predicting genetics achievement in nonmajors college biology

15
JOURNAL OF RESEARCH IN SCIENCE TEACHING VOL. 25, NO. 1, PP. 23-37 (1988) PREDICTING GENETICS ACHIEVEMENT IN NONMAJORS COLLEGE BIOLOGY ANGELA MITCHELL and ANTON E. LAWSON Department of Zoology, Arizona State University, Tempe,Arizona 85287 Abstract Students enrolled in a non-majors college biology course were pretested to determine their level of intellectual development, degree of field independence, mental capacity, amount of prior genetics knowledge, and amount of fluid intelligence. They were then taught a unit on Mendelian genetics. The only student variables found to not account for a significant amount of variance on a test of reading comprehension and/or a test of genetics achievement was amount of prior genetics knowledge. Developmental level was found to be the most consistent predictor of performance, suggesting that a lack of general hypothetico- deductive reasoning ability is a major factor limiting achievement among these students. Introduction Traditionally one of the most difficult topics in general biology courses is Mendelian genetics (cf. Stewart, 1982). The purpose of this study is to investigate sources of that difficulty. Our hope is that someday the practice of education will become more like the practice of medicine, where the causes of student learning difficulties can be identified in a way analogous to that used by physicians to diag- nose illnesses and their causes so that the proper steps can be taken to eliminate those difficulties and insure comprehension (e.g., Champagne & Klopfer, 1982; Fleming & Malone, 1983). From a theoretical point of view, at least four cognitive variables can be ar- gued to be of importance in learning genetics: (1) general hypothetico-deductive reasoning ability because the topic of Mendelian genetics is a classic example of a hypothetico-deductive system of scientific explanation (e.g., Lawson, 1958; Law- son, 1983; Oldham & Brouwer, 1984; Smith & Good, 1984; Gipson & Abraham, 1985; Costello, 1982), (2) domain-specific knowledge because genetics deals with the transfer of genetic information via the specific mechanism of meiosis and sexual recombination (e.g., Novak, 1979; Hackling & Treagust, 1984), (3) pattern recognition or disembedding ability because learning about specific patterns of gene behavior are embedded in a variety of complex and potentially misleading 0 1988 by the National Association for Research in Science Teaching Published by John Wiley & Sons, Inc. CCC 0022-43O8/88/0 10023 - 15$04.00

Upload: angela-mitchell

Post on 06-Jul-2016

214 views

Category:

Documents


1 download

TRANSCRIPT

JOURNAL OF RESEARCH IN SCIENCE TEACHING VOL. 25, NO. 1, PP. 23-37 (1988)

PREDICTING GENETICS ACHIEVEMENT IN NONMAJORS COLLEGE BIOLOGY

ANGELA MITCHELL and ANTON E. LAWSON

Department of Zoology, Arizona State University, Tempe, Arizona 85287

Abstract

Students enrolled in a non-majors college biology course were pretested to determine their level of intellectual development, degree of field independence, mental capacity, amount of prior genetics knowledge, and amount of fluid intelligence. They were then taught a unit on Mendelian genetics. The only student variables found to not account for a significant amount of variance on a test of reading comprehension and/or a test of genetics achievement was amount of prior genetics knowledge. Developmental level was found to be the most consistent predictor of performance, suggesting that a lack of general hypothetico- deductive reasoning ability is a major factor limiting achievement among these students.

Introduction

Traditionally one of the most difficult topics in general biology courses is Mendelian genetics (cf. Stewart, 1982). The purpose of this study is to investigate sources of that difficulty. Our hope is that someday the practice of education will become more like the practice of medicine, where the causes of student learning difficulties can be identified in a way analogous to that used by physicians to diag- nose illnesses and their causes so that the proper steps can be taken to eliminate those difficulties and insure comprehension (e.g., Champagne & Klopfer, 1982; Fleming & Malone, 1983).

From a theoretical point of view, at least four cognitive variables can be ar- gued to be of importance in learning genetics: (1) general hypothetico-deductive reasoning ability because the topic of Mendelian genetics is a classic example of a hypothetico-deductive system of scientific explanation (e.g., Lawson, 1958; Law- son, 1983; Oldham & Brouwer, 1984; Smith & Good, 1984; Gipson & Abraham, 1985; Costello, 1982), (2) domain-specific knowledge because genetics deals with the transfer of genetic information via the specific mechanism of meiosis and sexual recombination (e.g., Novak, 1979; Hackling & Treagust, 1984), (3) pattern recognition or disembedding ability because learning about specific patterns of gene behavior are embedded in a variety of complex and potentially misleading

