and confirmability. In this study, the tworesearchers knew Shu-May and had worked with herfrom her senior year to her first year of teaching. Thesecond author stayed in the field for half a year,providing an opportunity for the researchers to collectShu-May's PCK and to clarify the researchers' under-standing. This prolonged engagement should help toincrease the credibility of the findings. Triangulationmethods (Bogdan & Biklen, 1992) were also appliedin the data collection and analysis procedures, suchas collecting data from different sources and includingdifferent researchers in the analysis of the findings.Finally, the study provided the following: detaileddata analysis procedures, Shu-May's views of theresearch findings, comparisons between our findingsand previous research, detailed information on theresearch context and the roles of the researchers. Thesemethods were incorporated to increase the trustwor-thiness of the findings.

III. Findings

1. The Development of Shu-May's pedagogicalcontent knowledge

A. The main organization of Shu-May's science teach-ing strategies did not change throughout the study;she used a verification method to teach scienceconcepts listed in the textbook from the beginningto the end of the study. Toward the end of the study,Shu-May used more teaching strategies to helpstudents memorize science concepts.

Analysis of Shu-May's teaching from her senioryear to first year of teaching showed that her teachingactivities(Table 1, 2) and ways of organizing theseteaching activities were very similar to other pre-service Taiwanese chemistry teachers (Tuan, 1996a).These teachers' content teaching were characterizedas the verification way to teach science.

Shu-May had an opportunity to practice ideasintroduced in the methods course in her microteachingexperience during her senior year. However, she

139

160


Table 1. Shu-May's Teaching Activities for Each Topic

Classroom

Teaching Activity

Frequency

Topic Taught in Senior Year Topics Taught in the First Half Year of Teaching

Microteaching Lessons Field Experience Lessons

Linear Movement

in Velocity

(27 min)*

Elements with

Similar Properties

(29 min)*

Property ofHalogen

(50 min)*

Measurement Water Influence Elements

& Units ofMetric

(300 min)*

& Air

(150 min)*

of HeatMatter

(200 min)*

Compounds

(250 min)*

I. classroom management 1 32 17 12 17

2. review previous concepts 2 10 6 8 14

3. introduce formulas 2 1 5 4 4 5

4. ask for memorization 1 7 10 7 12

5. underlining important parts in

textbooks and/or taking notes 2 12 13 16 21

6. emphasize the content whichwould be convered on test 1 9 3 5 6

7. using added scores toencourage Ss 2

8. ask open-ended questions 3 4 5 9 4 8

9. ask individuals to respond 2 6 4

10. provide opportunities

for Ss to discuss 2 1 4 5 2

11. quiz 2

12. explain science concepts 4 6 29 I 1 26 28

13. summarize introduced concepts 3 8 17 4 20 15

14. demonstrats problem solving 3 11 2 10 3

15. students do problem

solving exercises 4 2

16. provide daily life examples 34 17 20 2

17. use analogies to explainconcepts 2 1 3

18. use illustration to explainconcepts 4 1 16 7 10

19. use graps to facilitate problem

solving 2 3 1 2

20. demonstrate concepts usingreal objects 1 2 3

21. demonstrate lab activities 1 2 3

22. Ss conduct lab activities 1 4

* Length of teaching time.

thought the best way to present the velocity topic wasa "traditional lecture." The strategies she used inintroducing velocity were to introduce the key con-cepts at the beginning of the lesson, review the pre-vious lessons in order to connect previous conceptsto this topic, define velocity, provide problem

solving techniques for both demonstrations andexercises for students to appreciate aspects of ve-locity, illustrate the difference between velocity andspeed, derive a velocity formula, and summarizeconcepts. In her twenty-seven-minute microteachingexperience, Shu-May's major emphasis was on ex-

140


Table 2. Example of Each Teaching Activity

Teaching Activities Examples

1. Classroom management2. Review previous concepts3. Introduce formulas4. Ask Ss to memorize

5. Ask Ss to underlineimportant points or takenotes

6. Address the content whichwould be covered on test

7. Apply additional score toencourage students

8. Ask open-ended questions9. Ask individual students

to respond10. provide opportunity for

student discussion11. Quiz

12. Explain and introduceconcepts from textbook

13. Summarize introduce labfinding and concepts

14. Demonstrate problemsolving

15. Let students do problemsolving exercise

16. Provide daily life examples

17.Use analogies to explainconcepts

18. Use illustration to explainconcepts

19. use graph to facilitateproblem solving

20. Demonstrate conceptsusing real object

21. Demonstrate with labactivities

22. Students conduct labactivities

T: Lin-Pin, do not talk in class, pay attention.T: Last period we studied page 34,Density. Density = Mass /volumeT: Today I will introduce pressure, pressure = Direct force/area.T: The definition of the law of constant composition, the mass of all elements in a compound, has a certain ratio. This

is very important, memorize it.T: Look at page 161, the third line, the law of constant composition, all elements combine into compounds,

their mass have the same ratio, this is very important, underline this paragraph.

T: "The law of constant composition is very important, it will be on the test."

T: If you answer this question correctly, I will add three points to your average score.

T: How to separate sugar and iron powder ?T: Ah-Lin, density equals what divided by what?

After asking questions, Shu-May gave time for students to conduct discussion.

After covering the content in the class, Shu-May asked the students to take out a piece of paper and pen to take a quiz.She then wrote down a question on the board: There is a fish tank on a square table. The table weights I200g. Thefish tank weighs 400g. Each side of the table is 10 cm in length. (I) What is the direct force? (2) What is the totalcontact area? (3) What is the pressure ?T: Let me introduce a term. The vertical direction of the force on the object, we called the vertical

force...How is the vertical force related to pressure? SS: Direct proportion.After explaining science concepts, Shu-May would summarize what she had just taught to the studentson the blackboard. For instance, after introducing pressure, she then wrote doWn the following notes on the blackboard.P (pressure) = F (vertical force) gw kgw / A (area) cm2T: look at page 45, there are two bricks. How much does each brick weigh? SS: 2000gT: How about the length of the brick? (She wrote on the blackboard)SS: 20 cmT: How about the width of the brick ?SS: 10 cm.T: How about the height of the brick ?SS: 5 cm.

[Wrote the following notes on the blackboard]F=2000g Length 20 cm, width 10 cm, height 5cm

T: Look at the brick lying on the table. What is it's pressure on the table?Does it relate to vertical force divided by area? How much is the vertical force?

SS: 2000gT: How much is area?SS:200 cm2T: Length times weight is 20 times 10 equals 200

[Wrote the following notes on the blackboard]2000

20x10 = (10 g / cm2)

Shu-may wrote problem on the blackboard and asked the students to do the problem solving exercise in the class.

After introducing how heat influenced matter, Shu-May talked about how water at 4 °C influences the ecology.T: How can fish live in a lake covered with ice on the surface ?T: Because at 4 °C, water has the largest density.. Thus, let's compare 4 °C water and ice. Which one is on the top;

which one is at the bottom?SS: 4 °C, water is on the bottom.T: So even if the lake is covered with ice, underneath the ice is the 4°C water; thus, the fish can stay alive.T: The diffusion rate of Hydrochloride and ammonium is like a fat guy and a skinny guy racing. Generally

speaking, the skinny guy runs faster than the fat guy. It is the same with the diffusion rate of gas.The lighter gas diffuses faster than the heavier one.T: Chemical reaction is the rearrangement of atoms, such as carbon and oxygen combusting into carbon

dioxide. The solid dot represent carbon. The hollow dots represent oxygen.C + 02 CO2

00 00When Shu-May asked ss how to measure the volume of Pine-Pong balls, she drew a graph on the boardand explained to ss how to do it.

20m1 100m1 80m1

Shu-May used a bottle of water to show ss the relationship between the depth of water and pressure.

Shu-May demonstrated lab activities from the textbook

Shu-May invited students to help her conduct lab demonstrations , or she let students do lab work in groups.

141

11. 62


plaining science concepts to the students, who werealso her classmates.

In Shu-May's first half-year of junior high schoolteaching, she used a similar teaching pattern in orga-nizing her lessons. The pattern followed a usual se-quence: (1) introducing science terminology; (2) usingnatural phenomena to illustrate science terminology;(3) explaining the definition of science concepts,principles, and/or formulas; (4) providing examples tore-emphasize concepts, principles and/or formulas;(5) providing problem solving exercises to help stu-dents understand previous concepts; and (6) usingquizzes to evaluate students' understanding. In gen-eral, Shu-May's teaching activities were mainly or-ganized to focus on introducing science concepts tostudents, and to help students memorize the conceptsin order to be successful in taking school examina-tions.

The literature of PCK has indicated (Geddis,1993; Shulman, 1986) that using conceptual changeteaching methods to overcome students' misconcep-tions is one important feature of PCK. This charac-teristic was not found in Shu-May's science teachingfrom her senior year to the first year of teaching. Shu-May's way of organizing instructional strategiesthroughout the study were characterized as using averification method to organize science concepts forstudents. Shu-May did make some changes in herinstructional strategies at the end of the study; thesewere: (1) increasing some representations to helpstudents appreciate science concepts; (2) increasingthe frequency of helping students to memorize scienceconcepts, such as underlining the important conceptsand focusing on the possibility of the concepts on tests;and (3) practicing problem solving techniques (Table1). These strategies were to help students get goodgrades on school tests.

B. Although toward the end of the study, Shu-Mayincreased her use of various representations, suchas providing daily life examples and using illustra-tions to represent science concepts, her majorrepresentations still relied on both linguistic ex-pression and calculation problems.

Many researchers(McDiarmind, Ball, & Ander-son, 1989; Shulman,1986) have indicated that PCKincludes how a teacher represents content to a student'slevel of understanding. In Shu-May's senior year, shechose two topics for microteaching teaching experi-ences; one was on the velocity of linear movement andthe other was elements with similar properties. Forvelocity, she spent the majority of her time in explain-

ing the definition of velocity, deriving the formula forvelocity, and using problem solving to help studentsappreciate and remember the concepts she had justtaught. In the second microteaching experience, Shu-May introduced important concepts related to "howto categorize the metal elements," reviewed the pre-vious lessons in order to remind students of the prop-erties of the Alkali metal group and Alkaline earthmetal group, and introduced the elements by showingand introducing each element to students. After that,she drew the same table listed in the textbook on theboard (the purpose of the table was to help studentsto see the differences in appearance and propertiesof elements before and after chemical reactions),reminded students of lab safety and some importantprocedures, demonstrated the first two procedures,and then divided students into three groups to conductthe lab activities. At the end of the lab, Shu-Maysummarized the students' lab results (Table 1).These teaching representations were very traditionaland mainly relied on Shu-May's linguistic expres-sions.

In her junior high school teaching, she liked touse linguistic expressions, problem solving, demon-stration, daily life examples and relevant examples byverbal or visual displays to help students quicklycomprehend and then memorize and apply physicalscience concepts.

Shu-May liked to use visual displays, either bydrawing or showing concrete objects to motivate stu-dents in learning new terminology and in appreciatingthe substance of abstract concepts. Figure 1 showshow Shu-May used drawings to express concepts relatedto pressure by drawing on the blackboard four bottleslaying on a sponge.

T: The first bottle contains less water, the second, third andfourth bottles are all filled with water, but the fourth oneis upside down on the sponge. Compare the depressionlevel of the sponge.

Shu-May used concrete objects to help studentsappreciate concepts. For example, Shu-May broughta plastic bottle filled with water, with three holes(stabbed with a sharp-pointed pen) covered with tapeon the side of the bottle. She asked the students topredict which holes would let water spurt thefurthest(the upper, middle, or bottom hole). She thenremoved the tape from the holes and let the studentssee the result, and asked them which factor influencedthe water spurt. The students stated depth of the wateras the answer, and Shu-May explained that the pres-sure of water is influenced by the vertical force on the

142

.263


(Note: These pictures were drawn on the textbook)

(Shu-May drew these pictures on the blackboard)

Fig. 1. The comparison between the picture of pressure in thetextbook and in Shu-May's teaching.

side of the bottle.Shu-May mentioned that the abstractness of

concepts was the indicator for her to use a visualdisplay; these visual displays ranged from real objectsto symbolic displays, such as diagrams. Shu-Mayexplained her timing in using a visual display:

If it (abstract concept) could be expressed by a real objectthat will be great...if they do not understand atom, or mol-ecule, I will draw some pictures on the board, or use Styrofoam

balls to build a model to explain abstract concepts. If theconcepts can be seen, such as the combustion of chemicals... I will use real objects to show students the results ofcombustion. Anyway, if there is a real object that candescribe the concepts, I will use the real object to showstudents.

Besides the above instructional representations,Shu-May would improvise with daily-life examples inher teaching. The purpose of using daily-life exampleswas to help her students appreciate the substance ofthe science concepts. Sometimes daily-life exampleswere used to help students understand the applicationof the concepts she had just taught. For instance, inintroducing "pressure", Shu-May first defined whatpressure was, and then she said:

"if two persons sat on a couch, one a fat guy such as HoungChin Paou ( a famous movie star), and the other is Lee-Pin(a skinny student in the class), will they make the couch sink

at the same level, or is there any difference in the sink levelbetween these two people ?" All the students laughed whenshe used names the students knew.

To Shu-May, these examples were relevant to thestudents so that they could not only appreciate theconcepts but also know how to apply them to a newsituation.

In another example, Shu-May used a concreteillustration to verify the students' appreciation of

science concepts. She asked the students to use twofingers to press the two ends of a ball point pen andthen asked the following questions:

T:Press the two ends of the ball-point pen. If the ball-pointpen holds steady within your two fingers, that means theforce from the right and left hand sides are equal. If itfalls from between your two fingers, which one has morepressure?

SS: The sharp side of the ball-point pen.T: The larger the area, the lower the pressure; so what is the

relationship between pressure and area?SS: Inverse ratio.

In summary, during the investigation period, Shu-May increased her use of visual displays and daily-life experiences to help students understand scienceconcepts (Table 1). However, these representationswere not the main themes in her teaching. In fact, Shu-May mainly used linguistic expressions in the class-room, such as "Let me introduce some terminology,the vertical direction of a force on an object is calleda vertical force" and "Today, I will introduce pressure,Pressure = Direct Force/ Area." The researchers be-lieved these linguistic expressions were too abstractand complicated for students to appreciate and tounderstand. Shu-May did not use more concrete waysto introduce and develop concepts. Besides linguisticexpression, Shu-May would demonstrate problemcalculations most of the time in order to show studentshow to apply the concepts to a calculation situation.In this kind of demonstration she expected studentsto distinguish and know all the variables involved inthe concept, and to know how to apply these conceptsto a new problem situation. To Shu-May, problemsolving exercises fulfilled two purposes: (1) to helpstudents have a better appreciation for the substanceof concepts or principles; (2) to help students applywhat they learned to a problem solving situation. Forinstance, when she taught about sodium, she wrotesome information on the blackboard.

143

Atomic numberAtomic numberelectronic

12Na23 neutral

Electronic number:Proton number:Atomic number:Neutron number:

T: So if the atomic number is 11, then what is the protonnumber?

SS: 11T: The proton number is the same as the atomic number 11.

Then how about the electronic number?

g64


SS:11

T: Did I say that sodium has electronic charge? No. So, theelectronic and proton number are all the same; that is 11.How about the neutron number...

SS.

T: What is the atomic weight?SS:23

T: As I have said before, almost all the mass is in the atomicnucleus, so all the proton number plus the neutron numberis the atomic mass number.

T: If the atomic mass is 23 and proton is 11, then what isthe neutron number ?

SS: 12 (Fn940113)

These findings were very similar to data reported byTuan (1996a) in a previous study; three of the pre-service chemistry teachers expressed the dual pur-poses of using problem solving as did Shu-May. Theyused problem solving not only to help students applythe principle/formula to problem situations, but alsoto help students memorize the principle/formula inorder to obtain good grades on examinations.

As summarized by Thorley & Stofflett (1996),a teacher's modes of representation for conceptualchange teaching may include linguistic expressions,criteria attributes, exemplars, images, analogies ormetaphors, kinesthetic or tactile, and other modes ofrepresentations. Although Shu-May used some of therepresentations, such as using daily-life examples andusing illustration to represent concepts, her represen-tations were primarily linguistic expressions ratherthan other representations. These approaches sup-ported her verification way of organizing her instruc-tional strategies.

