cognitive modeling of student learning in web-based ...plyons/papers (by others)/hci/web/chen... ·...

29
INTERNATIONAL JOURNAL OF HUMAN-COMPUTER INTERACTION. 17\3). 375-402 Copyright © 2004. Lawrence Erlbaum Associates, Inc. Cognitive Modeling of Student Learning in Web-Based instructional Programs Sherry Y. Chen Robert D. Macredie Department of Information Systems and Computing, Brunei University There has been tremendous growth in Web-based instruction over the past few years. Because the user group of Web-based instruction includes learners from heterogeneous backgrounds, individual differences become an important issue in the development of Web-based instructional programs. Among a variety of individual differences, cogni- tive style is a particularly important characteristic. This study aims to determine the re- lationships between learners' cognitive styles and their perceptions and attitudes to- ward the features of a Web-based instructional program. The results indicate that cognitive styles influence students' reactions to nonlinear interaction, independent learning, and navigation tools and the difficulties and problems that they encounter. The findings are applied to develop a learning model that can support the design of Web-based instructional programs. 1. INTRODUCTION As the World Wide Web has grown to become a major channel for business communi- cations, entertainment, and information exchange, it has also begun to be seen as a preferred technology to improve instruction in higher education (MacArthur & Lewis, 1996). Unlike other software with access confined to a rather homogeneous group of users. Web-based instructional programs are used by a population of learn- ers who have far more heterogeneous backgrounds, in terms of their preferences, skills, and needs. The diversity in the user population results in a new challenge for instructional design. In response to this challenge, researchers need todirect more at- tention toward seeing how diverse populations are learning, accessing, and using Web-based instructional programs (Zoe & DiMartino, 2000). Therefore, empirical evaluation of learners' individual differences becomes paramount because such evaluation can provide concrete prescriptions for developing learner-centered pro- grams that can be matched with the particular needs of each individual. Requests for reprints should be sent to Sherry Y. Chen, Department of Information Systems and Computing, Brunei University, Uxbridge, Middlesex UB8 3PH, UK. E-mail: [email protected]

Upload: vuongdang

Post on 22-May-2018

214 views

Category:

Documents


0 download

TRANSCRIPT

INTERNATIONAL JOURNAL OF HUMAN-COMPUTER INTERACTION. 17\3). 375-402Copyright © 2004. Lawrence Erlbaum Associates, Inc.

Cognitive Modeling of Student Learning inWeb-Based instructional Programs

Sherry Y. ChenRobert D. Macredie

Department of Information Systems and Computing, Brunei University

There has been tremendous growth in Web-based instruction over the past few years.Because the user group of Web-based instruction includes learners from heterogeneousbackgrounds, individual differences become an important issue in the development ofWeb-based instructional programs. Among a variety of individual differences, cogni-tive style is a particularly important characteristic. This study aims to determine the re-lationships between learners' cognitive styles and their perceptions and attitudes to-ward the features of a Web-based instructional program. The results indicate thatcognitive styles influence students' reactions to nonlinear interaction, independentlearning, and navigation tools and the difficulties and problems that they encounter.The findings are applied to develop a learning model that can support the design ofWeb-based instructional programs.

1. INTRODUCTION

As the World Wide Web has grown to become a major channel for business communi-cations, entertainment, and information exchange, it has also begun to be seen as apreferred technology to improve instruction in higher education (MacArthur &Lewis, 1996). Unlike other software with access confined to a rather homogeneousgroup of users. Web-based instructional programs are used by a population of learn-ers who have far more heterogeneous backgrounds, in terms of their preferences,skills, and needs. The diversity in the user population results in a new challenge forinstructional design. In response to this challenge, researchers need todirect more at-tention toward seeing how diverse populations are learning, accessing, and usingWeb-based instructional programs (Zoe & DiMartino, 2000). Therefore, empiricalevaluation of learners' individual differences becomes paramount because suchevaluation can provide concrete prescriptions for developing learner-centered pro-grams that can be matched with the particular needs of each individual.