0 1988 by the National Association for Research in Science Teaching Published by John Wiley & Sons, Inc. CCC 0022-43O8/88/0 10023 - 15$04.00

24 MITCIIBLL AND LAWSON

instructional contexts (e.g., Witkin et al., 1977; Ronning, McCurdy, & Ballinger, 1984; Lawson, 1983; Gosnell-Moses & Barufaldi, 1984; Jolly & Strawitz, 1983). and (4) information processing (mental) capacity because Mendelian explanations are of sufficient complexity to require the coordination of a relatively large num- ber of ideas for comprehension (e.g., Pascual-Leone, 1969; Scardamalia, 1977a,1977b; Lawson, 1983).

The present study represents an extension of the Lawson (1983) study in which the previous cognitive variables were found to predict various aspects of achievement of the topic of biological evolution. In general, Lawson (1983) docu- mented that there is no simple relationship between cognition and achievement. Rather, both cognition and achievement must be viewed as composed of a subset of independent and dependent variables each with their own unique set of inter- relationships. In other words, the extent to which general hypothetico-deductive reasoning ability predicts achievement depends to a great extent upon what aspect of achievement one is interested in. With regard to genetics achievement, for ex- ample, general reasoning ability is predicted to correlate highly with problem- solving performance but not with performance on multiple-choice items, simply because genetics problem solving presumably calls for the use of a hypothetico- deductive mode of reasoning (and its component parts such as combinatorial, proportional, and probabilistic reasoning), while what is required for answering typical multiple-choice items is primarily the recall of domain-specific facts (cf. Lawson, 1983; Smith & Good, 1984; Morgenstern & Renner, 1984; Stewart, 1982). Thus, the present study represents a partially descriptive attempt to iden- tify those interrelationships within an important and traditionally difficult area of the biological sciences. The following specific questions are raised:

1. To what extent does general hypothetico-deductive reasoning ability, prior domain- specific knowledge, pattern-recognition/disembedding ability, and mental capacity predict achievement of principles of Mendclian genetics as measured by multiple choice, classifying, and matching items and by success in solving textbook genetics problems?

2. To what extent is a student’s ability to comprehend textbook readings about genetics dependent upon the aforementioned cognitive variables?

Answers to these questions are of considerable interest from a theoretical point of view, but they have considerable practical significance as well. As mentioned, the identification of specific causes of student difficulty may lead to the prescrip- tion of specific “cures.” At one extreme, these cures may require only the insertion of a few key missing facts, as would be the case if domain-specific knowledge were found to be the cause. At the other extreme the cure may be to avoid intro- duction of the subject matter entirely, or at least to avoid specific aspects that prove unteachable, as would be the case if some immutable cognitive variable proved to be the cause. The extent to which the present cognitive variables of reasoning ability, pattern recognition/disembedding ability, and mental capacity are immutable is, of course, not entirely clear. It suffices to say that, although changes within students in these variables occur, they are generally of the long-term, gradual kind, thus are most likely influenced very little, if at all, by specific short- term teaching efforts.

PREDICTING GENETICS ACHIEVEMENT 25

Method

Students enrolled in three sections of a nonmajors undergraduate biological science course were pretested to determine their: (1) general ability to reason hypothetico-deductively, i.e., their level of intellectual development (concrete operational, transitional, formal operational), (2) ability to disembed relevant infor- mation from irrelevant background (i-e., their degree of field independence), (3) in- formation processing capacityhize of working memory, (4) amount of prior knowledge of genetics, and (5 ) fluid intelligence.

Following the pretesting, a series of three laboratory-oriented lessons were taught in which the concepts of characteristics, variables, constants, species, biological classification, normal variation, meiosis, gamete formation, random as- sortment, probability, gene, alleles, homozygous, heterozygous, blending in- heritance, dominant and recessive characteristics, genotype, and phenotype were introduced. instruction lasted for approximately 10 hours of class time and cul- minated in an explanation of observed ratios of offspring characteristics using principles of Mendelian genetics. During class time students were also asked to read pps. 229-237 from Concepts of Biology (Enger, Gibson, Kormelink, Ross, & Smith, 1982), which introduced relevant genetics concepts. Immediately follow- ing the reading, a five-question quiz over what was read was administered. At the conclusion of the instructional period, an achievement posttest was administered which consisted of multiple-choice, matching, classifying items, and standard textbook genetics problems.

Subjects

Subjects (Ss ) were 98undergraduate students (12 males and 86 females, age 18.0 years to 40.6 years, X = 23.1 years) enrolled in three sections of “Biological Science for the Elementary Teacher,” a course taught at Arizona State University.