C. Shu-May's knowledge of students related to bothstudents' science learning characteristic and abil-ity increased, this knowledge influenced her in-structional representations and concept emphasesand related variables.

During the study period, Shu-May gained knowl-edge about students' learning. In Shu-May's senioryear, her views of how students learned science werefocused on memorization and recitation. She thoughtstudents needed to use problem solving to becomefamiliar with concepts. In the teacher educationprogram, the educational psychology course hadstressed students' cognitive development. In theinstructional materials and teaching methods course,the importance of students' learning by doing scienceand by experiencing concrete objects had been empha-sized verbally. However, these emphases did not seemto have much impact on Shu-May. Her views of

students' science learning were very general, and theyalso reflected Shu-May's view of her own sciencelearning. One excerpt illustrated Shu-May's view ofstudents' science learning at the end of her senioryear:

In the beginning [of the semester], they read the textbookas if they were reading a story. Then thcy uscd a red pento underline the important points. After they had underlinedthe important points, they would try to understand the con-cepts. When they practiced problem solving, they were fullyable to understand what the concepts meant.

After Shu-May interacted with her students,she started to talk more about how the students' arith-metic and Chinese comprehension ability influencedtheir science learning. For instance, in a calculationsituation, Shu-May mentioned the students couldonly write "density equals mass divided by volume,"but did not know how to calculate the equation. ButShu-May treated this problem as a Chinese literacyproblem because the students did not understandwhat the "unit" of mass meant. To her, there was noway to help them understand the problem. She didnot recognize that even though the students hadmemorized the definition of density, since they didnot understand the concept, they could not applythe scientific terminology to a new situation. Thisrepresented a more fundamental learning problemthan what Shu-May considered. A group of studentsexplained to the researchers why they could notlearn well in physical science. One student men-tioned:

It (learning physical science) not only requires memorization

but also calculation. We are not good at mathematics andcannot solve calculation problems in physics, so we cannotlearn physical science well.As Shu-May described after the class: I gave them a test on"one meter is equal to how many kilometers." Almost halfthe class wrote 10km.

Besides being unaware of some of the students'problems in learning new science concepts, Shu-Maydid not know their previous conceptions or difficultiesin learning typical science concepts. Interviews withShu-May before teaching new topics indicated she didnot think of students' preconceptions; she did mentionsome learning difficulties related to some scienceconcepts. She judged the difficulty level of conceptsbased on the abstractness of the concepts, such as moleconcepts, atom, molecular, etc. However, interviewswith Shu-May regarding the teaching strategies she

144

16 5


would use to reduce students' learning difficulties, andobservations of her teaching, indicated she usedfew alternative ways to reduce student learningdifficulties. What Shu-May used were mainly (1)repetitions, such as using the same analogy (mole islike a dozen of eggs, etc.), or (2) consistently remind-ing students of the definition of the terminology,such as, "Do you remember I mentioned before,one mole of molecular mass is equal to all the atomsmass in the molecule." Shu-May could have used thestudents' language and familiar examples or objectsto help students comprehend the substance of conceptsinstead of using repetitious ways.

Shu-May did not really grasp the students' learn-ing difficulties. Only after seeing the students' testresults did Shu-May realize their difficulties in learn-ing particular science concepts. Teachers' unfamil-iarity with students' preconceptions were also foundin studies of other Taiwanese preservice chemistryteachers (Tuan, 1996a) and in U.S. elementary mathteachers (Marks, 1990). To Shu-May, the students'literacy (in this case their Chinese language compre-hension ability) and arithmetic abilities, and theirbackground in science were more important in influ-encing them in learning than their preconceptions. Shefelt they provided the foundation for acquiring newknowledge. In the beginning of the study, Shu-Maydid not judge whether students would like science andthe reasoning behind their science attitudes. Towardthe end of the study, Shu-May could explain the stu-dents' declining attitude toward science learning in-stead of just being aware of the phenomena. Shu-Maydescribed why the students' attitude toward sciencedecreased dramatically:

Physical science is difficult. In the beginning of the semester,

students liked physical science because of the lab activities.They were willing to listen, watch and memorize scienceterminology. But! When they faced calculation problems,they totally gave up. (Int 940106-3)

Interviews indicated students responded similarly byindicating "lab is fun" and they were very bad at"problem solving."

Toward the end of the study, Shu-May increasedthe variety of her concept representations and empha-sized their relationships to students.

In the beginning of the study, Shu-May's ideasabout the content of a particular topic mainly consistedof several important concepts listed in textbooks. Inher senior year, when teaching the unit of constantvelocity movement, she described how she wouldintroduce this unit:

I wrote down four important parts on this topic [in the lesson

plan]. First is why we need to learn Velocity and the definitionof constant velocity movement. Second is the differencebetween speed and velocity... Third, I will use illustrationon S-T and V-T to show students the relationship betweenposition and time, velocity and time, and tell students thedifference between these two. (Inv921012)

The observations of Shu-May's microteaching, de-scribed in the previous section, were the same as whatshe said in the above interview. Shu-May tried to helpstudents understand the variables in the concepts usingmany calculation problems; few representations wereused in her microteaching.

Toward the end of the study, Shu-May started tocare more about using different content representa-tions and helping students distinguish the differencebetween the variables of the concepts. For instance,in interviewing Shu-May on her knowledge of pres-sure:

Researcher: What do you think this topic (pressure) is about?What is the key point?Shu-May: This topic is about the concepts of pressure... I

will address this formula (P=F/A), help them appreciate thisformula, and explain to the students the relationship betweenpressure and areas... (Int931007-01)Researcher: How would you teach pressure to students?Shu-May: [in the textbook], it only described the relationshipbetween pressure and areas, but I think force also needs.to be addressed in this topic, so that they can appreciatepressure in relationship to area and force. This organizationis better [to show students the whole picture of pressure andthe variables related to pressure]. (Int931007-01)

The way Shu-May represented the concepts of pres-sure are listed in Fig. 1. She drew four pictures onthe board instead of the two pictures in the textbookand explained to students how direct force and areainfluence pressure.

In summary, Shu-May made several changes inher knowledge of students' science learning. In thebeginning, she only had general views of students'learning characteristics, and focused on memorizationand recitation. Toward the end of the study, herconstructed knowledge of students' science learningability became more focused, especially on how stu-dents' previous capacity influenced their sciencelearning. She also could explain students' decliningattitude toward science. Knowledge of students'science learning that influenced Shu-May's teachingwere that she increased her thinking about and applieddifferent instructional representations, and emphasizedconcepts and related variables to students in her teach-ing.

145

166


D. Shu-May's thinking on content knowledge becamefocused and more related to the goals of juniorhigh school physical science textbook. Based onthe low ability of the students she taught, she re-duced the amount of content covered in her teach-ing.

In Shu-May's mind, the purpose of her teachingwas to reach the goals of the junior high schoolcurriculum, which she believed should be very prac-tical for her students. Thus, all the content knowledgeshould be relevant and useful to the students in theirdaily lives.

In Shu-May's senior year, she spent 27 minutesin introducing five different types of velocity calcu-lation problems in class with an emphasis on calcu-lation. In asking her view on chemistry, she describedthe different disciplines, such as organic, inorganic,analytic, and physical chemistry, and provided de-tailed descriptions of each discipline.

Toward the end of the study, Shu-May constantlymentioned that the students' low ability made herselect a few simple concepts to teach. For instance,in introducing the law of constant composition, Shu-May taught students:

Shu-May: Let me introduce the law of constant coniposition.What is the law of constant composition? [She put thefollowing statement on the board]Law of constant composition: H20Tap water --- WaterSea water --- Water H:0River water Water Ratio of weightUnderground water --- Water[Shu-May started to explain to students]Shu-May: ... No matter where the original sources of thewater, as long as you extract it, pure water will alwaysbe H20. Do you understand? No matter what the originalsources of the water are, as long as we extract it, it willalways be pure water.... So the mass between hydrogenand oxygen is 1:8.... It is constant. No matter where thesources of the water are, the weight ratio between hy-drogen and oxygen is always 1:8.... Let's turn to page 164[she

read from the textbook]. We used experiment to prove theratio between hydrogen and oxygen in the water. What isthe ratio?SS: 1:8.

This example showed how Shu-May tried to reducethe difficulty level of the concepts and tried to simplifythe content for the students. In this case, she onlyasked the students to remember the ratio betweenhydrogen and oxygen in water, and carbon and oxygen

in carbon dioxide. These two examples were bor-rowed from the textbook. Except for teaching whatwas in the textbook, Shu-May did not use calculationproblems to show students the law of constantcompositions in other complicated chemical com-pounds.

Toward the end of the study, when interviewedabout what chemistry is, what the important con-cepts in chemistry are, and how chemistry relatedto other disciplines, she only described what waslisted in the junior high physical science curri-culum, and never mentioned the additional contentin her college chemistry courses. The same re-sponses were obtained when interviewing her onthe topic she was teaching. What concerned hermost was how to introduce science concepts in asimple way that could be appreciated by her stu-dents.

The above findings were also found in studiesof other preservice Taiwanese science teachers(Tuan, 1996b) during their senior year of practicumcourses. These preservice teachers related whatthey had learned in the chemistry department andthe topic they were teaching. However, after one-month of classroom teaching experience, theseteachers' responses to the previous questions shiftedto closely following the content knowledge listed inthe junior high school physical science textbooks;the relevance of college chemistry to junior highschool textbook topics did not seem to greatly influ-ence their science teaching. Shu-May and otherpre-service chemistry teachers indicated "what youhave learned in the chemistry department is tooabstruse to teach in junior high school physical sci-ence."

The changes in Shu-May's thoughts aboutcontent became more and more focused on the juniorhigh school curriculum. In the beginning of the study,although Shu-May thought teaching should relate tostudents' needs, her awareness was not reflected in herthinking or teaching of particular content topics.Toward the end of the study, Shu-May would con-stantly question the difficulty level of the concepts,and decided to teach simple concepts without exten-sive explanation.

2. Factors influencing Shu-May's PCK devel-opment

Analyzing Shu-May's data from her senior yearto first year of teaching, the researchers identifiedmany factors which influenced the development of herPCK.

146

167


A. Shu-May's perceptions toward science teaching andstudents' science learning influenced each otherand reinforced her presentation of science conceptsin a verification way.

Shu-May's teaching was characterized by usingthe verification method of introducing science con-cepts from her senior year to first year of teaching;this was influenced by her beliefs about science,epistemology, pedagogy and her perceptions of stu-dents' science learning.

Although she did not have any problems withclassroom discipline and students' learning ability inher microteaching experience, she thought the bestway to present content was by using a "traditionalteaching method." During Shu-May's internship ex-perience, she started to be concerned about students'learning abilities; however, her perceptions regardingthe students' learning and ways to maintain the stu-dents' motivation reinforced her continuous in usingverification ways in teaching.

In a low-ability learning environment whereclassroom discipline problems occurred easily,Shu-May believed the best way to present contentwas to use simple and repetitious ways to presenther content knowledge to the students. The students'achievement's in her science class declined as theamount of calculation problems increased on the schooltests, due to students' lack of arithmetic ability andtheir lack of motivation. Shu-May tried to solve theseproblems by spending time helping them to appreciatethe substance of science concepts and using verifica-tion instruction to have the students memorize thecontent knowledge to help them obtain higher achieve-ment scores. Achievement scores were the indicatorsShu-May used to check students' understanding; italso became the goal in Shu-May's teaching.

Shu-May: Their (students) backgrounds were not good, theydid not know. So, you have to force them to memorize some

facts.

Researcher: What is your orientation toward physical science

teaching?Shu-May: Helping them understand and then helping themget good grades. (Int 930929)

She seemed to treat science as a body of termi-nology (or factual knowledge) for students to know.She also thought that students' mastery of previousknowledge would influence their learning of newknowledge; this reinforced her use of teaching strat-egies to help students memorize scientific terminol-ogy.

B. Shu-May was insensitive in judging the appropriategoals of the lessons and the students' needs forunderstanding, which made her content represen-tations beyond the students' comprehension level.

An important feature of PCK is to identify theimportance of the goal of the topic to match the stu-dents' level of understanding, and then decide theappropriate instructional methods to reach the goalof the chapter (Shulman, 1986). Although Shu-Mayincreased her thinking and practicing regarding therepresentations of the content to students during theinvestigation period (see Table 1), she did not effec-tively judge the appropriate level of content knowl-edge, the goals of lessons, or identify the comprehen-sion levels of the students; these conditions made hercontent explanations and representations exceed thecomprehension level of her students.

When the researcher discussed with Shu-May theimportance of addressing categorizing skills for thestudents, she responded:

only taught them what was addressed in the textbook. I did

not tell them the goals of this chapter. (Int940321-07)

In the unit on density; Shu-May described:I kept teaching the students the formula D=M/V....My col-leagues did not tell me I have to teach M=DxV or V=M/D...I found a problem in a reference book about the density ofmixed solutions; I really don't know whether I should teachthis problem or not. If I did not teach it, then I did not address

any important part in this unit. If I taught it, I was afraidthat the students could not understand it. (Int930929-15)

If students understood the concept of density,they could apply the density concept to many situa-tions. Thus, the teaching emphasis should have beenplaced in various representations for students to con-struct density concepts instead of helping them memo-rize different forms of the density equation.

Several variables influenced her teaching. Shu-May's previous tutoring experience in college influ-enced her in constructing a novice view of the juniorhigh school physical science curriculum. She thoughtbeing familiar with the concepts covered and orga-nized in the junior high school textbooks was the goalof the curriculum. Helping students succeed on schoolexaminations was considered by Shu-May to be a moreimportant goal than helping students to understandscience concepts and applying them in their dailylives. These factors influenced her in defining hergoals for the lessons.

Another variable influencing her teaching was

147

68


her insensitivity in judging students' needs for under-standing. Analyses of Shu-May's questions in hersenior year and her quizzes in the first year of teachingindicated most of Shu-May's questions and quizzeswere at the memorization level, such as write downthe Chinese name for sodium, and write down thechemical formula for water. Fewer than half of thestudents could reach a passing grade (60 points) onher quizzes. When students faced calculation prob-lems, most of them gave up trying. We interviewedShu-May about the reaction students had, and sherelated students' motivation to their confidence inachievement in the science class. Her thoughts ofusing quizzes was that they could both help her checkstudents' comprehension and help students success onthe school examinations. She believed that after hav-ing success on school examinations, students' moti-vation toward learning science would be sparked.Therefore, Shu-May thought that frequently conduct-ing quizzes would push students to memorize scien-tific terminology in order to get higher scores on theexaminations. When Shu-May saw students' failingon these tests, she was unable to diagnose the reasonsstudents did not understand the concepts. Therefore,she could not identify more appropriate ways to presentcontent to these students.

C. Shu-May showed evidence of comprehending con-cepts and procedures related to pedagogical knowl-edge in her college methods courses. Her inabilityto apply them to her teaching situation and laterlack of recall of them, however, indicated she didnot develop a functional understanding or use ofmuch of the pedagogical knowledge taught.

Analysis of Shu-May's teaching performanceshowed that her teaching strategies did not changeduring the investigation period. The only change wasin the frequency of some of her teaching activities,such as showing more concrete objects and usingdrawings or daily-life events to explain concepts.

As mentioned before, the first author had coveredmany science teaching methods such as lab teaching,conceptual teaching, discrepant event teaching,learning cycle and creative science teaching in thechemistry teaching methods. The author tried to usediscussion and role modeling in order for the preserviceteachers to appreciate the meaning and application ofeach teaching method. When she was studying theseteaching methods, she could explain and discuss ideaswith her classmates. In Shu-May's report, it seemedthat she was familiar with these teaching methods.One example shows Shu-May's assignment on herperception of the learning cycle:

In planning the lesson, a teacher has to think of the students'perception, and then provide a learning environment andteaching resources, let the students conduct labs and learnconcepts, and provide opportunities for the students to dis-cuss ideas so that they can apply the concepts to new situ-ations... The role of the teacher is to diagnose, facilitate,and provide a learning environment. (assignment-014)

Unfortunately, toward the end the of the study, Shu-May could not remember most of the teaching strat-egies or pedagogical knowledge from her collegecourses related to thinking, talking and performingrelated to her teaching.