Requests for reprints should be sent to Sherry Y. Chen, Department of Information Systems andComputing, Brunei University, Uxbridge, Middlesex UB8 3PH, UK. E-mail: [email protected]

Chen and MacredJe

In this vein, the study reported in this article was designed to examine hovi' indi-vidual differences influence students' reactions to Web-based instructional pro-grams. Among a variety of individual-differences approaches, this study focusedon cognitive styles because they influence the effectiveness of teaching and learn-ing (Miller, 1987). The research addressed a specific question: "What are the effectsof students' cognitive styles on their perceptions and attitudes toward learningwithin a Web-based instructional program?" Answers to this question were soughtby analyzing students' responses to the features provided by a Web-based pro-gram. On the basis of the findings, the authors offer suggestions to improve the de-velopment of Web-based instructional programs.

This article begins by building a theoretical framework to present the relation-ships between cognitive styles and Web-based instructional programs, followed bythe proposition of five hypotheses developed from an analysis of previous studies.It then progresses to discuss an empirical study of students' learning experiences ina Web-based instructional program. Subsequently, the findings of this empiricalstudy are used to frame a design model, which illustrates how to integrate the pref-erences of each cognitive style into the development of Web-based instructionalprograms.

2. THEORETICAL RATIONALE AND HYPOTHESES

2.1. Web-Based Instruction

The values of Web-based instruction reflect its flexibilities. Traditional com-puter-based instructional programs present information in a linear fashion.Web-based instructional programs make use of hypermedia capabilities, whichpermit much more flexibility in the delivery of instruction by enabling users to se-lect hypertext links (Federico, 2000). Such flexibility offers learners a rich explora-tion environment and encourages them to navigate by association. In this way, theycan then construct their own individualized knowledge structure by cross-refer-encing related topics in their knowledge base. Therefore, learners are able to followpaths through the subject content produced by designers or to develop their ownroutes according to individually prescribed requirements (Large, 1996). The otherflexibility that Web-based instruction provides is that learners can read course con-tent through a computer network at any time and from different places (Chang etal., 1998), which is why many educators have tried to develop distance learningprograms on the Web. These flexibilities provide learners with many opportunitiesto explore, discover, and learn according to their own individual needs.

However, the freedom offered by Web-based instructional programs may comeat a price because flexibility increases complexity (Ellis & Kumiawan, 2000). Ngand Gunstone (2002) indicated that although students had positive perceptionsabout self-based learning provided by Web-based instruction, the unstructured na-ture of the Web made some students need more time to search for information.Power and Roth (1999) reported that Web-based instruction is more dynamic andricher than other learning material but that it creates new challenges related to the

Cognitive Modeling in Web-Based Instruction 377

effect on learners' comprehension. Andrew (2001) pointed out that use of the Webopens doors for students to explore a vast universal resource, but students still findit difficult to shed the hard copy habit. In addition, Y. Quintana (1996) found that al-though students gained the advantage of flexibility in time, pace, and distance withWeb-based instruction, many of them felt isolated, suffered from a lack of motiva-tion or lack of support, and found that the feedback provided was too limited, andconsequently they dropped out of their courses. These studies provide evidencethat not all types of students appreciate being given freedom in their learning pro-cesses. In particular, students who need more guidance through the learning pro-cess may meet an increased number of problems in using Web-based instructionalprograms.

To address this limitation, an instructional program should be developed to sup-port the unique needs of each individual learner (Carter, 2002). Only when learn-ers' needs are identified can developers of programs effectively enhance function-ality and increase learners' satisfaction (Ke, Kwakkelaarb, Taic, & Chenc, 2002). Inother words, accommodation of the learner is the central issue in the design pro-cess. This is the reason why recent research has seen a shift toward emphasizinglearner-centered design (Federico, 1999), which is theoretically motivated bysociocultural and constructivist theories of learning (Soloway et al, 1996). This de-sign approach claims that the development of an instructional program should bebased on the learner's point of view (Soloway, Guzdial, & Hay, 1994) and addressthe needs of learners (C. Quintana, Krajcik, & Soloway, 2000).