Predictor Variables

Developmental Level. Ss’ hypothetico-deductive reasoning ability, or develop- mental level, was assessed by use of a slightly modified version of the Lawson Classroom Test of Formal Reasoning (Lawson, 1978). The modified test includes 13 items requiring Ss to isolate and control variables and use proportional, correla- tional, probabilistic, conservation, and combinatorial reasoning.

Each test item involves a demonstration using some physical materials used to pose a question or call for a prediction. S s responded in writing in individual test booklets, which contained only the questions followed by a number of pos- sible answers or space to generate an answer. S s were instructed to respond by checking the box next to the answer they thought correct and to explain why they chose that answer.

Scores on the test can be treated as a continuous distribution or grouped as follows: 0-5 = concrete operational reasoning; 6-9 = transitional reasoning; 10-

26 MITCHELL AND LAWSON

13 = formal operational reasoning. The split-half reliability of the modified test with the present sample was 0.60.

Disernbedding Ability. Ss’ ability to disembed relevant information from ir- relevant background was assessed by means of the GEFT Group Embedded Figures Test (Witkin et al., 1971). The GEFT is a timed test in which the S’s task is to locate and outline simple figures concealed in complex ones. Split-half reliability of the GEFT was 0.65.

Mental Capacity. FIT The Figural Intersection Test (Burtis & Pascual-Leone, 1974) was used to assess mental capacity. The FIT is a group test consisting of 42 items. For each item the S must place a point marking the intersection of from two to eight overlapping figures. An item with eight overlapping figures theoreti- cally requires a mental capacity of seven for successful completion, while an item with seven overlapping figures requires a mental capacity of six and so on. Scor- ing procedures are detailed in Bereiter and Scardamalia (1979). Split-half reliability of the FIT was 0.80.

Fluid Intelligence. The verbal (abstractions) section of the Shipley Hartford intelligence scale was used to measure fluid intelligence (Shipley, 1940). The sec- tion consists of 20 word-series completion items administered within a 10-minute time limit. Items range in degree of difficulty [e.g., (a) white-black short-long down- ; (b) tar-pitch-throw saloon-bar-rod fee-tip-end plank- -meals]. An S’s score on the section is simply the number of items correctiy completed. Lawson (1982) determined the split-half reliability of the series section for a similar sample of Ss to be 0.70.

Prior Knowledge. Prior knowledge of genctics was assessed by means of a 12-item test. There were 10 multiple-choice items worth 1 point each (two of these items appear below), and two problem-solving items worth two points each. One example is given below.

If characteristics are inherited, why is it that you don’t look exactly l i e your mother or father?

a. Because your inherited characteristics come from your mother if you are a female. b. Because half your inherited characteristics come from your mother and half from your

father. c. Because your inherited characteristics come from your father if you are a male. d. It is merely chance that you do not look exactly like one of them.

If dark eyes (D) are dominant over light eyes (1). and a male with the following al- leles (Dl) was mated with a female having the same alleles, their offspring could have?

a. only dark eyes b. only light eyes c. dark or light eyes d. no way to tell

PREDICTING GIWBTICS ACHIEVEMENT 21

The gene R for red is dominant over the gene w for white. If two flowers with a genotype of Rw are crossed, what percentage of the flowers produced will be red.

Dependent Variables

Reading Quiz. The reading quiz consisted of five multiple-choice items written to directly assess knowledge of what had been read. Two example items appear below.

A little e represents the recessive allele for attached earlobes. A capital E represents the dominant allele for unattached earlobcs. If a person had a genotype of EE he would be:

a. Homozygous for unattached earlobes b. Homozygous for attached earlobes c. Heterozygous for unattached earlobes d. Heterozygous for attached earlobes.

What characteristics would show up in a person if the following alleles were present in a person’s genotype, DdEe. E=free earlobes, D=dark hair, e=attached earlobes, d=light hair.

a. The person would havc dark hair and frec earlobes. b. The person would have light hair and frec earlobes. c . The person would have dark hair and attached earlobes. d. The person would have light hair and attached earlobes.

Achievement Posttest. To assess student genetics achievement a posttest was given at the conclusion of instruction. The posttest consisted of nine multiple- choice items similar to the ones on the pretest, 10 matching items, and three items requiring the S to identify variables and their values of a sample of eight “bugs” drawn in the test booklet and to develop a hierarchical classification scheme of the bugs using the identificd variables and their values (the “bugs” varied with respect to size, color, and shape).