Researcher: Can you think of any teaching methods learnedin college for preparing you in teaching?Shu-May: All the theories learned in college seem remoteto me. Maybe I have integrated all the pedagogy learned incollege or I have totally forgotten it. I do not know... I

have forgotten the learning cycle you just mentioned.(Int940321-03)

Researchers could not trace the topics coveredin the methods courses to her classroom teaching.What she often taught in the class represented contentknowledge for students to quickly appreciate andremember. Any conceptual change teaching methods,discrepant events, or learning cycles were rarely foundin Shu-May's teaching. Shu-May did not develop afunctional use of most pedagogical knowledge taughtin her college courses; therefore, she was unable toapply it in her first-year of teaching.

D.The culture norms Shu-May perceived in differentteaching contexts either reinforced or impeded herperception and performance in science teaching.

In Cochran's model (Cochran, et al., 1993) teach-ers' knowledge of culture and environment also in-fluenced the development of their PCK. Feiman-Nemser & Floden (1986) defined teaching culture as"embodied in the work-related beliefs and knowledgeteachers share - beliefs about appropriate ways ofacting on the job and rewarding aspects of teaching,and knowledge that enables teachers to do their work(p.508)."

In Shu-May's senior year, she was enrolled ina context where chemistry content knowledge learningwas the top priority. In addition, all the students shefaced were her classmates; no learning problems orclassroom discipline problems occurred in the teach-ing context. Her perceptions of using the traditionallecture style to present science matched with theexisting teaching culture, which reinforced her to

148


mainly address in-depth content explanations in hertwo microteachings. Shu-May's idea of helping stu-dents overcome their learning difficulties at this timewas by reTteaching many times until students couldunderstand the meaning. She was not involved in acontext that had any pressure, time limitations, or aprescribed curriculum pace.

However, when she was enrolled in her practicumschool, the environment where examinations, compe-tition and covering a prescribed curriculum at a setpace were practices and norms very similar to thosedescribed in several research reports (Abell & Roth,1992; Brickhouse & Bodner, 1992; Sanford, 1988),Shu-May tried to adapt to these practices and culturenorms. As Shu-May explained to the author why shecould not re-teach to confirm students' understanding,she said:

[In the beginning] I was concerned about whether studentsunderstood the concepts or not. If they did not understand,I would teach them again and again, until they grasp at theconcepts. But I can not use this way now. I was too idealistic

before... I realize the students' ability was low beyond myimagination. No matter how simple a concept, there werealways some students who did not understand it... School

also provided a prescribed curriculum pace, and examina-tions, I have to catch up to the prescribed curriculum. Thatproblem is a reality problem. (Inv931104-02)

In addition, Shu-May's perceptions of using achieve-ment scores to motivate students to learn reinforcedher adaptation to the cultured norms. This teachingculture and her perceptions toward teaching and learn-ing further reinforced her thinking that helping stu-dents get good grades on examinations was her maingoal in teaching. These goals could also provideintrinsic motivation for teachers to choose teachingactivities designed to foster achievement (Mitchell,Ortiz, & Mitchell, 1982). These goals also influencedShu-May to help students memorize science conceptsinstead of helping them appreciate the substance ofthe concepts and how science concepts could applyto their daily lives..

IV. Discussion and Conclusions

Investigation of the nature of a beginning Tai-wanese science teacher's PCK revealed three features.These are integration, relevance, and specificity.Toward the end of the study, Shu-May's thinking andperformance in teaching had more integration ofcontent, students and pedagogy. She also tried to makecontent representation more relevant to the students.

j

Relevancy also influenced Shu-May to match contentteaching with the students' level by teaching simplercontent using repetition. However, Shu-May's think-ing regarding the importance of content relevancy didnot necessary reflect her ability to appropriately presentcontent to the students' level. Thus, there are somediscrepancies between what Shu-May thought and herappropriate teaching performance. In terms of thespecific nature of PCK, Shu-May's knowledge of stu-dents changed from a general view of students' learn-ing to a more specific understanding of their contentlearning ability. Her content knowledge also changedfrom a general view of the physical science disciplineto specific descriptions of what students should knowon each topic. Again, the above features of the de-velopment of Shu-May's PCK revealed the tendencyof Shu-May to think about and improve her teaching.There often is a discrepancy between what teacherssay they would like to do and their teaching perfor-mance, in other words, they lack the functional knowl-edge to transfer what they know into appropriateteaching performance. Science teacher educators needto help preservice and beginning science teachers notonly to think of the important issues related to PCKbut also to help them perform functional PCK in theirteaching situation.

The development of Shu-May's PCK from hersenior year to the first half of her teaching year waswhat Berliner (1988) termed, the "novice stage." Thatis, Shu-May still used certain rules the verificationteaching strategies to teach content, even though shewas aware of the low ability of her students. She triedto use different instructional representations on herstudents but lacked understanding and awareness forjudging the appropriateness of the representations. Onthe other hand, Shu-May's increasing knowledge ofthe students started her to think about the teachabilityof the content; her use of steady and simplified contentteaching strategies in the class were in line with Kagan's(1992) model on the acquisition of knowledge ofstudents, shifting attention to students, and developingstandard procedures. However, Shu-May's constructedknowledge that related to students' science learningwas not appropriate to what really caused the students'science learning problems.

Applying Tuan's (1996c) microscopic views ofthe development of Shu-May' s PCK, her teachingstrategies were organized into introducing science con-cepts from the textbook to help students memorizeconcepts. Toward the end of the study, her mainteaching strategies were the same, but she increasedher use of some strategies to help students memorizeconcepts. Shu-May's instructional representations

149

170


increased in both variety and frequency, such as pro-viding daily-life examples and illustrations to presentscience concepts. Shu-May's knowledge of studentsincreased both in students' science learning charac-teristics and ability. Her thinking and teaching ofcontent became more focused and related to the juniorhigh school physical science textbook. Other featuresof PCK such as curriculum knowledge and knowledgeof assessment did not show obvious development; thismight be related to the data collecting procedureswhich did not address these aspects of data. Futurestudy is needed to address these features of PCK. Interms of the last feature of a PCK -- teacher's knowl-edge of context, it became one of the factors influ-encing Shu-May's PCK development, and is discussedlater.

Shu-May lacked recall of functional pedagogyknowledge learned in college. Her thinking on contentknowledge became focused and more related to thegoals of junior high school physical science textbooks,and her knowledge about students increased in bothstudents' science learning characteristics and ability.These three domains of knowledge integrated togetherinfluenced her in thinking and performance in instruc-tion of the teaching context. Of course, there is nostrong evidence to say Shu-May's pedagogy and contentknowledge disappeared, but it is evident these twodomains of knowledge were not functional when shethought of and performed PCK. One way to think ofShu-May's decrease of thinking content and pedagogyknowledge might be due to her learning experiencein college. She memorized instead of making senseof this knowledge; therefore, it was easy for Shu-Mayto forget concepts and use was likely never functional.

Factors influencing the development of Shu-May's PCK were her perceptions toward science teach-ing and students' science learning, insensitivity injudging appropriate goals of lessons and students'understanding, declining retention of pedagogicalknowledge, and the different cultural norms she per-ceived in different teaching contexts.

Shu-May's PCK revealed her view of the natureof science. Her perceptions of science seemed to bethat of a logical positivist: that growth in knowledgeoccurs through accumulation (Abimbola, 1983). Thisperception reinforced her representation of the scien-tific knowledge in logical sequential ways and her useof the verification method in presenting science con-tent. A modern view of the nature of science, suchas temporary, inquiry and societal nature of scientificknowledge (AAAS, 1993) was not addressed in Shu-May's teaching. She did not present the syntactic andsubstantial nature of the content discipline (Grossman,

Wilson, & Shulman, 1989; Shulman, 1986) to her lowability students, which might have impeded her inusing a variety of ways to represent science contentinstead of simply using lecture style. The currentchemistry teacher education department did not offerexperiences with these aspects of science to beginningscience teachers. The teacher education program couldaddress the nature of science and the syntactic andsubstantive nature of science disciplines in the sciencecontent courses; this might help beginning scienceteachers or future preservice science teachers to gainmore understanding of the nature of science and thenature of content discipline and facilitate the devel-opment of their PCK.

Shu-May's inadequacy in judging the ability ofher students to learn science, such as what studentscan know, what they need to know, and how to helpthem learn science strongly influenced her teaching.Although Shu-May gained some knowledge of stu-dents, she still needed to gain understanding of someessential components that directly influenced the stu-dents' science learning in order to know how to useappropriate ways to teach content. Because Shu-Maycould not diagnose her students' learning ability andneeds, she had difficulty in selecting appropriate lessongoals.

Shu-May's declining retention of pedagogicalknowledge learned in college and her probable lackof functional pedagogical knowledge influenced herto use the same teaching strategies. If Shu-May couldhave remembered the teaching strategies taught in themethods course, then when she faced students' withlow motivation in science learning, she could haveapplied other teaching strategies to facilitate students'learning. We think, although Shu-May heard aboutthe kinds of science teaching strategies that could havehelped her, these teaching strategies were not neces-sarily meaningful for her in a functional way. Shu-May did not have opportunities to practice these teach-ing strategies in teaching contexts. Therefore, she didnot know how successful these teaching methods couldbe of the appropriate ways to use them.

The last factor influencing Shu-May's PCKdevelopment was the culture norms she perceived inher teaching. Feiman-Nemser & Floden (1986) ad-dressed a teacher's socialization, investigating thetransmission of beginning teachers' beliefs, knowl-edge, attitudes, and values to the norms of the teachingculture. Shu-May's college learning experience, es-pecially her pedagogical knowledge, seemed inactivein her school experience, corresponding to the findingsby Zeicher & Tabachnick (1981). Her view of thestudents' ability to learn also influenced her content

150


teaching, supporting previous research that pupils playan important role in influencing a teacher's behavior(Zeichner, 1983). Although Shu-May had pedagogicalknowledge and could verbalize this knowledge incollege, this pedagogical knowledge did not seem tohave real meaning for her in the classroom, nor didit become part of her root values; therefore it waseasily diluted by the teaching culture. However, asFeiman-Nemser & Floden (1986) have indicated,teachers do not have to passively adopt the norms ofthe teaching culture; they can also change the culturebased on their beliefs and practical knowledge.

Therefore, there are two ways to think about thisfinding. One is that when Shu-May or other beginningscience teachers enrolled in the teacher educationprogram, they had experience in memorizing all thefactual knowledge in both the content and pedagogycourses. This factual knowledge was not meaningfulfor Shu-May. Therefore, she did not apply it to herclassroom teaching. It is also possible that coursesoffered in science teacher education programs did notcreate an environment for preservice science teachersto question their previous perceptions on the natureof science and on science learning. Therefore, whenthey taught science in school, they applied what theyhad experienced as high school students, instead ofapplying what they learned in the science teachereducation program.

V. Suggestions

As we learned from Shu-May in the developmentof her PCK and the factors influencing her develop-ment, we generated the following suggestions for futurescience teacher education.

Research supported our finding that beginningscience teachers give more attention to their pedagogyinstead of content (Duffee & A ikenkead, 1992;Lederman & Gess-Newsome, 1991; Tuan, 1993). Therewere many opportunities for Shu-May to improve herscience teaching if she had thought more about thenature of science, and the substantial and the syntacticstructure of the content discipline.

Teacher education programs should address thenature of science and the syntactic and substantivenature of science discipline in the science contentcourses. Teaching methods courses need to addressthe nature of science and science teaching strategiesrelated to the substance of PCK, such as strategies,appropriate instructional representations, curriculumknowledge, ways to diagnosis students' understand-ing, and ways to determine students' learning char-acteristics on different subjects. These emphases would

help future preservice science teachers to develop theirPCK.

As the finding indicated during the internshipexperience, the beginning science teacher became awareof the importance of integrating her content knowl-edge to the students' level of understanding. Hope-fully, this teaching condition will create opportunitiesfor science educators to introduce and to help begin-ning science teachers to construct appropriate PCKbases during their supervision and their regularseminars. Topics could be covered in the internshipseminar that could relate to thinking of different in-structional strategies, appropriate representations, thenature of science related to science teaching, diagnos-ing students' understanding, and overcome students'learning difficulties.

To provide meaningful teaching methods for pre-service science teachers, we need to arouse theirattention to the importance of different domains ofknowledge in order to facilitate their developing PCK.As Feimer-Nemser & Folden (1986) illustrated, themost immediate genesis of the teaching culture is theclassroom context. A teacher's PCK is influenced bythe student-teacher interaction in the classroom. Thus,helping preservice teachers have early field classroomteaching experiences and more discussions on relatingcontent teaching and diagnosing students understand-ing are ways of improving a teacher's PCK develop-ment. Teachers need to have opportunities to practiceand reflect on what they have learned in their methodscourse in order to appreciate the substance of eachteaching strategy and various instructional represen-tations. What has not been learned in a useful waycan not be applied in the teaching context.

If we can educate preservice teachers such asShu-May and help them construct appropriate percep-tions of their content disciplines, the natureof science,and current science teaching strategies, and help themhave a successful experience in learning how to usethese science teaching methods in the classroom, wecan then expect our future teachers to change thescience teaching culture instead of the culture dilutingwhat they have learned in the university. Constructinga professional developmental school where the oppor-tunities, support and freedom to practice an appropri-ate science teaching culture could also help the be-ginning science teachers to have freedom to mastertheir science teaching.

Acknowledgment

This study was funded by the NSC under grant nos. NSC83-0111-S-018-004 and NSC 84-2513-S-018-003. Many thanks to

151

172


the National Science Council for financial support. The first authorwould also like to express her greatest appreciation to Dr. RobertHowe for his suggestions, which improved the quality of the paper.

Reference

American Association for the Advancement of Science (1993).Benchmarks for all science literacy. Project 2061. OxfordUniversity Press.

Abell, S. K., & Roth, M. (1992). Constraints to teaching elementaryscience: A case study of a science enthusiast student teacher.Science Education, 76(6), 581-595.

Abimbola, I. 0. (1983). The relevance of the "new" philosophyof science for the science curriculum. School Science andMathematics, 83(3), 181-192.

Barnette, C. (1991). Building a case-based curriculum to enhancethe PCK of mathematics teachers. Journal of Teacher Educa-tion, 42(4), 263-272.

Berliner, D. C. (1988). The development of expertise in pedagogy.Paper presented at the Annual meeting of the American Asso-ciation of college for Teacher Education, Los Angles.

Bogdan, R. C. & Biklen, S. K. (1992). Qualitative Research forEducation, (2nd ed.). Allyn and Bacon.

Brickhouse, N. W., & Bodner, G. M. (1992). The beginning scienceteacher: Classroom narratives of conviction and constraints.Journal of Research in Science Teaching, 29(5), 471-485.

Brunkhorst, H. K., Brunkhorst, B. J., Yager, R. E., Apple, D. M.,& Andrews, M. A. (1993). The salish consortium for theimprovement of science teaching preparation and development.Journal of Science Teacher Education, 4(2), 51-53.

Clermont, C.P. & Krajcik, J.S., & Borko, H. (1993). The influenceof intensive in-service workshop on pedagogical content knowl-edge growth among novice chemical demonstrators. Journalof Research in Science Teaching, 30(1), 21-43.

Clermont, C. P. & Borko, H., & Krajcik, J. S. (1994). Comparativestudy of the pedagogical content knowledge of experienced andnovice chemical demonstrators. Journal of Research in ScienceTeaching, 31(4), 419-441.

Cochran, K. F., DeRuiter, J. A., & King, R. A. (1993). Pedagogicalcontent knowing: an Integrative model for teacher preparation.Journal of Teacher Education, 44(4), 263-272.

Dana, T. M., & Dana, N. F. (1996). Issues new teachers face whileconsidering science education reform. Paper presented at theNationaLScience Teacher Association. St. Louis, Missouri.

De Jong, 0. (1997). The pedagogical content knowledge of pro-spective and experienced chemistry teachers: A comparativestudy. Paper presented at the National Association for Researchin Science Teaching, Chicago, Oak brook.

Duffee, L., Aikenhead, G. (1992). Curriculum change, studentevaluation, and teacher practical knowledge. Science Educa-tion, 76(5), 493-506.

Feiman-Nemser, S. & Floden, R. E. (1986). The cultures of Teach-ing. In Wittrock, M. C. (Ed.) Handbook of Research on Teach-ing (3rd), 505-526. Macmillan company.