Leaner-centered design is imperative, and understanding students' needs is ofprimary significance. In particular, it is critical to Web-based instructional pro-grams because much more heterogeneity exists among students (Soloway & Pryor,1996). A prominent issue is to recognize how the different needs of the variouslearners involved with a Web-based instructional program may influence their per-formance and satisfaction. To develop effective Web-based instructional programs,it is necessary to consider how to accommodate learners' differences. Therefore, in-dividual differences arguably become an important consideration in the researcharea of Web-based instruction. A number of learner-centered studies have shownthat individual differences have a strong impact on the use of instructional technol-ogy (Marchionini, 1995). An analysis of existing pedagogical studies also con-firmed that the successful usage of instructional technology depends on the tech-nology itself and the users' individual characteristics (H. Chou & Wang, 2000). Forthese reasons, research into individual differences on Wcb-based instruction hasmushroomed in the past decade. The examined differences include cognitive style(Durfresne & Turcotte, 1997; Shih & Gamon, 1999), gender differences (Ford &:Miller, 1996; Leong & Al-Hawamdeh, 1999), system experience (Chen & Ford, 1998;Reed & Oughton, 1997), and domain knowledge (Lawless & Kulikowich, 1998).Among these differences, cognitive style has been identified as one of the most per-tinent factors because it refers to a user's information-processing habits, represent-ing an individual user's typical mode of perceiving, thinking, remembering, andproblem solving (Messick, 1976). It has also been suggested that teachers should as-sess the cognitive styles of their students to design instructional strategies for opti-mal learning (Lee, 1992).

378 Chen and Macredie

2.2. Cognitive Styies

Cognitive style refers to an individual's preferred and habitual approach to organiz-ing and representing information (Riding & Rayner, 1998). Within the area of cogni-tive styles, field dependence versus field independence has emerged as one of themost widely studied dimensions with the broadest application to problems of edu-cation (Messick, 1976; Witkin, Moore, Goodenough, & Cox, 1977) because it reflectshow well a learner is able to restructure information based on the use of salient cuesand field arrangement (Weller, Repman, & Rooze, 1994).

Witkin et al. (1977) found that field dependent people have global perceptions,whereas field independent people are good at analytical thought. Previous research(e.g., Jonassen & Grabowski, 1993; Riding & Cheema, 1991) has noted conceptuallinks between field dependence-field independence and the holist-serialist classifi-cation proposed by Pask (1988). Field dependent individuals typically perceive ob-jects as a whole and approach a task more holistically; field independent individualsfocus on individual parts of the object and tend to be more serialistic in their ap-proach to learning. Similar to field dependent individuals, holists process informa-tion in relatively global ways in that they tend to concentrate first on building anoverall picture of the subject area, into which they subsequently fit procedural detail.Conversely, serialists have a learning pattern similar to that of field independentlearners; they tend to maintain a local focus, concentrating on one thing at a time andon building up procedural understanding step by step. Evidence can be found inFord and Chen (2001), who developed two Web sites with two different navigationpaths: One used a depth-first path, and the other one used a breadth-first path. In thecase of the depth-first path, each topic was presented in detail before the next topic(i.e., the serialistic condition), whereas the breadth-first path gave an overview of allmaterial prior to introducing detail (i.e., the holistic condition). Field dependent us-ers performed better following the breadth-first path, whereas field independent us-ers did better following the depth-first path.