The posttest also included seven genetics problems. The problems were slightly modified versions of those found in the problem section of Chapter 12 “Mendelian Genetics” from the student text Concepts in Biology (Enger et al, 1982). Examples of the multiple-choice items and the problems appear below.

If D for dark hair is dominant over d for light hair, which statement would be cor- rect?

a. a person with a DD genotype could have light or dark hair. b. a person with a Dd genotype would have dark hair. c. a person with a Dd genotype could have light or dark hair.

28 MITCHELL AND LAWSON

d. a person with a dd genotype would have dark hair.

If two organisms are not able to mate and produce viable offspring they are by definition not in the same:

a. species b. class c. phylum d. genus

A sperm or egg is referred to as a:

a. zygote b. gene c. chromosome d. gamete

If E = unattached earlobes, e = attached earlobes, W = wavy hair, w = straight hair, which statement is true about a person with the following genotype EeWw?

a. their phenotype would be attached earlobes and wavy hair b. their phenotype would be unattached earlobes and straight hair c. their phenotype would be attached earlobes and straight hair d. their phenotype would be unattachcd earlobes and wavy hair

If an offspring has the genotype Dd, what possible combinations of parental genotypes can exist?

In certain bean plants, the gene L for large pods is dominant over the 1 for short pods.

a. If both individuals are heterozygous, what will be thc genotypic and phenotypic ratios of the offspring? Show your work.

b. If a homozygous (i.e., both genes are the same) long and homozygous short are crossed, what will be the genotype and phenotype of the offspring? Show your work.

Eye color of the imaginary Grizzlcy Gronk population of the White Mountains varys. Some Gronks have purple eyes, some have white eyes and some have orange eyes. Professor Greengenes has discovered that whenever two purple-eyed Gronks mate, they always produce purple-eyed offspring. Likewise, whenever two orange-eyed Gronks mate they always produce orange-eyed Gronks. But when white-eyed Gronks mate they are able to produce offspring with all three colors of eyes.

a. Use Mendelian theory to explain how eye color is determined among Gronks (i.e., what is the genotype of the white-eyed Gronks and how can they mate to produce off- spring of all three colors of eyes?)

PREDICTING GENETICS ACHIEVEMENT 29

b. Use your theory to predict the phenotypic ratio of offspring if purple-eyed and white- eyed Gronks were mated. Show your work.

An achievement total score was recorded for each S consisting of the sum of the scores of all the items on the test. The split-half reliability of the posttest was calculated to be 0.84.

Results

On the Classroom Test of Formal Reasoning, 15.7% of the Ss scored from 0-5 and were classified as concrete operational, while 53.0% scored from 6-9 and were classified as transitional and 31.3% scored from 10-13 and were classified as for- mal operational.

On the Group Embedded Figures Test, 33.7% of the students scored from 0-6 and were classified as field dependent, 41.9% of the Ss scored from 7-13 and were classified field intermediate, and 24.4% of the Ss scored from 14-20 and were classified as field independent.

Application of the Bereiter & Scardamalia (1979) classification procedure to scores on the Figural Intersection Test resulted in 6.3% of the Ss with an assessed mental capacity less than six, 26.6% with an assessed mental capacity of six, and 67.1% with a mental capacity of seven.

On the Shipley Hartford test of fluid intelligence 9.3% of the Ss scored from 10-14 and were classified as having low fluid intelligence, 36.0% of the Ss scored from 15-17 and were classified as having intermediate fluid intelligence, and 54.7% of the Ss scored from 18-20 and were classified as having high fluid intel- ligence.

On the test of prior genetics knowledge 33.7% of the Ss scored from 2-6 and were classified as having low prior knowledge, 59.3% of the Ss scored from 7-10 and were classified as having an intermediate amount of prior knowledge, while 7.0% of the Ss scored from 11-14 and were classified as having a relatively high amount of prior knowledge.

Relationships with Achievement

Table I shows the percent success of each student group on the Reading Quiz and on each of the four sections of the achievement posttest (Multiple-choice Items, Classification Items, Matching Items, and Genetics Problems). One-way analysis of variance was used to determine whether group differences reached statistical significance. As can be seen, developmental level group differences reached significance (pc0.001) on the Reading Quiz, the Classification Items, and the Genetics Problems. Prior Knowledge group differences were significant (p0.05) for only the Genetics Problems. Disembedding Ability group differences reached significance Q ~ 0 . 0 5 or p0.01) on all but the Matching Items, while Men- tal Capacity group differences were significant ( ~ ~ 0 . 0 5 ) for only the Genetics Problems. Fluid Intelligence group differences were significant (p<O.OOl) for all but the Matching Items. Notice also that level of achievement is about the same on

TABL

E I.