Finley, F., Lawrenz, F., & Heller, P. (1992). A summary of researchin science education-1990. Science Education, 76, 239-254.

Fuller, F. F. & Brown, 0. H. (1975). Becoming a teacher. In K.Ryan (Ed.), Teacher Education (74'h Yearbook of the NationalSociety for the Study of Education, Pt, II, pp. 25-52). Chicago:University of Chicago Press.

Geddis, A. N. (1993). Transforming subject-matter knowledge: therole of pedagogical content knowledge in learning to reflect onteaching. International Journal of Science Education, 15(6),

673-683.Geddis, A. N., Onslow, B., Beynon, C., & Oesch, J. (1993).

Transforming content knowledge: Learning to teach aboutisotopes. Science Education, 77(6), 575-591.

Gess-Newsome, J. & Lederman, N. G. (1993). Preservice biologicalteachers' knowledge structure as a function of professionalteacher education: A year-long assessment. Science TeacherEducation, 77(1), 25-43.

Gilbert, S. W.(1994). NSTA in NCATE: Some results from folioreviews and directions for the future. Journal of Science TeacherEducation, 5(4), 131-138.

Gold, R. L. (1958). Roles in sociological field observations. SocialForces, 36, 217-223.

Grossman, P. L., Wilson, S. M., & Shulman, L. S. (1989). Teachesof substance: subject matter knowledge for teaching. In M.C.Reynolds (Ed.), Knowledge base for the beginning teacher. (pp.23-36). New York: Pergamon Press.

Hashweh, M. Z. (1987). Effects of subject-matter knowledge inthe teaching of biology and physics. Teaching and TeacherEducation, 3(2), 109-120.

Hoz, R., Tomer, Y.,& Tamir, P.(1990). The relations betweendisciplinary and pedagogical knowledge and the length ofteaching experience of biology and geography teachers. Jour-nal of Research in Science Teaching, 27, 973-985.

Kagan, D. M. (1992). Professional growth among preservice andbeginning teachers. Review of Educational Research, 62(2),129-169.

Kennedy, M. M. (1990). A survey on recent literature of teachers'subject-matter knowledge. Issue paper 90-3. NCTRE, Michi-gan State University, East Lansing, Michigan.

Klienfeld, J. (1992). Can cases carry pedagogical content knowl-edge? Yes, but we've got signs of a mathew effect. Paperpresented at the annual meeting of American EducationalResearch Association, San Franciso, CA.

Lederman, N. & Gess-Newsome, J. (1991). Metamorphosis, ad-aptation, or evolution?: Preservice science teachers' concernsand perceptions of teaching and planning. Science Education,75(4), 443-456.

Lederman, N. G., Gess-Newsome, J., & Latz, M. S. (1994). Thenature and development of preservice science teachers' concep-tions of subject matter and pedagogy. Journal of Research inScience Teaching, 31(2), 129-146.

Lederman, N. G. & Latz, M. (1995). Knowledge structures in thepreservice science teachers: Sources, development, interac-tions, & relationships to teaching. Journal of Science TeacherEducation, 6(1), 1-19.

Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. Sagepublications.

Marks, R. (1990). Pedagogical content knowledge: From aMathematical case to a modified conception. Journal ofTeacherEducation, 41(3), 3- I 1.

McDiarmind, G. W., Ball, D. L., & Anderson, C. W. (1989). Whystaying one chapter ahead doesn't really work: subject-specificpedagogy. In M.C. Reynolds (Ed.), Knowledge based for thebeginning teacher (pp. 185-192). Oxford: Pergamon.

Miles, M. B. & Huberman, A. M. (1994). Qualitative data analysis.Sage company.

Mitchell, D, Ortiz, T, & Mitchell, T. (1982). Final report oncontrolling the impact of rewards and incentives on teacher taskperformance. Riverside: University of California.

Ormrod, J. E. & Cole, D. B. (1996). Teaching content knowledgeand pedagogical content knowledge: A model from geographiceducation. Journal of Teacher Education, 47(1), 37-42.

152


Sanford, J. P. (1988). Learning on profession development ofbeginning science teachers. Science Education, 72(5), 615-624.

Schtin, D. A. (1986). Educating the reflective practitioner. Jossey-Bass Inc., Publishers.

Shulman, L. S. (1986). Those who understand: knowledge growthin teaching. Educational Researcher, 15(1), 4-14.

Shulman, L. S. (1987). Knowledge and teaching: Foundationsof the new reform. Harvard Educational Review, 57(1),1-22.

Thorley, N. R. & Stofflett, R. T. (1996). Representation of theconceptual change model in science teacher education. Science

Education, 80(3), 317-339.Tippins, D., Nichols, S., & Tobin, K. (1993). Restructuring science

teacher education with communities of learners. Journal ofScience Teacher Education, 4(3), 65-72.

Tuan, H. L.(1993). A case study of a preservice chemistry teachers'thinking. Chinese Journal of Science Education, 2(2), 115-141.

Tuan, H. L. (1996a). Exploring the nature of preservice chemistryteachers' pedagogical content knowledge case studies. Pro-ceedings of the first conference on math and science teachingand teacher education conference, Changhua, Taiwan, p. 144-177.

Tuan, H. L.(1996b). Investigating the nature and development ofPreservice chemistry teachers' content knowledge, pedagogicalknowledge, and pedagogical content knowledge. Proceedingsof the National Science Council, Part D: Mathematics, Science,& Technology Education, 6(2), 101-112.

Tuan, H. L. (1996c). A revelation of pedagogical content knowledgeon future science teacher education. Proceedings of the firstconference on math and science teaching and teacher educationconference, Changhua, Taiwan, p. 118-142.

Yang, J. H. & Wu. I. L.(1996). A beginning biology teacher'sknowledge of students - A case study. Paper presented at theannual meeting of the National Association for Research inScience Teaching. St. Louis, Missouri, USA.

Zeichner, K. M. (1983). Individual and instructional factors relatedto the socialization of beginning teachers. Paper presented atthe conference "First year of teaching: what are the pertinentissues?". University of Texas, R & D Center for teacher edu-cation, Austin.

Zeichner, K. M. & Tabachnick, B. R. (1981). Are the effects ofuniversity teacher education washed out by school experience?Journal of Teacher Education, 32(3), 7-11.

Appendix 1: Interview Protocol

Interview questions at the beginning and at theend of the semester:1. What does an excellent science teacher mean to

you? How do you become an excellent scienceteacher ?

2. How did you become a teacher? What are yourfuture plans?

3. What kind of teacher do you think you are? Youcan use a metaphor to explain your answer.

4. What does teaching mean to you? What issues inteaching do you think of often? What issues aboutstudents do you think of often?

5. What do you think of often about chemistry? Couldyou express your thinking about chemistry to me?

6. What do you think of often about physics? Couldyou express your thinking to me ?

7. What do you think of often about physical science?Could you express your thinking about physicalscience to me?

8. What are your pedagogical beliefs? What does yourteaching look like in an ideal situation?

9. What is your perception of students' physical sci-ence learning? Do you have any strategies forchanging students' learning habits?

10.What are your expectation on how students shouldlearn?

11.What are the goals students need to reach aftertaking your physical science courses?

Interview questions before each class teaching:1. Could you explain the content of this unit? Please

explain in detail the knowledge you have on thistopic.

2. How do you teach this unit? What's your teachingflow?

3. How do you prepare this unit? What factors did youconsider? What kinds of difficulties might youface? How do you plan to solve these difficulties?

4. What do you expect students to learn from this unit?Why?

5. Do you have other ways of presenting this unit?

Interview questions after each classroom teach-ing:1. What do you think was covered in this unit? What

change would you like to made about the responsesyou provided in previous interview?

2. How do you view your own teaching?3. How did students learn in this unit? How did you

help them?4. Do you have other ways of presenting this unit?

153


fillA1feNN=jfklAW*4tc*UMZWM3f A

&[44 * Ag6-it **

*MAZMOVT.t**31-*fttliJMW

**1":;*thlfit.1211=1

*WEnROA100tWIttfttIgatKi*MMR,A0M043ft**WALA*04ffAMAMVM,RMEAEMin*4ft*OM.iffARWOWJFA0 **ftVO'OVAY43(411kV0 FAAMMil,difWMW4ntiNft#WAYMWM,tH_LAMMIE..*g5)4R±WU514figaM&.?ff910METTC-,4*114J*4,1*Wffitl*X olft*mmn*.mroit*m IWKAJONAOAMM44igit0 WIZOt*VikM,1=10AA'AVAM 'tMVA'*-01D116Am7T-i4A.WAM*4ft*OMORIgM#AVAPP'01AvntiVit.AffAM*4ftnft*EAECEOLAWfa**4-ft**43MU 4*40MA*--tWIINIONA0 *Aft.*MRYJVC5t144:NiCit'fAiALLOM*Pf*tWVATVft0 kftM*431-0MVIM04.031-MVittMAA,M*AA4VJEtYAMWITMnIAN0 A*,WWMX*4ft4-40MRWZIAMAV**V*AV*t.00t,tttiVeloNVaA*A-MXPIFINT%tA,MAKA.tft*OMMO,R*EROM*30-EV1+1004AMI,bAP-03t*IRIE4*0-6**WtitM°

154


Identification of the Essential Elements and Developmentof a Related Graphic Representation of Basic Concepts in

Environmental Education in Taiwan

Ju Chou

Graduate Institute of Environmental EducationNational Taiwan Normal University


(Received December 18, 1996; Accepted May 12, 1997)

ABSTRACT

The objectives of the study were: 1. to identify the essential elements of the basic concepts inenvironmental education for Taiwan; and 2. to develop an easily-understood graphic model to expressthe interrelationships among the identified essential elements. This study's target population (N=1398)consisted of the Society of Science Education, the National Park Society, the Society of EnvironmentalProtection, and many other concerned individual subscribers of the Quarterly Journal of EnvironmentalEducation. The Q-Sort technique was used to collect research subjects' attitudes toward environmentalconcepts. A group of experts assisted in establishing the instrument's face and content validity. A Test-Retest method was used to decide the reliability of the Q-set. A Q-set including 44 concepts was established(mean of Pearson's r=.55). The research instrument was sent by mail to 280 systematically randomlyselected samples. A total of 196 (70%) research subjects responded to the survey. R-factor analysiswas used to analyze the Q-Sort data response. Six underlying essential elements were extracted andidentified. They are: 1. Interaction and Interdependence, 2. Resource Conservation, 3. EnvironmentalEthics, 4. Environmental Management, 5. Ecological Principles, and 6. Carrying Capacity and Qualityof Life. A graphic representation describing the interrelationships among the identified elements wasalso developed via two rounds of discussion involving an interdisciplinary panel of seven experts inscience education, environmental protection, conservation, and environmental education.

Key Words: environmental education, concept, element, Q-Sort technique

I. Introduction

The development of appropriate and effectiveenvironmental education programs is one of the mostimportant efforts being made by both the governmentand society in tackling environmental problems inTaiwan. To convey the entire picture of the status ofthe environment to the general public and students,the decision makers and planners of environmentaleducation programs should have a clear idea of themajor foci and directions for their programs beforethey formally launch extensive efforts. Similar effortshave been made by western environmental educatorssuch as Entwistle (1981), Stapp (1982), & Roth (1989).Therefore, there is a necessity to provide a simple andvivid graphic representation that can reveal the inter-

relationship among the essential elements underlyingbasic concepts (Chou, 1994; Chou & Roth, 1992) usedin environmental education in Taiwan. With a rep-resentation, environmental education programmers anddecision makers can grasp the major directions andincorporate them into their education programs moreeasily. Moreover, this effort can help environmentaleducation practitioners conduct and implement envi-ronmental education programs more efficiently andeffectively.

To develop environmental education at either theformal or non-formal level has become a basic stepfor an integrated effort made by any country intackling various environmental problems. It is alsobelieved that the development of environmental edu-cation in different countries will "require the devel-

155

176

J. Chou

opment and teaching of the environmental philosophyand related concepts at every point in the formal andnon-formal education process" (Schmieder, 1977).However, before meaningful approaches can be ini-tiated to educate children and adults about environ-mental matters, questions such as "what should themain focus and major concerns of environmentalprograms be?" and "what should students know aboutthe environment?" must be clearly answered (Chou,1992; Townsend, 1982; Yang et al., 1988).

Basically a concept is a way of grouping objectsor events in terms of essential similarities. However,concepts are not evaluative (Entwistle, 1981), espe-cially abstract concepts related to the environmentwhich are very difficult to define clearly by simpleconcepts. Thus, learners have to "build up their ownunderstanding, which thus acquire a personal mean-ing" (Entwistle, 1981) of those concepts relating tothe total environment. Therefore, it is necessary toidentify the most essential elements beyond thoseenvironmental concepts for appropriate use in envi-ronmental education. For example, tremendousprogress has already been made in the U.S. in iden-tifying the basic concepts that can be used in envi-ronmental education at different levels (Allman, 1972;Ballard, 1989; Brennan, 1986; Ronfeldt, 1969; Roth,1969; Visher, 1960; White, 1967). Moreover, knowl-edge of the underlying structure beyond those con-cepts has also been accumulated (Bowman, 1972; Chou,1992; Isabell, 1972; Townsend, 1982). However, whenapplying the research results in developing practicalenvironmental education, the specific conditions indifferent countries, such as the economy, politics,culture, etc., should be carefully considered (UnitedNations Educational, Scientific, and Cultural Organi-zation [UNESCO], 1990).

The necessity of developing environmentaleducation in Taiwan did not become an important issueuntil the late 1980s. Accompanied by the rapiddevelopment of the economy, environmental degrada-tion gradually emerged and attracted the attention ofthe whole society. In addition to large scale pollutionabatement measures launched by the government, acomprehensive research program on environmentaleducation was launched by the National ScienceCouncil (NSC) of the Republic of China (R.O.C.) inJuly 1987 (Wang et al., 1987). Many studies have beenundertaken in Taiwan since then (NSC, 1989; 1990;1992). However, very little effort except Chou's (1992)study has been directed toward determining the essen-tial elements that are appropriate for the environmentof Taiwan (Chou, 1989; NSC, 1990). It is also be-lieved that with a simple and easily-understood rep-

resentation, environmental educators should be ableto get the whole picture of the concerns related toenvironmental education. Although some studies haveattempted to identify basic concepts in environmentaleducation (Chou, 1989; NSC, 1990; 1992), researchspecifically focusing upon identifying the essentialelements of environmental education and its relatedgraphic representation have been very rare in Taiwan.Such information is essential for environmental edu-cation related instruction, teacher training, curriculumdevelopment, program planning, and policy formula-tion in Taiwan.

1. Statement of the Problem

This study is focused on understanding the per-ceptions that are held by experts in and practitionersof environmental protection, conservation, scienceeducation, and environmental education in Taiwanregarding the basic concepts that are appropriate forenvironmental education. The question posed in thisstudy was: What are the essential elements and theirrelated graphic representation appropriate for environ-mental education (K-16) as perceived by randomlyselected samples of concerned professionals in Tai-wan?

2. Objectives of the Study

The objectives for this study were:(1)To identify the essential elements underlying the

important environmental concepts as perceived byexperts in environmental protection, conservation,science education, and environmental education inTaiwan.

(2) To develop an easily-understood graphic represen-tation describing the interrelationships among theidentified essential elements.

II. Methods

1. Population and Sampling

The target population of this study was profes-sionals who were concerned about the environmentespecially with regard to education. Owing to the greatdifficulty in identifying all such individuals in society,names listed on the directories of The Society ofEnvironmental Protection, the National Park Society,the Society of Science Education , and individualsubscribers of the Environmental Education Quarterlywere used as the accessible population to form thesampling frame (n=1398). A systematic random

156

177

Elements of Environmental Education Concepts

sampling method was adopted to select samples frOmthe frame. Two hundred and eighty subjects wererandomly selected for participation in this study.