In addition to the difference between global and analytical approaches, anotherdistinction is that field dependent individuals rely more on external references; incontrast, field independent individuals rely more on internal references(Goodenough, 1976). This is reflected in the differences of their behaviors in the in-terpersonal domain (Witkin & Goodenough, 1981). Field dependent individualspay more selective attention to social cues; they favor situations that bring theminto contact with others and have the ability to get along with others. Conversely,field independent individuals tend to be more autonomous and show initiative,self-reliance, and the ability to think for themselves. The tendencies to rely primar-ily on external or internal references also affect their performance on cognitive re-structuring tasks. Field dependent individuals are more influenced by for-mat-structure, whereas field independent individuals are less affected byformat-structure (Jonassen & Grabowski, 1993). Davis and Cochran (1989) ex-plained this phenomenon as field dependent learners being more reliant on salientcues in learning. Conversely, field independent individuals tend to sample morecues inherent in the field and are able to extract the relevant cues necessary for thecompletion of a task.

Cognitive Modeling in Web-Based Instruction 379

The other differences between field independent and field dependent individu-als are that the former are likely to use active cognitive strategies, and the latterhave a tendency to use passive strategies such as rehearsal (Frank & Keane, 1993).According to Witkin et al. (1977), field independent individuals adopt a hypothe-sis-testing role in learning; conversely, field dependent individuals adopt a specta-tor role in learning. Even (1982) noted that people who exhibit field independentstyles are likely to benefit from a self-directed emphasis. Yet, field dependent learn-ers also tend to prefer more structured learning environments, suggesting that fielddependent learners may benefit from reorganization of the learning material tomake the organizational structure more explicit (Chapelle &: Jamieson, 1986).

The aforementioned differences in learner characteristics (global vs. analytical,external vs. internal, passive vs. active) may influence individuals' learning strate-gies in Web-based instruction. Several studies have shown that there is an interest-ing correlation between Reid dependence and learning behaviors in Web-based in-structional programs. For example, Chen and Ford (1998) conducted a study inwhich they presented a Web-based instructional program with nonlinear formatsto introduce students to artificial intelligence. They administered Riding's (1991)Cognitive Styles Analysis (CSA) to assess each participant's level of field depend-ence. The results indicated that field dependent students used the mairi menu moreoften than field independent students. Furthermore, Kim (2001) investigated theeffects of cognitive style on Web searching experience. The Group Embedded Fig-ures Test (GEFT) by Witkin, Oltman, Raskin, and Karp (1971) was administered toidentify participants' cognitive styles. They found that field dependent novicestended to follow links prescribed by the Web page and to experience more disorien-tation problems. As a result, they suggested that field dependent individuals, espe-cially when novices, might need special attention from interface designers andthose who train Web users.

Results from these studies suggest that different cognitive style groups use dif-ferent learning strategies in Web-based instruction. These studies also indicate thatfurther empirical work is needed to identify the learning preferences of differentcognitive style groups, the results of which might help to guide the developmentand evaluation of Web-based instructional programs. This article presents such astudy, which aims to examine how cognitive styles influence students' responses tothe interface features of a Web-based instructional program, and subsequently de-velop support n:\echanisms for designers.

2.3. Developing a Rationale for the Hypotheses

As discussed in Section 2.1, Web-based instruction programs make use ofhypermedia techniques; in turn, the features of hypermedia learning may also influ-ence learners' reactions to Web-based instruction, including nonlinear learning,learner control, and multiple tools. Chen and Macredie (2002) presented a compre-hensive review of empirical studies concerning the effects of learners' cognitivestyles from 1989 to 2001, particularly focused on Witkin's field dependence-field in-dependence, on the effectiveness of hypermedia learning. On the basis of an analysis

Chen and Macredie

of the results of previous studies, they built a learning model that suggests that fielddependent learners prefer guided navigation and field independent learners preferfree navigation, and that field dependent learners need more guidance and tend touse maps to process information globally, whereas field independent learners enjoyindependent learning and tend to be analytical, locating specific information usingan index. The model revealed that learners with different cognitive styles showeddifferent reactions to nonlinear interaction and independent learning, and they fa-vored using different navigation tools. Thus, this research ieads to three hypotheses:

• Hypothesis 1: Cognitive styles will significantly influence learners' attitudestoward nonlinear interaction.

• Hypothesis 2: Significant interactions will be found between learners' cogni-tive styles and their reactions to independent learning.