Perc

ent S

ucce

ss o

n R

eadi

ng Q

uiz and

Ach

ieve

men

t Pos

ttest

by

Stud

ent G

roup

Ach

ieve

men

t Pos

ttest

Item

s St

uden

t R

eadi

ng

Gro

up

Quiz

Mul

tiple

C

lass

ifica

tion

Mat

chin

g Pr

oble

ms

Cho

ice

w 0

Dev

elop

men

tal L

evel

Tran

sitio

nal (n=44)

63

91

89

94

62

Form

al (n

=26

) 77

93

97

95

73

Con

cret

e (n=13)

54***

85N

S 66***

93NS

43***

Prio

r K

now

ledg

e Lo

UI (n

=29)

In

term

edia

te (n

=5 1

) H

igh

(n=6

) D

isem

bedd

ing A

bilit

y Lo

w (n

=29)

In

term

edia

te (-36)

Hig

h (n

=21

) M

enta

l Cap

acity

M

< 6

(n=

5)

M =

6 (n

=21

) M =

7 (n

=53

) Fl

uid

Inte

llige

nce

Low

9 (n

=8)

Inte

rmed

iate

(n=

31)

Hig

h (n=47)

61NS

89Ns

86"'

93NS

58*

69

90

89

94

63

70

95

100

97

85

59*

86*

76**

50**

68

93

92

93

67

72

92

97

98

71

5ZNS

91NS

73Ns

9 gNS

51*

65

86

86

90

54

69

92

93

95

69

55**

* 63***

36**

* 58

91

85

96

58

73

92

94

93

70

*p<0

.05.

p<

O.O

l. p<

0.00

1.

**

***

PREDICTING GENETICS ACHIEVEMENT 31

all the dependent measures for all the student groups with the possible exception of the high Prior Knowledge group, which was 12 percentage points higher than the next best group on the Genetics Problems.

Intercorrelations among Predictor Variables

Table I1 shows the intercorrelations among the predictor variables of Develop- mental Level, Prior Knowledge, Disembedding Ability, Mental Capacity, and Fluid Intelligence before and after correction for attenuation. For purposes of this analysis, Ss’ raw scores, instead of student “types,” were used. As can be seen, in- tercorrelations ranged from -0.09 between Prior Knowledge and Fluid Intelligence to 0.7 1 between Disembedding Ability and Fluid Intelligence. This coefficient ac- tually reached greater than unity (1.06) after correction for attenuation. A coeffi- cient greater than unity is, of course, theoretically impossible, thus this indicates that the reliability estimates were too conservative for these two measures. Never- theless, the high coefficient suggests that, for the present sample at least, the two measures are assessing the same cognitive dimension. Other moderate coefficients indicate a substantial amount of nonindependence of the predictor variables.

Multiple Regression Analyses

Due to the fact that the predictor variables were not entirely independent of one another, stepwise multiple regression analyses were performed to determine which of the predictor variables proved to be the best predictor(s) of achievement score variance. The results of these analyses are shown in Table 111.

In all, six stepwise multiple regression analyses were performed, one for the Reading Quiz, one for each of the four types of Achievement Posttest items, and one for the Achievement Posttest total score. As shown in Table 111, Developmen- tal Level and Fluid Intelligence were the only significant predictors of Reading Quiz score. Together they accounted for 21.4% of score variance. Due to the rela- tively high level of student performance, and generally similar performance of all student groups on the Multiple-Choice and Matching Items, none of the cognitive variables proved to offer a unique and significant contribution to predicting these scores, and thus are not included in Table 111. Disembedding Ability was the only variable entered into the regression equation for the Classifying Items (12.2% of variance accounted for), while Developmental Level and Mental Capacity were the only variables entered into the regression equation for the Genetics Problems. Together they accounted for 20.6% of score variance. Likewise only Developmen- tal Level and Mental Capacity were found to account for a significant amount of variance in the prediction of total score on the Achievement Posttest (21.2% of variance).

W

hl

TABL

E II.

In

terc

orre

latio

n M

atrix

for P

redi

ctor

Var

iabl

es

Var

iabl

e Num

ber a

nd N

ame

1

2 3

4 5

1. D

evel

opm

enta

l Lev

el (

0.60

)a

1 .oo

2. Pr

ior

Kno

wle

dge

(-)

3. D

isem

bedd

ing

Abi

lity

(0.6

5)

4. M

enta

l Cap

acity

(0.8

0)

(0.4

8)b

(0.4

1)

(0.1

7)

0.26

0.