2. Instrumentation

After extensive review of the related researchesand considering the purpose of the study, it was decidedthat a structured type of Q-sort that used forced sorting(a normal distribution with a specified number of cardsin each pile) and R factor analysis could appropriatelymeet the needs of this study. Several interdisciplinarypanels of experts with expertise in environmentalmanagement, conservation, environmental education,and science education were consulted during the re-search instrumentation stage and assisted the researcherin developing the research instrument. At the begin-ning, six interdisciplinary review panels from the areasof environmental management, conservation, andenvironmental education were asked to help solidifythe initial concept base which was originally devel-oped by the researcher (containing 57 concepts).Through discussion; the panelists reached agreementon reducing the number of concepts in the pool from57 to 50. Afterwards, a pilot test was conducted. Atthis stage, eleven experts from different fields but allconcerned about the environment were asked to usethe Q-sort technique to sort the cards twice (with aone week break in between). Six concepts which hadlow or negative values of the calculated Pearson's rwere deleted from the original Q set. The final Q setcontaining 44 concepts was used in the formal researchinstrument during the following data collection stage.Experts also helped to review the content of thequestionnaire and the sorting instructions. Throughthe field test and pilot test (test and retest method wasused) mentioned above, the face validity, content va-lidity, and reliability (mean of Pearson's r=.55) wereestablished. The basic components of the researchinstrument used in the data collection stage included:1. a complete Q set (44 different environment relatedconcepts, each printed on a card), 2. clearly stated andeasy to follow sorting instructions, and 3. a question-naire inquiring about the background of each respon-dent.

3. Data Collection and Analysis

The data collection process of the survey wasconducted following Dillman's (1978) suggested stepsin the Total Design Methods (TDM). All the mailingswere sent out by regular first class mail. Totally, threefollowups were sent to the subjects in order to increase

the response. One week after the first mailing, thefirst followup was delivered. A two-week break wasallowed between the later followups. Totally, 196subjects (70% of the samples) responded to the surveyand returned their research packet.

Research subjects were asked to follow the sort-ing instructions to sort the 44 small cards in the Q set.The easily followed sorting instructions were a veryessential tool in order for obtaining data which couldbe analyzed and were necessary to meet the criteriaand satisfy the purpose of the study. By consultingrelated researches (Chitwood, 1977; Chou, 1992;Townsend, 1982), the researcher of this study devel-oped sorting instructions which were used in the datacollection stage. The instructions directed the re-search subjects to manipulate and sort the 44 cards inthe Q set. These cards were to be placed into sevendifferent piles (labeled 0 to 6) in terms of their relativeimportance ("6" being the most important pile, "0"being the least important pile) for environmentaleducation, with a specified number of cards in eachpile. One key question they were to ask themselveswhen they sorted the cards was: "What should ourstudents know about the environment?" The score foreach statement was calculated according to the nu-meric label of the pile to which it belonged.

The major analytical instrument which was usedin the data analysis stage was a statistical packageprogram, used on a Macintosh computer, calledSYSTAT (Version 5.2 SYSTAT, Inc.). R Factoranalysis with a VARIMAX rotation on the data col-lected from the samples was used to obtain essentialelements.

Ill. Results

1. Factor Analysis Results

The collected responses of the Q-sort results (thescores of each statement) were used as the basis forR factor analysis. In order to get a significant factorthrough the calculation process, several standards wereconsidered and applied in selecting the significantfactor (Chou, 1992). First, to be a significant itemin a factor, its factor loading had to be equal to orgreater than three times the standard error of a zerocorrelation, i.e., 3Nr7 where n is the number of itemsin the Q set (Schlinger, 1969). That is, for this study,the value of a significant factor loading had to be equalto or greater than 3//7-14 . Thus, a decision was madeto use .45 as the minimum standard for this study todetermine whether a factor loading was significant.Second, a significant factor had to have at least two

157

7 8

J. Chou

or more statements. The third consideration was theproportion of the total variance explained by the lastfactor to be retained. Researchers have set one, five,or ten percent as the standard in different studies, andit was further stated that "...one may set the criterionat whatever level is considered substantively impor-tant" (Kim and Mueller, 1978). Four percent wasestablished as the minimum for this standard. Thefourth consideration was the criteria of interpretabil-ity, which is a crucial step (Kim and Mueller, 1978).Using factor analysis with an orthogonal (Varimax)rotation on data obtained from Q-sort responses ofresearch subjects, six factors were identified and treatedas essential elements met the first objective set forthfor this study.

Overall, six factors were extracted from the factoranalysis. These factors were then treated as essentialelements underlying the 44 environmental concepts inthe Q set. Altogether, they could explain 35.0 percentof the total variance (Table 1).

The six extracted factors and the correspondingconcept statements with rotated factor loadings areshown in the following tables.

Table 2 provides information about the conceptstatements and corresponding rotated factor loadingsof Factor I. Factor I could explain 9.4 percent of thetotal variance.

Table 3 provides information about the conceptstatements and corresponding rotated factor loadingsof Factor II. Factor II could explain 7.07 percent ofthe total variance.

Table 4 provides information about the conceptstatements and corresponding rotated factor loadingsof Factor III. Factor III could explain 5.60 percentof the total variance.

Table 5 provides information about the conceptstatements and corresponding rotated factor loadingsof Factor IV. Factor IV could explain 4.83 percentof the total variance.

Table 6 provides information about the conceptstatements and corresponding rotated factor loadings

Table 2. Factor I and the Corresponding Concept Statements withRotated* Factor Loadings

ConceptRotated

Factor Loading

40. All living organisms have the right to liveon earth and compete to survive. This rightcan not be denied, no matter whether theycan be uscd by humans or not. -.58

10. During different kinds of human settlementand development, human have the responsibilityto maintain the living environment in order toupgrade the welfare of all people. .54

9. Humans are also a part of the nature. We arenot superior to other living organisms. -.53

17. Responsibility for conservation should beshared by individuals, business and industries,special interest groups, and all levels ofgovernment and education. .51

23. Living things are interdependent on each otherand their environement. -.49

41. Energy, its production, use, and conservation,are essential in the maintenance of our society. .47

*Varimax rotation

Table 3. Factor II and the Corresponding Concept Statements withRotated* Factor Loadings

ConceptRotated

Factor Loading

39. Natural resources are limited and shouldnever be wasted.

33. Most natural resources are vulnerable todepletion in quantity, quality, or both.

6. To manage environmental pollution, preventionbeforehand is more effective than compensationand treatment afterwards.

25. Natural resources affect and are affected by thematerial welfare of a culture and directly orindirectly by philosophy, religion, government,and the arts.

28. The natural environment is irreplaceable.

.59

.54

.54

-.51.47

*Varimax rotation

Table 1. Results of Factor Analysis of the Q-sort Resultsof Factor V. Factor V could explain 4.09 percent ofthe total variance.

Table 7 provides information about the conceptstatements and corresponding rotated factor loadingsof Factor VI. Factor VI could explain 4.01 percentof the total variance.

2. Naming the Essential Elements

Obtaining six essential elements which wereextracted from the 44 environmental concepts in the

Factor EigenvalueAdded (%)Variance

Coumulative Percent (%)of Total Variance

IIIIIIVVVI

4.133.112.472.131.801.77

9.407.075.604.834.094.01

9.4016.4722.0726.9030.9935.00

158

179


Table 4. Factor III and the Corresponding Concept Statementswith Rotated* Factor Loadings

ConceptRotated

Factor Loading

4. We are ethically responsible to other individualsand society, and to that larger community, thebiosphere.

11. Humans have a moral responsibility for theirenvironmentally related decisions.

35. Forests are an important part of the globalecosystem that supports us and of which weare a part. Deforestation (clearing an area ofall trees) will cause an immediate loss ofwildlife habitat and natural resources. Long-termeffects may include desertification and climatechange.

-.69

-.65

.55

*Varimax rotation

Table 5. Factor IV and the Corresponding Concept Statementswith Rotated* Factor Loadings

ConceptRotated

Factor Loading

30. The management of natural resources to meetthe needs of successive generations demandslong-range planning.

1. Humans should play a role in understandingnature and to cooperating with otherconstitutents of nature.

-.62

-.60

*Varimax rotation

Table 6. Factor V and the Corresponding Concept Statements withRotated* Factor Loadings

ConceptRotated

Factor Loading

3. In any environment, one component such asspace, water, air or food may become alimiting factor.

13. The environment is the sum of all externalconditions and influences affecting organisms.The environment may be divided into biotic(living) and abiotic (non-living) components.

-.63

-.56

*Varimax rotation

Q set was only half the work required to reach ob-jective one set forth for this study. In order to makethe results of factor analysis meaningful, the secondstep--a very crucial and essential step--was to giveeach identified essential element a name. Withmeaningful names given to the identified elements,researchers and environmental educators could then

Table 7. Factor VI and the Corresponding Concept Statementswith Rotated* Factor Loadings

ConceptRotated

Factor Loading

37. There is a maximum human populationcorresponding to each resource base. Thepopulation cannot exceed this level if asatisfactory standard of living for allpeople is to be maintanined. -.66

44. Increasing population and per capita use ofresources have brought about changedland-to-people or resource-to-population ratios. -.65

27. Family planning and the limiting of family sizeare important if overpopulation is to be avoidedand a reasonable standard of living assured forfuture generations. -.59

26. Humans should never do whatever they want.Everything we do will have an unexpectedinfluence on other people and living organimsmeither now or in the future. Therefore, humansshould always use environmental ethics as abasis for making value judgments. .48

*Varimax rotation

communicate effectively and efficiently with targetaudiences about the major concerns/directions ofenvironmental education in terms of the content core.Five professionals from the areas of the environment,education, and conservation were asked to provideassistance during the factor-naming stage. Two roundsof written discussion among panelists were adoptedat this stage. They carefully looked into the meaningsand implications of the concepts of each element andexpressed to each other their opinions about the ap-propriate names and their meanings. After two roundsof exchanging and sharing opinions, they finallyreached a consensus on the names of the essentialelements. They were named as follows:

(1) Interaction and Interdependence;(2) Resource Conservation;(3) Environmental Ethics;(4) Environmental Management;(5) Ecological Principles; and(6) Carrying Capacity and Quality of Life.Furthermore, the identified essential elements'

implied meanings were listed and discussed as fol-lows:1. Interaction and Interdependence

Within the environment, living and non-living el-ements always interact and are interdependent witheach other. Moreover, living organisms, humans,and even different segments of the human societyare interacting and interdependent with each other.Therefore, we have to understand the unique char-

159

t8 0

J. Chou

acteristics of interaction and interdependence ex-isting in the environment.

2. Resource ConservationHuman society and its cultural development areclose tied to natural resources. Since the naturalresources on earth are limited, humans have to usethem carefully by considering the sustainability ofnatural resources.

3. Environmental EthicsHumans are also members of the whole ecosystem.Therefore, we have to have a positive attitude, properbehavior, and a sense of responsibility with respectto the earth and all other living organisms.

4. Environmental ManagementHuman welfare depends upon the appropriatemanagement and maintenance of the environment.We have to use the environment carefully in orderto avoid damaging the earth and harming futuregenerations.

5. Ecological PrinciplesUnderstanding the nature and principles of ecologyis essential to foster positive attitudes and actionsamong people to maintain the proper balance, sta-bility, and integrity of the ecosystem.

6. Carrying Capacity and Quality of LifeThe size of the human population and human be-havior have direct influence on natural resources,environmental quality, and the quality of life.Therefore, we have to adequately regulate the sizeof the human population and control humanbehavior's impact on the environment.

3. Developing the Graphic Representation

The second major focus of this study was todevelop a simple and easily-understood graphic rep-resentation which could be used as a communicationtool to help concerned educators understand the majorconcerns of environmental education and the interre-lationships which exist among them. It was hoped thatthrough the model's graphic representation of the in-terrelationships existing among the identified under-lying essential elements, environmental education pro-grammers and policy makers could then grasp thewhole picture of the major concerns of environmentaleducation easily. In order to reach this goal, theresearcher invited seven experienced professors fromthe areas of environmental management, environmen-tal education, conservation, and education to form aninterdisciplinary panel to participate in a panel dis-cussion on the model-development stage in this study.During this stage, each panelist fully expressed his orher ideas on the paper in and returned them to the

researcher by mail without meeting the other panelistspersonally. The researcher was the one responsiblefor compiling all the ideas and feedback then sendingto each panelist to serve as a basis for next round ofdiscussion. Each expert was given all of the writteninformation about the results of factor analysis forreference. Experts were asked to draw a diagram onpaper to illustrate the interrelationships that existedamong all the identified essential elements. They werealso asked to write down their explanation of themeaning expressed by the representation and theirreasons for drawing it the way they did. All the resultswere integrated and compiled by the researcher andprovided to each expert in the panel. After two roundsof extensive written discussion, the panelists finallyreached on agreement concerning the graphic repre-sentation. This is shown as follows (Fig. 1).

IV. Discussion

Apparently, the results of this research essen-tially reveal the need for a comprehensive and sys-tematic understanding of the entire environment whichemphasizes learning about all the facets of the envi-ronment and related issues, and does not only focuson teaching and learning compartmental and partialknowledge of the environment.

The meanings conveyed by the essential ele-ments and graphic representation may be directedtoward the development of a basic understanding ofthe earth's environment and related ecological con-cepts. The developed model emphasizes that under-standing should be based upon the characteristics ofinteraction and interdependence existing among theliving organisms, the environment, humans, and dif-ferent sectors of society. Humans should also estab-lish a new ideology that includes a positive attitude,value judgments and treatment of the earth and all

teractionInterde endence

ResourceConservation

EnvironmentalManagement

EnvironmentalEthics

CarryingCapacity

Quality of Life

Ecological principles

Fig. 1. Graphic representation of the essential elements of hte en-vironmental education related concepts.

160

181


organisms with equality and respect. Only through theintegration and incorporation of the knowledge, atti-tudes, and values mentioned above can we then effi-ciently use limited natural resources by consideringthe earth's sustainability, and manage and solve en-vironmental problems in an effective manner. Onlyby solving environmental problems at their root causesand lessening humans' negative effects on the envi-ronment can the quality of the environment be main-tained. In this way the end goal of achieving a betterquality of life can be reached.

V. Implications and Recommendations

Corresponding to the processes, technique, andresults of this study, several implications and recom-mendations for environmental education, future study,and other researchers are given as follows:1. The six essential elements identified in this study

are a very important means of clarifying the majorfocus of environmental education development.They imply that the major task of environmentaleducation is to investigate all facets of the environ-ment instead of being limited to one or two direc-tions' narrow perspective.

2. The graphic representation developed here can beprovided as valuable reference for the developmentof future environmental education curriculum andinstructional materials for different age groups. Theessential elements in the graphic representation canbe used as a major axis leading toward the directionof developing different levels' environmental cur-riculum and instructional materials.

3. The findings of this study may be provided to theconcerned governmental agencies and non-govern-mental organizations as a good reference for thedetermination of the major directions of environ-mental education efforts such as environmentaleducation policy planning, program planning,teacher training (both pre- and in-service training),and curriculum development.

4. Because of the limited time and resources available,it is suggested that use more diverse groups ofpeople as the population. Further study shouldconsider different target groups such as:a. government officials in education and environ-

mental conservation;b. school teachers (primary and secondary levels);c. representatives of non-governmental organiza-

tions;d. representatives of industry.

5. Further study may also use Q-factor analysis (treat-ing the subjects as the independent variables) to

analyze the data in order to understand any differenttypes of attitude underlying the research subjects'views concerning the environment.

6. Further research can incorporate on investigationof the association between the attribute variablesand the elements identified into the research designin order to provide more information on the fieldof environmental education.

7. Comparing the normal distribution method used inthe forced-sorting Q-sort process of this study tothe rectangular distribution Q-sort method used inan earlier similar study (Chou, 1992), the researcherdiscovered that there is no great difference betweenthe two kinds of method when considering thesurvey's response rate. Similar future research cantake either normal or rectangular distribution as thesorting method in the Q-sort process.

Acknowledgment

The author would like to express appreciation to Dr. Kuan-Jen Yang of National Taiwan Normal University for his full supportand encouragement during this study. Thanks also go to all respon-dents and panelists, and also the National Science Council, R.O.C.for its financial support for this study.

References

Allman, S. A. (1972). Identification of environmental concepts forinclusion in an elementary school curriculum. DissertationAbstracts International. 33, 2137B (University Microfilm No.72-27, 378, 121).

Ballard, M. (1989). Defining the universal in ititernational en-vironmental education through a content database. Master the-sis, State University of New York, College of EnvironmentalScience and Forestry, Syracuse, New York.

Bowman, M. L. (1972). The development and field validation ofan instrument to assess college students' attitudes toward thedeterminants of environmental issues. Doctoral dissertation,The Ohio State University, Columbus, Ohio.