• Hypothesis 3: Different cognitive style groups will favor different types ofnavigation tools provided by the Web-based instructional program.

According to the available literature, cognitive styles may influence learners' re-quirements related to content presentation. Chen and Ford (1998) examined the re-lationships between students' cognitive styles and learning strategies (described inSection 2.2). In addition to the link between students' cognihve styles and thechoice of navigation tools, the research also showed that students with differentcognitive styles showed different requirements in relation to content presentation.Field independent students preferred content to be presented in a logical way, in-termediate students regarded the presented content as too superficial, and field de-pendent students judged the content to be too detailed. Hence, one may hypothe-size as follows:

• Hypothesis 4: Learners with different cognitive styles will have different re-quirements in terms of content presentation within Web-based instruction.

Previous studies have also indicated that cognitive styles might have significanteffects on students' learning performance. Field independent students, who have ahigher ability to engage in independent learning, perform better than field depend-ent students, who are less capable of learning independently (Boyce, 1999; Chuang,1999; Umar, 1999). In this study, learning performance was measured by the per-ceived confidence that the learner has in his or her understanding of subject con-tent. Thus, a final hypothesis might be proposed:

• Hypothesis 5: Cognitive styles will have a significant effect on the confidence re-lated to the understanding of the subject content within Web-based instruction.

3. METHODOLOGY DESIGN

The aforementioned five hypotheses were tested using quantitative measurementin which the data obtained from the closed questions of the questionnaire were sta-

Cognitive Modeling in Web-Based Instruction 381

tistically analyzed to build up a picture of the mapping of the relationships be-tween students' cognitive styles and their learning preferences. In addition, datafrom open-ended questions were used for qualitative evaluation to complement amore detailed analysis of the diversity in students' learning preferences accordingto their cognitive styles. The intention of using both quantitative measurement andqualitative evaluation was to overcome the individual weaknesses and biases of ei-ther approach used in isolation (Rossman & Wilson, 1985).

3.1. Participants

This study was conducted at Brunei University's Department of Information Sys-tems and Computing. A total of 61 master's students participated in this study, rep-resenting 68*>1> of the population of the master's courses. All participants had thebasic computing and Internet skills necessary to operate a Web-based instructionalprogram. At the outset, they were inexperienced in the content domain of HTMLauthoring. Despite the fact that the participants volunteered to take part in the ex-periment, they were evenly distributed in terms of cognitive styles based on theCSA test results {see the Cognitive Styles Analysis section). In addition, the partici-pants were an almost equal mix of male and female within each cognitive stylegroup. Table 1 illustrates the distribution of the sample of this study.

3.2. Task Design

In this study, participants were asked to complete specific tasks by themselves. Thetask activities involved constructing a home page using Notepad, to measure thelevels of understanding of HTML. Before the participants interacted with theWeb-based instructional program, they each received a paper-based task sheet. Onthe basis of this task sheet, the participants were asked to locate specific informa-tion to complete the tasks successfully. Scoring consisted of summing the success-fully completed items. The participants were allowed to decide the order in whichthey attempted the tasks on the sheet. One and a half hours were allocated for eachparticipant to use the program and complete the task activities.

Researchers designed the task to address 10 activities covering different areas,such as creating hypertext links, changing background colors, and producing tables.The aim was to have tasks be at a level of complexity that would maintain motivationin the participants {Scanlon, 2000). Furthermore, these 10 activities were aligned

Cognitive Style

Field independentIntermediateField dependentTotal

Table 1: The Distribution

Women

10109

29

of the Sample

Men

13118

32

Total

23211761

382 Chen and Macredie

with key areas of content within tbe Web-based instructional program, which meantthat participants were forced to encounter all interface features. In this way, the au-thors were able to identify how members of different cognitive style groups per-ceived various interface features provided by the Web-based instructional program.

3.3. Research Instruments

Research instruments work as a guide to make sure that the same information isobtained from different students. The research instruments used in this study in-cluded a Web-based instructional program to teach students how to use HTML,Riding's CSA to measure students' cognitive styles, and an exit questionnaire toidentify students' perceptions and attitudes toward the Web-based instructionalprogram. The following sections introduce and explain these three instruments.