30**

* 0.

28**

* 0.

11

1.00

0.

12

1 .oo

0.27

**

0.30

(0

.421

**

1 .oo

-0.9

(1

.06)

0.

71*’

* (0

.40)

0.

30**

*

5. FI

uid

Inte

llige

nce

(0.7

0)

aSpl

it-ha

lfrel

iabi

lity.

bF

igur

es in

par

enth

eses

repr

esen

t coe

ffic

ient

s cor

rect

ed fo

r atte

nuat

ion.

*p

<0.0

5.

1 .oo

** p<

O.O

l. p<

O.O

Ol.

***

PREDICTING GENETICS ACHIEVEMENT 33

Discussion and Implications

Stewart (1982) argues that a major source of student difficulty in solving genetics problems stems not from lack of reasoning skills (e.g., inability to use combinatorial reasoning) but from lack of adequate knowledge of genetics and meiosis. Due to lack of specific details in the Stewart (1982) article, this may cer- tainly have been the case for his students. Yet for students in the present study, it seems clear that lack of appropriate hypothedco-deductive reasoning skills (e.g., combinatorial, probabilistic, and proportional reasoning ability) as reflected by per- formance on the test of developmental level, was the major source of difficulty, not only in solving genetics problems and in constructing classification schemes, but in interpreting text material on genetics as well.

A word of caution must be voiced, however. As this study was descriptive in nature, the results at best show a correlation between developmental level and genetics problem-solving skill, not a cause-effect relationship. Nevertheless, the assumption of a causal relationship seems to us to be very justifiable, simply. be- cause the measure of developmental level tested students’ ability to solve problems utilizing schemes of proportions, combinations, probability, etc. in familiar contexts, while the genetics problems required use of those schemes in hypothetical contexts. Students failing to apply these schemes in familiar concrete contexts can hardly be expected to apply them in the context of a hypothetical system of gene transfer. Thus it only seems reasonable to conclude that lack of ability to use these schemes is a cause of failure in solving genetics problems. This is not to say that lack of these reasoning schemes is the only cause of failure. Certainly, as Stewart and others have pointed out, domain-specific knowledge is also necessary. However, for the present students the results point to the con- clusion that lack of reasoning skills, rather than lack of domain-specific knowledge, is limiting performance. Note not only the failure of the test of prior knowledge to contribute to the regression equation for genelics problem solving, but also the relatively high level of performance of all students on the multiple- choice and matching items. It appears as though the instruction was fairly suc- cessful in transferring the domain-specific knowledge required to respond successfully to these items but this knowledge was not sufficient to “cause” suc- cessful performance on the genetics problems. The concrete operational students were successful on 85% of the multiple-choice items and 93% of the matching items but only 43% of the genetics problems.

Findley (1983) cites schema theory in support of the position that construct- ing meaning from text depends upon student’s prior knowledge of the subject mat- ter. Certainly there is nothing in the present results to suggest that prior knowledge is irrelevant; however, the results clearly indicate that with respect to comprehension of text material in genetics, developmental level plays a very im- portant role (more so than prior knowledge as presently assessed). Taken together the results tend to be consistent with the view that the application of specific prin- ciples from written text or from classroom instruction requires a general operativity which serves as a framework for the assimilation of specific input. Students lacking such a framework are simply unable to correctly assimilate and apply the specific principles.

The issue of just what unique cognitive dimensions are being measured by the

34 MITCHELL AND LAWSON

tests of fluid intelligence, disembedding ability and mental capacity is clouded by the present results as they tended to not be independent of one another among the present students. This issue is central to the general question of just what one means by the term “intelligence”; thus a detailed discussion lies beyond the scope of the present article. Yet the results are at least suggestive of a particular view and worthy of comment. Developmental level, fluid intelligence, disembedding ability, and mental capacity all contributed to the prediction of some aspect of achievement (see Table 111); thus they are presumably worthy of attention on the part of teachers and curriculum developers. The fact that lack of reasoning skills appears to be a cause of student difficulties suggests that measures should be taken to improve those skills (cf. Shymansky, 1984; Lawson, 1985; Karplus et al., 1977).