Brennan, M. J. (1986). A curriculum for the conservation of peopleand their environment. The Journal of Environmental Educa-tion, 17(4), 1-12.

Chou, C. H. (1989). Environmental education researches fundedby the National Science Council: A status report (in Chinese).Taipei, Taiwan: National Science Council.

Chou, J. (1994). Major concerns of environmental education in theRepublic of China. In: R. Mrazek (Ed.),Pathways to partner-ships: Coalitions for environmental education , Proceedings ofthe 22nd Annual Conference of The North American Associationfor Environmental Education, September 23-27, Big Sky,Montana, U.S.A. (pp. 220-228). Troy, Ohio: The North AmericanAssociation for Environmental Education.

Chou, J. (1992). A determination of the underlying constructs ofbasic concepts in environmental education as perceived byfaculty of The Ohio State University and National TaiwanUniversity. Doctoral dissertation, The Ohio State University,Columbus, Ohio.

Chou, J., & Roth, R. E. (1992). A determination of the underlying

161

182

J. Chou

constructs of basic concepts in environmental education. Paperpresented at A World Congress for Education & Communicationon Environment & Development, October 16-21, Toronto,Canada.

Dillman, D. A. (1978). Mail and telephone survey: The total designmethod . New York: John Wiley & Sons.

Entwistle, N. (1981). Concept formation and intellectual develop-ment. In Style of learning and teaching (pp. 161-177). NewYork: John Wiley and Sons.

Isabell, L. 0. (1972). Identification of concepts that will serve asa basis for development of environmental education instruc-tional materials. Dissertation Abstracts International, 33, 3424A(University Microfilm No. 73-56, 103).

Kim, J. 0., & Mueller, C. W. (1978). Number of factors problemrevisited. Factor analysis: Statistical methods and practicalissues (pp. 41-45). Beverly Hills, California: Sage Publica-tions.

National Science Council (1989). Mid-range plan on environmentalprotection and technology: The demand of research related toenvironmental protection. Science & Technology of Environ-mental Protection (in Chinese), 2(3), 29-30. Taipei: NationalScience Council.

National Science Council (1990). Abstracts of the researchesfunded by the National Science Council in 1989/1990 on en-vironmental education (in Chinese). Taipei, Taiwan: NationalScience Council

National Science Council (1992). Abstracts of the researchesfunded by the National Science Council in 1990/1991 on en-vironmental education (in Chinese). Taipei, Taiwan: NationalScience Council.

Ronfeldt, L. L. (1969). A determination of basic urban environ-mental understandings important for inclusion in the elemen-tary school curriculum. Dissertation Abstracts International,30, 3649A (University Microfilm No. 70-5314).

Roth, R. E. (1969). Fundamental concepts for environmentalmanagement education (K-I6). Doctoral dissertation, Univer-

sity of Wisconsin, Wisconsin, Madison.Schlinger, M. J. (1969). Cues on Q-Technique. Journal of

Advertising Research, 9(3), 53-60.Schmieder, A. A. (1977). The nature and philosophy of environ-

mental education: Goals and objectives. In: UNESCO (Ed.),Trends in environmental education (pp. 23-34). Paris, France:United Nations Educational, Scientific, and Cultural Organi-zation.

Stapp, W. B. (1982). An instructional model for environmentaleducation. In: J. J. Kirk (Ed.), Facets and faces of environ-mental education (pp. 81-100). Branchville, New Jersey: NewJersey School of Conservation.

Townsend, R. D. (1982). An investigation into the underlyingstructure of the domain of environmental education concepts.Doctoral dissertation, The Ohio State University, Columbus,Ohio.

United Nations Educational, Scientific, and Cultural Organization(1990). Planning environmental education at the national level.Connect, 1-2.

Visher, H. H. (1960). A determination of conservation principlesand concepts desirable for use in the secondary schools.Dissertation Abstracts International, 20, 4005 (UniversityMicrofilm No. DA6000312).

Wang, S., Lu, K. Y., Chou, C. H., Yen, H. W., Kuo, Y. W., & Yang,K. S. (1987). The scope and framework of research in en-vironmental education: A comprehensive plan (in Chinese).Taipei, Taiwan: National Science Council.

White, R. C. (1967). A study associating selected conservationunderstanding with available community resources for gradefour through twelve, Dissertation Abstracts International, 28,1638A (University Microfilm No. DA6713474).

Yang, K. J., Shih, T. T., Yang. K. S., Chou, J., Lu, K. Y., Huang,C. C., & Fang, C. C. (1988). A plan for implementing envi-ronmental education in Taiwan (in Chinese). Taipei, Taiwan:Environmental Protection Administration.

162

3


1001111(III_VALti143:41444EICKA-4111NIZIAM1Mt

1U1:11±N. ts4--4-3RIttt-Vf9-tfi

#0 N

2PifF9.-En reJAIVTOMM/IftRIAIRiD191R411-14fft: 1-0421KW* ( essential elements ) Ak#TAUg-4-g:PATMM*INA*.W51,IWRYON(survey)frAMaAt*MANNPRAJA,M*

f4,64cttv* igm/L.; ** igitg-g**wsmitwoRltm-g-N04144* Q-Sort Technique*iikZ1n4VIIII#A0Wkag'lit#14*figMa3Klik*YeA40434*Wii/goTWAIAQsetfait 4*_Sgq----0*-fiTAfficl1 (face validity ) ( content validity ) Itg pilot testtRII-AAYJAPk 44 11M4itt&Atfilf,I1Uka (Pearson's-fA =0.55 ) ° VFAIAA111titiSiMag*e iUUii#44ffix4zi-t*A-m- 280 rik4it$4 70% 0 1:1113043**-1-1V11 RHfrI ( R-factor Analysis ) 4.491a4C192:tia21Kw* , I. 11,yq.tA ( Interaction and Inerdependence ) 2 . Tin 1143 ig ( ResourcesConservation ) , 3. M1143{I ( Environmental Ethics ) , 4. Mira W ( Environmental Management ) 5. lag ( Ecological Principles ) 1),J/A 6. *gtxtitrm mlig-m fgt. IgItt-4-4-ffirt<J4vligirgm*ona,ff WTWalau,ffl*Air:A%MMIASIATI4M4AKW*MMnflif.

163

484


A Beginning Biology Teacher's Professional Development-- A Case Study --

JoNG-HsIANG YANG AND I-LIN WU

Department of BiologyNational Taiwan Normal University

Taiwan, R.O.C.

(Received April 26, 1997; Accepted September 12, 1997)

Abstract

The purpose of this case study was to describe and identify if knowledge of students was an importantfoundational component of PCK (Pedagogical Content Knowledge) for a beginning biology teacher.Interpretive research methods described by Erickson (1986) were applied in this study. Multiple datagathering methods, both qualitative and quantitative, were used to collect data from 55 consecutive periodsof classroom observation across a complete academic year. Data obtained and used included field-notes,after-class interviews, classroom interaction analyses, and student questionnaires. Triangulation tech-niques were employed to enhance the validity of the findings. Data indicated that the novice teachercould create 'an active classroom environment, but that her poor knowledge of students prevented herfrom managing more effective communication in the classroom. In the instructional activities, the teacheralways failed to find appropriate analogies, metaphors, and representations for students' constructionof knowledge. For this novice teacher, knowledge of students seemed to be an important foundationalcomponent; her lack of this knowledge limited her PCK growth.

Key Words: beginning biology teachers; knowledge of students; professional development; pedagogicalcontent knowledge; secondary school science; junior high schools

I. Introduction and Background

What does it mean to be a qualified beginningteacher? In Taiwan, we still know very little aboutthe differences between effective beginning teachersand effective experienced teachers. In the early his-tory of science education in this country, teachereducation only emphasized knowledge of subjectmatter. During the past few decades, student teachershave been told to put emphasis on the effectivenessof general pedagogical methods, such as questioningskills, wait time, small-step questioning, inquirytraining, open-ended laboratories, and the assessmentof student performance independent of subject-matter.

Shulman (1986, 1987) suggested that teachingexpertise should be described and evaluated in termsof pedagogical content knowledge (PCK). Cochran,DeRuiter and King (1993) applied their model ofpedagogical content knowing (PCKg) to teacher edu-cation to generate hypotheses for future applied andtheoretical research. According to this model,

PCKg includes four components: knowledge ofpedagogy, knowledge of subject-matter, knowledgeof environmental contexts, and knowledge ofstudents. Content-specific pedagogy has been definedto understand the knowledge, skills, abilities, and in-terests students bring to the study of subject matter(Arreaga-Mayer & Greenwood, 1986; Confrey, 1990;Corno & Snow, 1986; Swisher & Deyhle, 1987;Wallace, 1986). Reynolds (1992) indicated that thisincludes students' conceptions and possible miscon-ceptions of particular topics in the subject matter,students' beliefs about their ability to succeed in thesubject matter, students' academic self-concept, stu-dents' nonverbal and sociolinguistic backgrounds, theexpectations students' families have, and students'personal experiences.

Other researchers see PCK as comprising themost appropriate forms of representations for thesubject matter for a given group of students (meta-phors, explanations, illustrations, examples, etc.),which make the subject matter understandable andinteresting to students (McDiarmid, Ball, & Anderson,

164

185

Case Study of Beginning Biology Teacher

1989; Wilson, Shulman, & Richert, 1987). Reynolds(1992) suggested that at the time of formal summativeevaluation for licensure, we should expect beginningteachers to be able to:

1. Plan lessons that enable students to relate newlearning to prior understanding; 2. develop rapport'with students; 3. establish and maintain rules androutines; 4. arrange the physical and social conditionsin the classroom; 5. represent and present subjectmatter in ways that enable students to relate newlearning to prior understanding; 6. assess studentlearning using a variety of measurement tools; 7. reflecton their own actions in order to improve teaching(P.7).

Measurement of these expectations was not withinthe scope of their study. However, they provided abasis for describing a beginning teachers' professionalgrowth.

The purpose of this study was to investigate abeginning biology teacher's professional growth andto describe and determine if "knowledge of students"was a critical component of pedagogical contentknowledge for this teacher.

II. Methodology and SettingInterpretive research methods described by

Erickson (1986) were used. Multiple qualitative andquantitative data gathering methods and triangulationwere used to enhance the validity of the findings.

1. Participant

Miss Cho (anonymity), a female beginningteacher, graduated from the Department of Biology,National Taiwan Normal University (NTNU), in 1992.She taught four grade-7 classes of biology in a juniorhigh school located in a western suburb of Taipei, thecapital city of Taiwan. She had a three-week student-teaching experience with grade-7 biology classes inthe university laboratory school before graduation.She comes from an "educational family" of a large cityin Southern Taiwan; both parents were school teach-ers. Miss Cho was enthusiastic about teaching biol-ogy.

2. Site

The junior high school was three years old. Manyschool buildings were still under construction. Thegrade-7 students used classrooms borrowed from anelementary school nearby. There was a fence sur-rounding the building to separate students of these two

schools. There were no laboratories or audio-visualstudios in this school. Classrooms were equipped withonly basic equipment, such as several sets of micro-scopes, petri-dishes, and beakers. Many items, suchas models and audio-visual equipment, were not avail-able. There were 50 students in the observed classwith approximately equal numbers of male and femalestudents. The majority of the students came frommining or laborer families. A majority of them wereeager to seek opportunities for further schooling aftergraduation from junior high school.

3. Data Collection

The qualitative data were collected mainly bymeans of participant observation, teacher interviews, and document gathering following the proceduresintroduced by Gallagher (1991). Miss Cho was ob-served for 55 lessons spread across a completeacademic year, 1992-1993. All observations werevideo-tape recorded. Before or after class interviewswere conducted with the teacher to obtain informationabout her beliefs related to her instructional behavior.The interviews were also tape-recorded. Teachingplans, teaching aids, students' work sheets and testpapers related to the observed instruction were pre-served in the form of photocopies or photo picturesto be analyzed. Some quantitative data were alsocollected. Teacher-student interactions and the fre-quency of student-initiated questions were recorded,and the teacher's patterns of response to the student-initiated questions were also analyzed. The teacher'sre- sponding patterns were categorized into fourcategories based on a model which was used to analyzeteachers' questioning skills in the a previous study(Taiwan Provincial Kaoshiung Teachers' College,1974):

165

A. Gives definitive strict responses, such as defi-nitions, principles or rules, as defined in text-books or encyclopedia without any consider-ation of students' prior knowledge or everydaylife experience.

B. Gives analogies or metaphors which relate tothe student's everyday life experience.

C. Gives direct yes/no responses to confirmstudent's ideas but without any further expla-nation.

D. Confusions or algorithms given for solutionswithout an understanding of the logical relationbetween the problem and solution.

At the end of the academic year, a questionnaire

1d6

J.H. Yang and I.L. Wu

was administered to students to assess the student'sattitudes toward the teacher.

4. Data Analysis

The data base consisted of both qualitative andquantitative data collected from field-notes of class-room observations, interviews and documents.Vignettes were obtained to describe the teacher's be-havior and possible explanations or tentative asser-tions of the behaviors. Then, the vignettes were sorted,and frequency analysis was proceeded. Finally, as-sociated with quantitative data collected from theclassroom interaction analyses and the results of aquestionnaire on students' attitudes toward the teacher,analytical reduction and triangulation were employedto enhance the validity of findings.

5. Triangulation

The validity of a qualitative study can be en-hanced by triangulation ( Patton, 1990). The primaryform of triangulation for this study was " triangulationof sources" (Patton, 1990; p.464). Data triangulationwas accomplished mainly by using after-class inter-views. By comparing data, including both qualitativeand quantitative data collected from classroom obser-vations, interviews, and documents (e.g., test papers,students laboratory notes and worksheets of classactivities), the consistency of statements waschecked.

Ill. Results and DiscussionIn this section, the performance of participants

in the observed classes are discussed. The partici-pants' concerns and perceptions as expressed by theparticipants are presented and supported through in-terview data and the results of document reviews. Theresults and assertions are discussed in four parts ofprofessional components as below:

1. Alternative Representations

Assertion 1: The beginning teacher appeared to un-derstand the need for creating lessonsthat were appropriate for the subjectmatter; however, she did this in rathersuperficial ways due to her lack ofknowledge of students.

PCK is a type of knowledge that is unique to ateacher and also is the form of knowledge that makes

science teachers rather than scientists (Gudmundsdottir,1987a,b). According to the academic record providedby the Department of Biology, NTNU, Miss Cho alwayshad excellent performance in both subject matterknowledge (in the biological sciences) and pedagogi-cal knowledge (in science education courses) duringher four years of study. However, at the beginningstage of her teaching career, she could not alwaysinterpret her students' questions properly. Her stu-dents always expressed their questions freely in herbiology classes, but Miss Cho seemed to have diffi-culty in answering students' questions at critical points.An example from a Photosynthesis lesson (S=Student;T=Miss Cho) is quoted:

S: Miss Cho, there is water existing in the firststage ("light" reaction). Why does a planthave to produce water in the second stage("dark" reaction) 9 Why can't water beutilized directly?

T: (repeating student's question in her ownwords) Do you mean why can not the waterin the second stage be utilized in the firststage?

S: No, I meanT: (restating the question in other words) Why

is water produced when we have water at thebeginning?

S: Yes,T: (smiling) I see

(Field Note: 10/21A)

Miss Cho explained in a rather strict, definitiveway; she repeated a paragraph from the textbook. Aspredicted, students were not satisfied. Then she usedan analogy:

T: Three good friends (an analogy of a watermolecule, 2 H atoms and one 0 atom), arestanding together hand in hand; after thechemical reaction, one of them (an H) leavesand a new friend (another H) joins to the othertwo.

(Field Note 10/21B)

The students seemed to be satisfied with thisexplanation. She mentioned "changing head" as ananalogy. Ebert (1993) argued the importance of theteacher's knowledge about the learner as a constructorof knowledge and how this knowledge has the poten-tial to impact instruction and, subsequently, learning.The grade 7 students in Taiwan did not have back-ground knowledge about chemical reactions. This

166

rip 1-7:


"change head" analogy was not an appropriate one.Miss Cho may have had some difficulty thinking ofan appropriate representation at that moment. Thiskind of difficulty was observed often in the " Cell(10/7)", "Leaf and Conduction (11/13)", "Tropism ofPlants" (12/23), and "Heredity (4/16)" lessons. In the"Tropism of Plants" lesson, Miss Cho explained thetropism of plants by using a "running track in a sta-dium" analogy:

T: Look, one side of a plant stem grows faster.the other grows slower, then, what will hap-pen? Another example, (drawing a pic-ture of a running track in a stadium) , now,you have physical training. Given theright to choose, which runway will youselect?