Web-based instructional program. Students interacted with a Web-basedinstructional program entitled How to Use HTML. The teaching-learning approachin this program was based on the concept of self-organized learning. In other words,students were given freedom to choose their own navigational routes through thesubject matter. Students could stud y topics and subtopics in any order. Three types ofnavigation control were available in this program as shown in Table 2.

The program began by introducing its learning objectives and explaining theavailable navigation approaches; it was presented in 82 pages using text, tables, anindex, and maps. The content was divided into three sections: Section 1, What isHTML?; Section 2, Working With HTML; and Section 3, Relations With SGML andWWW [World Wide Web]. Section 2, which covered 12 subtopics of HTMLauthoring, was the key element of the Web-based instructional program. Eachsubtopic was further split into five subject categories, consisting of (a) overview, (b)

Table 2: Three Types of Navigation Control

Contra! Purpose Tool

Sequence control To allow students to decidethe sequence of subjects tobe learned.

Content control To allow students to controlthe selection of thecontents they wish tolearn.

Display control To allow students to chooseone of several displayoptions that cover thesame concept.

Hierarchical maps: to show all topics andsub-topics in a hierarchical way.

Alphabetical index: to list keywords inalphabetical order.

Back-forward buttons: to see the page previouslyvisited.

Section buttons: to choose from three sectionsthat hold the main content.

Main menu: to present the main topics.Hypertext links: to connect relevant concepts.Display options: to include an overview,

examples, detailed techniques, and so forth.

Cognitive Modeling in Web-Based Instruction 383

detailed techniques, (c) examples, (d) related skills, and (e) references, so that anal-yses of students' preferred content presentation could be undertaken by examiningthe students' navigation paths and their replies to items in the questionnaire.

Interface elements included the following: (a) a title bar located at the top of thescreen showing the section name being viewed and the other available section but-tons; (b) a control panel with the choices for menu, map, index, and quit buttons;and (c) the main body of the program, providing referenced links and subject cate-gories for selection. Figure 1 shows the screen design of this program.

Cognitive Styles Analysis. The cognitive style dimension investigated inthis study was the level of field dependence. A number of instruments have beendeveloped to measure field dependence, including the GEFT by Witkin et al. (1971)and the CSAby Riding (1991). The GEFT derives scores for field independence byrequiring participants to locate simple shapes embedded in more complex geomet-rical patterns. However, a criticistn of this approach is that levels of field depend-ence are inferred from poor field independence performance (Ford & Chen, 2001).

The CSA differs from the GEFT in that it includes two subtests. The first presentsitems containing pairs of complex geometrical figures that the individual is re-quired to judge as either the same or different. The second presents items each com-prising a simple geometrical shape, such as a square or a triangle, and a complexgeometrical figure, as in the GEFT, and the individual is asked to indicate whetheror not the simple shape is contained in the complex one by pressing one of twomarked response keys (Riding & Grimley, 1999).

These two subtests seem to have different purposes. The first subtest is a task re-quiring field dependent capacity, whereas the second subtest requires thedisembedding capacity associated with field independence. In this way, the CSAovercomes the GEFT limitation that affects the measures of field dependence andfield independence, because field dependent competence is positively measuredrather than being inferred from poor field independent capability (Ford & Chen,2001). In addition, the CSA offers computerized administration and scoring. There-

FIGURE 1 Screen design ofthe Web instruction program.

Chen and Macredie

fore, the authors selected the CSA as the measurement instrument for field depend-ence in this study.

The CSA measures what Riding and Sadler-Smith (1992) referred to as aWholist/Analytic (WA) dimension, noting that this is equivalent to field depend-ence-independence. As Witkin et al. (1971) argued, a field independent individ-ual is capable of a more analytical cognitive function than a field dependent indi-vidual, who uses a more glohal approach. Riding's (1991) recommendations arethat scores below 1.03 denote field dependent individuals; scores of 1.36 andabove denote fieid independent individuals; students scoring between 1.03 and1.35 are classed as intermediate. In this study, categorizations were based onthese recommendations. Table 3 presents the overall range of the WA scores inthis study.