When teaching genetics the teacher shouId be aware of the reasoning abilities of his or her students. Certainly if students are predominantly concrete operation- al, extreme care must be taken not to quickly gloss over problems such as the generation of combinations of gametes, the ratios, proportions, and probability of occurrence of offspring with specific traits. Rather than giving quick treatment or algorithmic strategies such as Punnett squares, the teacher should be sensitive to these sources of difficulty and, whenever possible, use the situations to help stu- dents gain a better grasp of the reasoning processes required. Compare, for ex- ample, the careful treatment of Mendel’s work and his thinking found in Starr &

TABLE ID. Stepwise Multiple Regression Summary for the Prediction of Reading Quiz Performance

and Achievement Posttest Performance

Significant Multiple R R Square Cumulative Predictor % of Variance Variables Explained

Reading Quiza 1. Developmental Level 0.38 0.14 14.3% 2. Fluid Intelligence 0.46 0.21 21.4% Classfiing Itemsb 1. Disembedding Ability 0.35 0.12 12.2% Genetics Problems’ 1. Developmental Level 0.39 0.15 15 .O% 2. Mental Capacity 0.45 0.21 20.6% Achievement Posttest Totap 1. Developmental Level 0.38 0.14 14.4% 2. Mental Capacity 0.46 0.21 21.2%

aReading quiz score = 1.08 + 0.47(DL) + 0.47(FI). bClassifying items score = 2.16 + 0.28(DA). ‘Genetics problems score = 4.03 + 2.96(DL) + 2.02(MC). dAchievement posttest total score = 19.39 + 3.57(DL) + 2.64(MC).

PREDICTING GENETLCS ACHIEVEMENT 35

Taggart (1981, pp. 162-163) to the cryptic treatment in Mader (1985, pps. 135- 136). By carefully detailing Mendel’s observations, questions, hypotheses, predic- tions, and results, Starr & Taggart give the student an opportunity to “see” Mendel’s reasoning, which should help students not only better appreciate the na- ture of scientific thought but should help them develop more advanced patterns of reasoning themselves. On the other hand, the cryptic treatment df Mendel’s “Laws” offered by Mader would seem to only further confuse students about how science operates, and thus only make it more difficult to use the topic of genetics as a vehicle for advancing student reasoning.

Fluid intelligence/disembedding ability (taken as one’s ability to recognize patterns of environmental regularity which are embedded in potentially mislead- ing contexts) not surprisingly seems to play a role in text reading, comprehension, and application. Presumably the teacher has little or no control of this student variable, and thus should be aware of instructional situations that are potentially confusing and should, whenever possible, simplify them by the elimination of un- necessary details or help students recognize the key patterns by the use of careful- ly chosen analogies with familiar situations. In theory these instructional strategies should help students overcome difficulties due to a smaller than normal mental capacity as well. On the other hand, an argument could also be made for the inclusion of more situations and problems that are potentially confusing due either to the inclusion of too much or too little information or to unfamiliarity. Working through such situations and problems may help students improve their disembedding and pattern-recognition skills. At present, evidence suggests that al- though dramatic improvements cannot be expected quickly, there is no reason to believe that they are not possible (cf, Whimby & Whimby, 1978). This appears to be a fertile area for future research.

In conclusion, evidence has been gathered which indicates that a general lack of hypothetico-deductive reasoning skills limits students’ ability to comprehend text material and apply principles of Mendelian genetics in the solution of problems. Lack of reasoning skills does not appear to significantly impede perfor- mance on matching and multiple-choice items that require, in the Bloom (1956) sense, lower-level cognitive abilities. The cognitive variables of fluid intelligence, disembedding ability, and mental capacity were also found to significantly predict certain aspects of achievement, while level of prior domain-specific knowledge did not.

References

Bereiter, C. & Scardamalia, M. (1979). Pascual-Leone’s M construct as a link between cognitive development and psychometric concepts of intelligence. Intel- ligence, 3 , 4 1 4 3 .

Bloom, B.S. (1956). Taxonomy of educational objectives: Cognitive domain. New York: Longmans, Green and Co.

Burtis, P.J. & Pascual-Leone, J. (1974). FIT: Figural interaction test. A group measure of M-space. Unpublished manuscript, York University.

Champagne, A.B. & Klopfer, L.E. (1982). A causal model of student’s

36 MITCHELL AND LAWSON

achievement in a college physics course. Journal of Research in Science Teach- ing, 19(4), 299-309.

Costello, S.J. (1982). The relationships among logical and spatial skills and understanding genetics concepts and problems. Paper presented at the Annual Convention of the National Association for Research in Science Teaching, Lake Geneva.

Enger, E.D., Gibson, A.H., Kormelink, J.R., Ross, F.C., & Smith, R.J. (1982). Concepts in biology (3rd ed.). Dubuque, IA: Wm. C. Brown Co.