S: Inside one!T: Yes! You will pick the inside track, because

the distance of the inside track is shorter thanthat of the outside one. Then, think about this,this side of a stem with less growth hormonegrows slower, and the other side has moregrowth hormone. Thus, one side grows faster,the other side slower, both sides grow, now,they are growing at same time, one faster,the other slower , then this 'person' grow-ing faster has to bend over to this side, theother 'person' grows slower, therefore, itbends

(Field Note:12/23)

The analogy employed may be appropriate forsome imaginative students, but Miss Cho neitherclarified the similarities and differences between thesetwo ideas, nor did she make connections to students'preconceptions; therefore, the students were even moreconfused.

Assertion 2: Teacher knowledge of students' priorknowledge and life experience mayform a foundation for the teacher tocreate appropriate analogies to presentconcepts for her students in an intel-ligible way.

PCK includes knowledge of the conceptual andprocedural knowledge that students bring to the learn-ing of a topic. It also includes knowledge of tech-niques for assessing students' understanding anddiagnosing their misconceptions (Carpenter, Feonnema,Peterson, & Carey, 1988). A field-note taken from

-the Photosynthesis lesson is an example of this prob-

lem:

S: Miss Cho, why does a plant need energy fromthe sun instead of doing the job by itself?

T: Because without chemical energy, there is nopower to separate the three good friends

S: That is right, I mean, if there is no chemicalenergy, can plants do the job?

T: No, plants can not do it.S: Does the plant need only water made of two

H atoms and one CO2?T: Do you mean water in the second stage?S: Can't that be reused?T: Do you mean that the produced water could

be reused by plant?S: Yes.S: Miss Cho, I meant, does it need only the water

made of two H atoms and one CO2?T: Who is "it"?S: The plants. (Miss Cho was confused, so she

checked the textbook and tried to make senseof the student's question.)

S: He asked, "Do the two H have to bind withCO2 and then be utilized?"

S: No, I meant, "Does the plant need only thewater made of two H atoms and one CO2?"(Students laugh)

T: I know what is your question now. The waterproduced can be reused by the plant

(Field Note: 10/21B)

At the beginning, Miss Cho could not understandwhat the student was asking. After a while, Miss Chodecided the question was, "Can the produced water bereused?" Then she gave the students an answer.However, in the student's mind, the produced waterwas made of two H and a CO2!

(Interview 10/21C)

In the after-class interview, she was astonishedwhen the researcher pointed out this misconception ofher student. She said:

T: Really! Did he say that? (water made of twoH atoms and one CO2). No wonder studentslaughed at him!

T: You must know their (her students') strangetalk with funny gestures and accents!

(Interview 10/21 D)

Most of her students came from miner families.For the city-born Miss Cho, the students were talkinga strange language. She always had difficulties

167

188


understanding them. Miss Cho urged the studentsto know if the water produced ( through photosynthe-sis) could be used by the plant (for "next" photosyn-thesis). In the student's mind, the water producedwas a different kind of water (made of two H and aCOD!

(Interview 10/21 C)

This kind of misunderstanding, caused by herconfusion over what the student was saying, occurredduring almost every lesson across the whole schoolyear. An effective teacher creates lessons that enablestudents to connect what they know to new informa-tion. The teacher also knows how to tailor the subjectmatter to the students. Photosynthesis is one of themost abstract subjects in grade-7 biology. This subjectcontains the concepts of "energy transformation","chemical synthesis" and even "thermodynamics". Itis not an easy task for even a veteran teacher to createappropriate representations. Nevertheless, an effec-tive teacher should , at least, be sensible to students'misconceptions.

2. Knowledge of Students

Teachers differ from scientists, not necessarilyin the quality or quantity of their subject matterknowledge, but in how that knowledge is organizedand used. An experienced science teacher's knowl-edge of science should be organized from a teachingprospective and be used as a basis for helping studentsto understand specific concepts. One of the fourcomponents of teacher knowledge described by Cochran(1991) is the teachers' knnwledge of students, includ-ing knowledge of students' abilities, learning strate-gies, ages and developmental levels, attitudes,motivations, and prior knowledge of the concepts tobe taught (Cochran, 1991).

Assertion 3: The beginning teacher successfullycreated a classroom climate in whichstudents were motivated to learn. Shedeveloped an excellent rapport withmost of her students. However, shesaw the students as a group and ne-glected individual needs.

Miss Cho established an open classroom climatethat enabled her to collect students' ideas and infor-mation. The students felt free to ask questions andspeak during class. Some students asked questionsfrequently in class, and others would talk to Miss Chodirectly. The researchers tabulated the frequency of

selected classroom interactions.The frequency of student-initiated questioning

increased through the first month of school (0.15-0.71times per minute) and stayed at around 0.3 times perminute (0.23-0.69 times per minute) in the followingmonths during the first semester. She had much morestudent-talk time in her observed class compared tothe average teacher-talk time of 70 percent reportedby Delamont (1990).

Miss Cho created a classroom climate in whichstudents wanted to learn. She developed excellentrapport with her students. Her students' images of agood teacher were like someone who knew everythingabout the subject matter. It may have made Miss Choact as a knowledgeable scholar who knew all theanswers to questions raised by students.

The questionnaire administered at the end of thesemester also indicated that students felt free andwilling to communicate with Miss Cho (Table 1).Although Miss Cho always misjudged some student-talk, she was quite willing to listen to students, andenriched her knowledge of students.

However, she always saw students in the classas a group and neglected individual needs. In the after,class interviews, concerning interaction problems inher biology class, she always complained that somestudents would not cooperate with her. During the firstsemester, one of the students, David, frequently askedquestions that made Miss Cho very uncomfortable.The researchers interviewed David after class andmade sure that he did not do that on purpose (FieldNote: 4/16). He indicated that he asked questionsbecause he was really confused and eager to clarifyhis concepts (Interview: 6/25).

From one of the after-class interviews in thesecond semester, we sensed Miss Cho's negativeattitude toward David's questioning behavior. In May,during the second semester, we found that David was

Table 1. Students' Metaphors of Miss Cho

Metaphors Number of Students

The source of knowledge--a master mind--a dictionary--an encyclopedia--a mine of wisdomA kind lady full of patience--a mother--a nursemaid--an elder sisterA shy pretty ladyA lady of responsibility

31/50

23/50

2/501/50

168

-


not asking questions anymore. This made Miss Chofeel better. The researchers talked to Miss Cho andexpressed concern that David was a slower learner andthat was why he asked questions. However, Miss Chostill insisted on her perception that David didnot carefully listen to the lectures and asked sillyquestions without thinking. In her mind, Davidwas one of the students who never cooperated withher.

We thought from the data analyzed that Miss Chodid establish an open classroom for students to exploretheir problems but only created a group of 10-12 veryactive target students. She did not always make senseof the information she received from other students.Sometimes, she misjudged the information from thosestudents. Miss Cho formed her knowledge of studentsthrough classroom interaction. But her knowledge ofstudents was not related to reality and, thus, affectedher teaching as Elbaz (1983) described.

3. Values in PCK

Assertion 4: The students' social background andvalue of schooling, e.g., learning forentrance examinations, affected thebeginning teacher's value of educa-tion and, consequently, influenced herteaching practices.

The role of values in the development of PCKwas examined through interviews and classroomobservation of four expert high school teachers byGudmundsdottir (1990). Analysis indicated that theteachers' value orientations to their subject matterinfluenced their choice of content pedagogical strat-egies and their perception of students' instructionalneed.

In the Heart, Vein, and Circulation lessons. MissCho talked about her experiences in previous teachingabout the structures of the human heart. She realizedthat the students only tried to memorize the terms forthe structure of the heart and ignored which bloodvessels connected with which parts of the heart. Ifthe figure of the heart in the textbook was altered, thestudents would be confused (interview on 11/20).Miss Cho pointed out this phenomena to her studentsand said, "Don't memorize only the locations";however, the students were still confused by what shesaid.

T: When you are taking an examination, thefigures on the examination sheets are notnecessarily the same as this one. Therefore,

don't just memorize the locations; instead,you have to memorize each part (of the heart)and how they are connected to each other. Allthose (blood vessels) connecting to ventriclesare arteries, such as the aorta which connectsto the left ventricle (sticking "Aorta" on theblackboard). The aorta here is the same asthe "main artery" you have learned in HealthEducation, The aorta connects to the leftventricle, and the pulmonary artery connectsto the right ventricle (sticking "pulmonaryartery" on the blackboard) This one. Allthe blood vessels which connect to the atriumare veins, such as the pulmonary vein and thesuperior and inferior vena cava. That's it(Miss Cho explained the figure without eyecontact with any student.)

S: I don't understand T (sticking "pulmo-nary vein" on blackboard): Can't you get it?

T: Now, see this figure is very importantwhen you are preparing for an examination;however, the figure in an examination willnot always be the same as this one, line onein your textbook You, sit down (to astudent who tried to ask a question). You cannot just memorize where the veins, aorta,pulmonary vein are located; instead, you mustknow which parts of the heart they connectto

(Field Note: 1/20)

At the beginning, Miss Cho drew a figure whichwas the same as that in the textbook on the blackboard,with the aorta and pulmonary artery crossed to eachother. After some dialogue between the teacher andstudents, Miss Cho drew another figure (with no crossbetween the aorta and pulmonary artery) and asked oneof the students to fill out the names of the structuresof the heart. As predicted, the student made manymistakes. Miss Cho reminded the students, "Don'tmemorize the locations"; however ,the students failedto figure out what she meant. Obviously, the concernsof the teacher and the students were quite different;thus, the students were confused and off-track. In theafter-class interview, we found that Miss Cho couldnot make sense of this problem. She insisted onexpressing her concern about the examination. Inresponding to the researcher's suggestion of " Whydidn't you use a three-dimensional mock-up model fora more concrete representation?", she said: " They arenot likely to have a three-D model in an examination.I have to teach them how to respond correctly inexaminations."

169


The knowledge base includes the different waysof knowing that are important for teachers and nec-essary for successful practice. Education, accordingto Peter (1970), implies transmitting something thatis worthwhile in a morally acceptable manner, andbeing educated means having changed for the better(p.6). Educational philosophers recognize thatthere is no such thing as value-free education (Gud-mandsdottir, 1990).

Miss Cho's values which were related to herPCK influenced her teaching practices. As an ex-ample, she provided algorithms to help her studentsbe good test takers. Due to the severe competitionto obtain further schooling, she had pressure fromstudents and the school administration. The pressuremay have influenced her choice of instructionalstrategies or representations. She administeredshort tests or quizzes to her students during almostevery class period. The test-items for a class alwaysincluded the same test-items administered in pre-vious classes. She explained that she had to do thatbecause she was deeply concerned that her studentsbe successful test-takers. The results suggestedthat Miss Cho's beliefs influenced her teaching prac-tices.

The case below was taken from the Hereditylesson. It shows another problem in instructionalrepresentation.

T: O.K., another question, a couple would liketo have another baby. What is the probabilityof having a boy with attached earlobes?

5- 1/2T: Lily, what's the probability?S: 1/2T: 1/2, is that right? Look, what is the probabil-

ity of having attached earlobes? That is1/2 (Miss Cho points to the Punnett Squareon the board). The probability of havingattached earlobes and of having free earlobesare equal, both are 1/2. But the couple wanta boy,... the probability of having a boy ora girl is 1/2; therefore, if this couple wantto have a boy with attached earlobes, theprobability is 1/4. That means when bothpropositions are true, we multiply, this con-cept

S: Why is the probability of gender uncertain?T: No, the probability of having a boy or a girl

is exactly 1/2.S: Then, why do we get 1/4?T: I added one more proposition in this case: the

baby must have attached earlobes. Another

example may be helpful to explain this con-cept. One department store is having a 20%off sale, and you go shopping. Another smartstudent will wait for an eitra 20% off. Whenwill you go shopping?

S: When there is an extra 20% off.T: That's right, with an extra 20% off, now we

have 36% off. Why? 20% off and another20% off (writing down 0.8x0.8) is 36% off.Isn't that the better time to go shopping? Bothpropositions are true, so we multiply themGot it? You must remember this concept

S: I don't get the pointT: Hands up if confusing, the reason why we

multiply, hands up, if you can't figure it out!You are falling asleep! (Many students puttheir hands up.)

S: Why with an extra 20% off do we get 36%off?

T: 0.8 times 0.8 equals 0.64!(Field Note: 4/16)

In this case, Miss Cho provided an algorithm toterminate the argument. She explained how to cal-culate the probability of "both propositions are true"by using the analogy of an extra 20% off in a depart-ment store. The students knew that an extra 20% offis cheaper than only 20% off, but they could not figureout that an extra 20% off meant a total of 36% off.In addition to the original problem, "having a boy withattached earlobes", the students had another confusingconcept.

The subject matter which the students were ar-guing about was a concept related to proportionalreasoning. According to Piaget's cognitive develop-mental theory , it is a formal operation skill not usuallyprocessed effectively by students until the stage offormal operation. Based on research conducted by Lin(1982), only one-third of the grade-7 students in Taiwanhave developed this skill. This means that the ma-jority, two-thirds of the students, probably have dif-ficulty trying to understand these ideas. The teachermerely provided a recipe, or an algorithm, withoutunderstanding.

7 170f.)

4. Professional Growth

As pointed out in assertion (1), this beginningteacher appeared to understand the need for creatinglessons that were appropriate for the subject matterand students, but she accomplished this task only insuperficial ways. However, Miss Cho still made someprogress in her instructional representations.

3191


Assertion 5: Through active classroom interaction,the teacher developed her understand-ing of students, and this understand-ing may have contributed to herprofessional growth.

Her students felt free to express their problemsduring the biology classes. As a matter of fact, thisis rather unusual in most classes in Taiwanese schools.Taiwanese students seldom ask questions duringtheir classes. An international study found thatTaiwanese students had a strong tendency to relyon " taught method" for problem solving and wereseldom encouraged to ask questions in their classes(Yang, 1993). However, in Miss Cho's classes, asdescribed in the previous paragraphs, the averagenumber (times) of student-initiated questioning(S-i-Q) during the first semester was 10 times perperiod (50 minutes). As shown in Table 2, morethan one-third of the questions in the biology classeswere S-i-Q. Regrettably, Miss Cho always gavedirect answers to almost all S-i-Qs. No question wasfollowed with another question by the teacher. Shetried to answer all the questions raised by the students.

Table 2. Student-Initiated Questions in a Class Period (50 minutes)

This probably is one of the reasons why studentssaw her as a "master mind" (a resourceful person)(Table 1).

The teacher's patterns of response to the S-i-Qwere analyzed. As shown in Table 3, 19% of theresponses to S-i-Q were strict definitions or principlesas defined in textbooks. An example shownbelow was the typical "strict definition" responsepattern:

(In responding to a student's question aboutPhotosynthesis) The 0 (oxygen) in the initial H20(water) is hit by chemical energy; the rest of two H(hydrogen) atoms bind together with 0 in CO2 and,thus, produce C6H1206 (glucose) and a new watermolecule.

(Field Note 10/21A)

36% of the S-i-Q were answered using analogiesand metaphors employed from the students' everydaylife experiences, such as the "changing head" analogyfor the chemical reaction in Photosynthesis and"running track in a stadium" analogy for tropism asdescribed in Assertion 1. Another 25% of the

Month September October November

Average Q+A Period (%)aAverage Number of S-i-Qb% of S-i-Q of all questions raised

298

38

4410

23

3813

35

a. Average time allotted for "questions and answers" (per 50-minutes-class period).b. Average number of Student-initiated questions raised.