Exit questionnaire. A paper-based questionnaire was used to examine theresearch question "What are the effects of students' cognitive styles on their per-ceptions and attitudes toward learning within a Web-based instructional pro-gram?" This instrument was chosen because it has the potential to collect cognitiveand affective data quickly and easily (Kinshuk, 1996). Another advantage of ques-tionnaires is that the data may be both qualitative and quantitative, allowing themto play a part in both quantitative and qualitative studies (Su, 1991).

Several well-known questionnaires are available in the field of human-com-puter interaction, such as Questionnaire User Interaction and Satisfaction, devel-oped by the University of Maryland (1988), and Purdue Usability Testing Ques-tionnaire, developed by Purdue University (1997). However, this study examinedstudents' responses to a particular Web-based instructional program that pro-vided multiple navigation tools and description formats, and the questionnairewas developed to test specific hypotheses related to cognitive styles. Therefore,we decided to design a questionnaire specifically for this study, instead of usingan existing questionnaire. The questionnaire was divided into two parts. The firstcontained information regarding biographical data relating to the student and hisor her experience of using computers, the Internet, and HTML. The second,which was the main focus, consisted of four open-ended questions and 47 closedstatements to collect students' responses to the Web-based instructional program.It took students approximately 20 min to respond to all of the questions.

The open-ended questions were used to explore students' experiences of do-ing the tasks, their opinions about the strengths and weaknesses of theWeb-based instructional program, and the difficulties that they met. Students

Table 3: The Overall Range of Wholist/Analytic Scores in This Study

Cognitive Styles

Field independent (N = 23)Intermediate (N - 21)Field dependent (N = 17)Overall

Mean

1.561.150.821.21

SD

QM0010.130.32

Minimum

1.361.030.610.61

Maximum

1.851.351.001.85

Cognitive Modeling in Web-Based Instruction 385

were free to describe their experiences and opinions in their own words, with rel-evant space provided, supplying a useful source of qualitative information. Theclosed statements were designed to test the hypotheses described in Section 2.3by gathering a large amount of specific quantitative information about students'comprehension, preferences, and satisfaction or dissatisfaction with theWeb-based instructional program. Five sections were included: Section A, level ofunderstanding; Section B, content presentation; Section C, interaction styles; Sec-tion D, functionality and usability; and Section E, difficulties and problems. Table4 illustrates the relationships between the hypotheses testing and the question-naire design.

Each closed statement could be classed as either in favor or not in favor of theprogram. The number of "favored" statements was almost equal to the "not-fa-vored" statements (20 favored statements and 27 not-favored statements), in anattempt to reduce bias in the questionnaire. All statements used a 5-point Likertscale consisting of the response options strongly agree, agree, neutral, disagree, andstrongly disagree. Students were required to indicate agreement or disagreementwith each statement by placing a check mark at the response alternative thatmost closely reflected their opinion. Their perceptions and attitudes were mea-sured by the following:

1. Positive Perceptions: the total score for all favored statements on the exitquestiormaire with the same Likert scale score;

2. Negative Attitudes: the total score for all not-favored statements on tbe exitquestionnaire with the same Likert scale score.

Table 4: The Relationships Between Hypotheses Testing andQuestionnaire Design

Foe MS of Hypothesis Questionnaire Examples of Closed Statcmciits

HI: Non-linear interaction Section BSection E

H2: Independentlearning

H3: Navigation tools

H4: Content presentation

H5: Confidence inunderstanding

Section C

Section D

Section B

Section A

1 iike the fact that it allowed me to learn topics in any order.I was confused which options I wanted, because it

provided too many choices.i like the fact that this tutorial allowed me to work at my

own pace and direction.I would prefer to learn from human tutors than Irom this

tutoriai.The map in this tutorial gives a meaningful framework of

HTML.It is easy to find a route for a specific task with the index.The content of this tutorial is too superficial.I would like to have more detailed expianations.At the beginning, I was not sure how HTML worked, but

I have gained a clear understanding by using thisprogram.