Finley, F.N. (1983). Students’ recall from science text. Journal of Research in Science Teaching, 20(3), 247-259.

Fleming, M.L. & Malone, M.R. (1983). The relationship of student charac- teristics and student performance in science as viewed by meta-analysis research. Journal of Research in Science Teaching, 20(5), 481495.

Gipson, M. & Abraham, M.R. (1985). Relationships between formal-opera- tional thought and conceptual difficulties in genetics problem solving. Paper presented at the Annual Convention of the National Association for Research in Science Teaching, French Lick.

Gosnell-Moses, D.J. & Barufaldi, J. (1984). A comparison of cognitive development, field independence/dependence cognitive style, and academic suc- cess of baccalaureate nursing students. Paper presented at the Annual Convention of the National Association for Research in Science Teaching, New Orleans.

Hackling, M.W. & Treagust, D. (1984). Research data necessary for meaning- ful review of grade ten high school genetics curricula. Journal of Research in Science Teaching, 21(2), 197-209.

Jolly, P.E. & Strawitz, B.M. (1983). Teacher-student cognitive style and achievement in biology. Paper presented at the Annual Convention of the National Association for Research in Science Teaching, Dallas.

Karplus, R., Lawson, A.E., Wollman, W., Appel, M., Bernoff, R., Howe, A., Rusch, J.J., & Sullivan, F. (1976). Science teaching and the development of reasoning: A workshop. Berkeley: Regents of the University of California.

Lawson, A.E. (1978). The development and validation of a classroom test of formal reasoning. Journal of Research in Science Teaching, 15(1), 11-24.

Lawson, A.E. (1982). Formal reasoning, achievement, and intelligence: An issue of importance. Science Education, 66( l), 77-83.

Lawson, A.E. (1983). Predicting science achievement: The role of develop- mental level, disembedding ability, mental capacity, prior knowledge, and beliefs. Journal of Research in Science Teaching, 20(2), 117-129.

Lawson, A.E. (1985). A review of research on formal reasoning and science teaching. Journal of Research in Science Teaching, 22(7), 569-6 18.

Lawson, C.A. (1958). Language, thought, and the human mind. East Lansing: Michigan State University Press.

Mader, S . S . (1985). Biology: Evolution, diversity, and the environment. Dubuque, IA: Wm. G. Brown.

Morgenstern, C.F. & Renner, J.W. (1984). Measuring thinking with stand- ardized science tests. Journal of Research in Science Teaching, 21(6), 639-348.

Novak, J.D. (1979). The reception learning paradigm. Journal of Research in Science Teaching, 16(6), 48 1-488.

PREDICTING GENETICS ACHIEVEMENT 37

Oldham, V. & Brouwer, N. (1984). Mendelian genetics: Paradigm, conjecture, or research program. Journal of Research in Science Teaching, 21(6), 623637.

Pascual-Leone, J. (1969). Cognitive development and cognitive style: A general psychological integration. Unpublished doctoral dissertation. University of Geneva.

Ronning, R.R., McCurdy, D., & Ballinger, R. (1984). Individual differences: A third component in problem-solving instruction. Journal of Research in Science Teaching, 21(7), 71-82.

Scardamalia, M. (1977a). Information processing capacity and the problem of horizontal decalage: A demonstration using combinatorial reasoning tasks. Child Development, 48,28-37.

Scardamalia, M. (1977b). The interaction of perceptual and quantitative load factors in the control of variables. Ontario: York University, Department of Psychology Reports.

Shipley, W.C. (1940). A self-administering scale for measuring intellectual impairment and deterioration. Journal of Psychology, 29, 371-377.

Shymansky, J. (1984). BSCS programs: Just how effective are they? The American Biology Teacher, 46(1), 54-57.

Smith, M.V. & Good, R. (1984). Problem solving and classical genetics: Suc- cessful versus unsuccessful performance. Journal of Research in Science Teach- ing, 21(9), 895-912.

Starr, C. & Taggart, R. (1978). Biology: The unity and diversity of life (3rd ed.). Belmont, CA: Wadsworth.

Stewart, J.H. (1982). Difficulties experienced by high school students when learning basic Mendelian genetics. The American Biology Teacher, 44(2), 80-84, 89.

Whimby, A. & Whimby, L.S. (1978). Intelligence can be taught. New York: Dutton & Co., Inc.

Witkin, H.A., Moore, C.A., Goodenough, D.R., & Cox, P.W. (1977). Field-de- pendent and field-independent cognitive styles and their educational implications. Review of Educational Research, 47( l), 1-64.

Manuscript accepted July 30,1987