December January Average

23 23 31.412 7 10.048 33 35.4

Table 3. Teacher's Responding Patterns to Student-initiated Questions During 13 Consecutive Class Hours

Pattern Percentages per total S-i-Q raised

1. Strict Definitions

2. Analogies or Metaphors

3. Confirmation

4. Confusionsand Gave Algorithms

Date

60 25

50

25

40

34 33 10

17 60

33 25 10

33 25 20

4 9 18a*

6

25

25

44

36

21

36

7

18b* 21 6 13

Month Sep Sep Sep Sep Oct

LivingWorldLessons Cell

20 14 17

50 60 57 41

38 29 50

13 20 16

20 11 23a* 23b*

40 50

20 40

40 10

6 8

Nov Nov Nov Dec Dec Dec Jan Jan

Digestion Leaf

Biosphere Organ Photosynthesis

Stimulus Irritability Respiration

Circulation Behavior Homeostasis

*18a,b & 23a,b: 2 consecutive class hours in the same day (September 18 & December 23)

171

1,92


questions were answered using yes/no or right/wrongpattern. Responses to 20% of the S-i-Q were con-fused and, in most cases, algorithms were used asdescribed for the heart structure and Hereditylessons in Assertion 4. This performance was rather"normal" for a beginning biology teacher (TaiwanProvincial Kaoshiung Teachers' College, 1974; Yang,1993).

In the very beginning stage of her teaching, somedegree of confusion and inappropriate representationswere likely. She always responded to her student-initiated questions by giving strict definitions quotedfrom the students' textbooks and in some cases evenquoted from college textbooks! This made the stu-dents more confused. In other cases, she just askedstudents to memorize something, or gave them analgorithm to satisfy them. After several weeks, sheseemed to recall what she was taught in her scienceeducation courses at NTNU and tried to employ analo-gies or metaphors for the S-i-Q. As shown in Fig. 1,her strict definition responding pattern decreased; onthe other hand, application of analogy and metaphorpatterns increased across the whole semester. Someanalogies used were not very appropriate for her stu-dents and were presented in superficial ways. Dataanalyzed indicated that in spite of her problems shedid achieve professional growth. Also, she reducedthe number of strict definitions and appeared to in-crease the number of confirmations. These data mayindicate that she was gaining confidence in her S-i-Qs.

IV. Conclusions and Implications

Data from 55 periods of classroom observationand many interviews indicated that Miss Cho devel-oped rapport with her students and established anactive open-classroom. She allowed greater freedom

9/4 9/9 9/18 9/18 10/21 11/6 11/13 11/20 12/11 12/23 12/23 1/6 1/8

Date

Fig. 1. Teacher's patterns of response to student-initiated ques-tions.

of action during the class. Students were comfortableexpressing their ideas and problems during thebiology classes. However, very frequently, Miss Chomisjudged the information she received throughstudents' questions and statements. In many cases,she ignored her students' prior-knowledge or experi-ences that would have helped them acquire newknowledge.

Constructivism suggests that teachers should firstteach students how to learn, and, secondly, what tolearn. It has also been suggested that language cannotbe a means of transferring information in teaching; itshould be used to cope with our environment and helpus make sense of our world (Lerman, 1989; Fosnot,1989; Steffe, 1991; Reynold, 1992; von Glasersfeld,1984, 1989, 1991). Obviously, Miss Cho and herstudents frequently failed to communicate with eachother effectively. Although Miss Cho created an activelearning climate, her poor knowledge of studentsinhibited her PCK performance as identified byCochran, DeRuiter, and King (1993). She frequentlydid not use appropriate representations to help stu-dents master concepts as indicated by Zheng (1992).Miss Cho achieved some professional growth in herpatterns of responding to student-initiated-questions.However, her ability to seek appropriate representa-tions and help students develop metacognitive strat-egies remained superficial.

Reynolds (1992) indicated that a beginning teacherneeds time to learn about the school culture and stu-dents; a teacher's knowledge develops through theprocesses of continuous learning and reflection. MissCho needed a longer period of time to increase hercompetence; she also would probably have gainedfrom more guidance. Wallace and Louden (1992) haveindicated that knowledge development in teachers isa gradual process, rather than one composed of suddenleaps of insight. A teachers' repertoire of instructionalrepresentations constantly expands as a result of"thousands of hours of instruction and tens of thou-sands of interactions with students" as described byBerliner (1987).

Effective teaching is the product of self-organi-zation. Data obtained here suggest that a teacherwithout substantial knowledge of her students is notlikely to employ her taught subject matter knowledgeand pedagogical knowledge effectively in a real class-room situation or to create a favorable learning en-vironment for her students. Knowledge of studentsthat can aid effective teaching should be learned throughreal classroom experiences. Knowledge of studentsseems to be a foundation needed for Miss Cho's PCKgrowth (Fig. 2).

172

tvr 193

TaughtSubject-Matter

Knowledge

TaughtKnowledge of Student


TaughtPefiaSoffwalKnowledge

Substantial Knowledge of Studentsthrough

Real Classroom Experiences

Communication Skills

LanguagePrior-Knowledge and ExperienceEvectations

Social Background

Value Orientation

Establishes II Reflection

Favorite Classroom Climate

Teacher-Student Rapport

Frequent Student-initiatedQuestionStudents' Attitude towardClass and Teacher

PROFESSIONALGROWTH

Reflection

Reflection oPs

Appropriate InstnictionalRepresentations

InterOws

DefinitionsAnalogies or MetaphorsMgorithmsExplanationsQuestioning Skills

Fig. 2. The beginning biology teacher's knowledge of student andprofessional groieth model.

Acknowledgments

This study was funded by the NSC under grant number NSC82-0111-S-003-037.

References

Arreaga-Mayer, C., & Greenwood, C. R. ( 1986). Environmentalvariables affecting the school achievement of culturally andlinguistically different learners: An instructional perspective.NABE: The Journal for the National Association for BilingualEducation, 10(2), 113-135.

Berliner, D. C. (1987). Ways of thinking about students and class-rooms by more and less experienced teachers. In J. Calderhead(Ed.), Exploring teachers' thinking. London: holt, Rinehart &Winston. -

Carpenter T. P., Fennema, E., Person, P. L., & Carey, D. A. (1988).Teachers' pedagogical content knowledge of students' problemsolving in elementary arithmetic. Journal for Research inMathematics Education. 19(5), 385-401.

Cochran, K. F. (1991). Pedagogical content knowledge: Teachers'transformations of subject matter. NARST News, No.27.

Cochran. K. F., DeRuiter, J. A., and King, R. A. (1993). PedagoicalContent Knowledge: An integrative model for teacher prepa-ration. Journal of Teacher Sducation, 44(4), 263-272.

Confrey, J. (1990). A review of the research on student conceptionsin mathematics, science, and programming. In C.B. Cazden(Ed.), Review of research in education, (Vol. 16, pp. 3-56),Washington, DC: American Educational Research Association.

Corno, L., & Snow, R. E. (1986). Adapting teaching to individualdifferences among learners. In M.C. Wittrock (Ed.), Handbookof research on teaching (3rd ed., pp.605-629). New York:Macmillan.

Delamont, S. (1990). Let battle commence: Strategies for theclassroom. In: Delamont, S. (Ed.), Interaction in the classroom

(2nd Ed.). New York: Routledge.Ebert, C. L. (1993). An assessment of prospective secondary teach-

ers' pedagogical content knowledge about functions and graphs.Paper presented at the meeting of the American EducationalResearch Association, Atlanta.

Elbaz, F. (1983). Teacher thinking: A study of practical knowledge.New York: Nichols Publishing Company.

Erickson, F. (1986). Qualitative method in research on teaching.In: M.C.Wittrock (Ed.), Handbook of research on teaching (3rd.edition). London: Collier Macmillan Publishers.

Fosnot, C. T. ( 1989). Enquiring teachers, enqu'iring learners: Aconstructivist approach for teaching. New York: TeachersCollege.

Gallagher, J. J. (1991). Uses of interpretive research in scienceeducation. In: J.J.Gallagher (Ed.). Interpretive research inscience education. NARST monograph, No.4.

Gudmundsdottir, S. (1987a). Learning to teach social studies: Casestudies of Cathy and Chris (Knowledge Growth in a ProfessionPublication Series). Stanford, CA: Stanford University, Schoolof Education.

Gudmundsdottir, S. (1987b). Pedagogical content knowledge:Teachers' ways of knowing (Knowledge Growth in a ProfessionPublication Series). Stanford, CA: Stanford University, Schoolof Education.

Gudmundsdottir, S. (1990). Values in Pedagogical content Knowl-edge. Journal of Teacher Education. 41(3),44-52.

Lerman, S. (1989). Constructivism, mathematics and mathematicseducation. Educational Studies in Mathematics, 20, 211-223.

Lin, P. J. (1982). A study on cognitive development of junior andsenior high school students in Taiwan. Science EducationMonthly, 51, 12-22. (In Chinese)

McDiarmid, G. W., Ball, D. L., & Anderson, C. W. (1989). Whystaying one chapter ahead doesn't really work: Subject-specificpedagogy. In M.C. Reynolds (Ed.), Knowledge base for thebeginning teacher (pp.193-205). Oxford: Pergamon.

Patton, M. Q. (1990). Qualitative evaluation and research methods(2nd ed.). Newbury Park, CA: Sage Publications.

Peters, R. (1970). Ethics and education. London, England: GeorgeAllen & Unwin.

Reynolds A. (1992). What is competent beginning teaching? Areview of the literature. Review of Educational Research.62(1),1-35.

Shulman, L. S. (1986). Those who understand: Knowledge growthin teaching. Educational Researcher, 15(2), 4-14.

Shulman, L. S. (1987). Knowledge and teaching: Foundations ofthe new reform. Harvard Educational Review, 57(1), 1-22.

Steffe, L. P. (1991). The constructivist teaching experiment: Il-lustrations and implications. In E. von Glasersfeld (Ed.), Radicalconstructivism in mathematics education (pp.177-194). Bos-ton: Kluwer Academic Publishers.

Swisher, K., & Deyhle, D. (1987). Styles of learning and learningof styles: Educational conflicts for America Indian/Alaskannative youth. Journal of Multilingual and Multicultural De-velopment, 8(4), 345-60.

Taiwan provincial Kaoshiung Teachers' College (1974). A Serveyon Junior High School Mathematics and Science Education.Kaoshiung, Provincial Kaoshiung Teachers' College (in Chi-nese ).

von Glaserstfeld, E. (1984). An introduction to radicalconstructivism. In P. Watzlawick (Ed.), The invented reality(pp. 17-40). New York: Norton.

von Glasersfeld, E. (1989). Knowing without metaphysics: aspectsof the radical constructivist position. (ERIC Document Repro-

173

"184


duction Service No. ED 304 344)von Glasersfeld, E. (1991). Radical constructivism in mathematics

education. Boston: Kluwer Academic Publishers.Wallace, B. (1986). Creativity: Some definitions: The creative

personality; the creative process; the creative classroom. GiftedEducation International, 4, 68-73.

Wallace, J., & Louden, W. (1992). Science teaching and teachers'knowledge: prospects for reform of elementary classrooms.Science Education, 76(5), 507-521.

Wilson, S., Shulman, L. S., & Richert, A. E. (1987). "150 differentways" of knowing: Representations of knowledge in teaching.In J. Calderhead (Ed.), Exploring teachers' thinking (pp.104-

124). Eastbourne, England: Cassell.Yang, J. H. (1993). A comparative study of Taiwanese and Scottish

students' problem solving skills as performed in the interna-tional assessment of educational progress 1991. Proceeding.National Science Council, Republic of China, Part D: Math-ematics, Science and Technology Education 3(2), 80-93.

Zheng, J. (1992). An exploration of the pedagogical content knowl-edge of beginning and experienced ESL teachers. A thesissubmitted to the Faculty of Education in conformity with therequirements for the degree of Master of Education. Queen'sUniversity, Canada.

+)3111.1tMIAILliiiRZIKI;+<,6irA

f',1140-0 t'f-tItt*A

trk N

21KTiF4saiffMnitliff9VA 1).tiff V.U41±110-11-1M434Mg , *t.OME**=1-tic*OrAA-Mwaviri.re]twIt ishilmmma 55 ailMtrfflailltaU.EcafillaalklYJIVA , AVM ttf.t-Wc440IMIlkAW0(3 II MIE*Itth3-}-ika'64tP.4,14M4-t , ff91MXPA7Titi-fSLOT-traYfOgaliiit*IMM nek*ml w A-1*Pft.1;rWOM IstMTMAIJIfAlfrikaggfc*Ic AW-MhLfulitakUlt*t11193VXMIYJOSt 1)..(1-elltiVgarffri AVV,t690.111;t4-ViAtl*X4ROgRAU9110C1141310

-ii4 4.L.A

41-it Fir; 1111g C D ) : ittfProceedings of the National Science Council (Part D):


ifq*A7E-1-4.+fii1-1q-**-4-**41%,t**AA-ItWalAc hliCt°

o 2534ffl I4 M-7614 is!J 129041tIk t *4- 1-45-Rtt#04St.

F*Et lt-ifL*VIL**74 , VAL ERIE

W-Fi r7 MA"DiSciplines: science education, mathematics education, environmental education,



Domain: learning and learner, curriculum and materials, instruction (includingcomputer assisted instruction), assessment and evaluation, history and


etc.

07-***f144tg'-4-=*"P**flA4+AR°aVIIA*JtAitlt-A4AJA-14-f*TI-ASLAAigtakitAlffl*.

RT. * AR 4tAk.

*4-1:7 t*-4---=-R.106-g,174-417011.7 4*: 02-7377595 Email: [email protected]

44 : 02-7377248

Call for Paper

Proceedings of the National Science Council (Part D):Mathematics, Science, and Technology Education

invites you to contribute Original or Reviewpapers in the following:

Disciplines: science education, mathematics education, environmental education,technology education, special education, health education, medical edu-cation, and information education, etc.

Domain: learning and learner, curriculum and materials, instruction (includingcomputer assisted instruction), assessment and evaluation, history andphilosophy of science, teacher preparation and professional development,etc.

Manuscripts should be written in English and sent in triplicate to:

Proceedings of the NSC, R.O.C., Part D

Room 1701, 17F, 106 Ho-Ping East Road, Sec. 2

Taipei 10636, Taiwan, Republic of ChinaTEL: (02) 737-7595

FAX: (02) 737-7248

Email: [email protected]

97

I

fiLf-T A: EVitia-W tt& fieR121**=1-hit

* A I*,

AW"-.1 tRR

*1=11 pt cam James Bensen

Brmidji State University,Bemidji, MN, U.S.A.






alL110636f-Pf-*MT.11106V,170(02) 737-7594 737-7595(02) 737-7248

*M

Alan J. BishopMonash UniversityVictoria, Australia






Email : [email protected]@nscpo.nsc.gov.tw

tiPi&elmkffiimw=g4-293-3v-F-ftz wit : (02)732-1234

IAMINMMEst61V,1MItY111041384at

T41-rA PAY4203fi

HMCita* 44-14111i MT*

ffill li-VA AtE ZikAYE&MITTit*** AtE : AtifiM**Y4141V,

V S tJ1J± *ft V)N:MAVACa

qt4 : (02)361-7511: (02)361-7511: (02)3821394

St4 : (04)2260330(04)725-2792

: (07)332-4910

Ntap*IIffl ...loilloRatigpoim222-591 )=1 FrOrdal*W44-0*2. DjMiTPrIUM-FilMfr:21K-er °

ItrAM-2 : (02)737-7764

1-7 * tiE * fit LEEt ff f\T --L f!fi rt

1'9

tlan1524g. OR2

a

024068860025

Ste SaMcXADA-1-ivAXA

UnilMIR-rmistrza (14A11k

MMffftNIA=gtaK0444a00MtlitEekRA,W A*A

MMI*02NO*n04*N*IRMAMINtEgAMA A

UffteROX0A-Rt11llAWA*V*

123

135

155

164

igq

ON.E4A-*

U.S. Department of EducationOffice of Educational Research and Improvement (0ERI)

National Library of Education (NLE)Educational Resources Information Center (ERIC)

NOTICE

REPRODUCTION BASIS

1-,-.-. This document is covered by a signed "Reproduction Release(Blanket) form (on file within the ERIC system), encompassing all

or classes of documents from its source organization and, therefore,does not require a "Specific Document" Release form.

This document is Federally-funded, or carries its own permission toreproduce, or is otherwise in the public domain and, therefore, maybe reproduced by ERIC without a signed Reproduction Release form

(either "Specific Document" or "Blanket").

EFF-089 (9/97)

document resume - eric · 2014-06-02 · document resume. se 060 701. guo, chorng-jee, ed....

Documents