After using this program, I can easily use my knowledgeto design home pages.

Note. H = Hypothesis.

386 Chen and Macredie

3.4. Procedures

The experiment was conducted using a Web-based instructional program accessedthrough Microsoft's Internet Explorer They applied the following procedures:

1. The CSA was used to classify students' cognitive styles as field independ-ent, intermediate, or field dependent.

2. Students were asked to interact with the instructional program on How toUse HTML,

3. Students were assigned a practical task, which involved constructing a Webpage using HTML.

4. Students were asked to reflect on their opinions of the Web-based instruc-tional program by completing a paper-based exit questionnaire.

3.5. Data Analyses

To investigate the students' views of the Web-based instructional program, the datacollected from the closed statements on the questionnaire were coded for quantita-tive analysis. The independent variable was the participants' cognitive style. Thedependent variable was the participants' responses given on the Likert scale {5 =strongly agree, 4 = agree, 3 ^ neutral, 2 = disagree, 1 ^ strongly disagree). TheKruskal-Wallis test—a nonparametric statistical test equivalent to the one-way be-tween-groups analysis of variance and suitable for comparing three or more inde-pendent groups of sampled data (Hatch & Lazaraton, 1991)—was applied to ana-lyze participants' responses to the closed statements. A significance level of p < .05was adopted for the study. Tables of frequency counts and percentages were pro-duced for the students' responses to each question {see Section 4).

Also analyzed were the qualitative data collected from the open-ended ques-tions contained in the exit questionnaire. The students' responses were dividedinto three cognitive style groups, with the responses of each group coded under thefollowing categories: (a) strengths of the Web-based instructional program, (b)weaknesses of the Web-based instructional program, (c) experience of doing thetasks, and (d) difficulties and problems met. Such qualitative approaches wereused to illuminate the phenomena identified in the quantitative data.

4. DISCUSSION OF THE RESULTS

We applied the data obtained from the exit questionnaires to identify students' per-ceptions and attitudes toward the Web-based instructional program. As describedin the Exit Questiormaire section, positive perceptions and negative attitudes wereseparately measured by the favored statements and the not-favored statements.The results indicated that field independent students had higher scores in the fa-vored statements. In contrast, field dependent students obtained higher scores inthe not-favored statements. In other words, field independent students signifi-

Cognitive Modeling in Web-Based Instruction 387

cantly displayed tnore positive perceptions toward the Web-based instructiotialprogram, whereas field dependent students showed more negative attitudes to theprogram (see Table 5).

In addition, the three cognitive style groups showed different perceptions andattitudes to the program features. Sections 4.1 to 4.5 present the quantitative resultsbased on the hypotheses described in Section 2.3. Section 4.1 begins by discussingstudents' responses to the key feature of Web-based instruction, nonlinear interac-tion; it then progresses to consider the second related issue, independent learning{Section 4.2), where it is suggested that not all learners can accept independentlearning. Section 4.3, which moves on to address another feature of Web-based in-struction, multiple navigation tools, indicates that different cognitive style groupsfavor different tools for navigation. This is followed by a discussion of how cogni-tive styles influence learners' requirements for content presentation (Section 4.4).Subsequently, Section 4.5 examines whether cognitive styles influence learners'confidence in relation to understanding the subject content. This article then goeson to discuss qualitative results. Section 4.6 discusses whether learning by doingcan enhance student learning in a Web-based instructional program, and finally theproblems and difficulties that different cognitive style groups met are discussed inSection 4.7.

4.1. Nonlinear interaction

Only a portion of the learners in this study favored nonlinear interaction, which is akey attribute of the Web-based instructional program. Field independent studentsappreciated the fact that this program allowed them to study topics in any order (seeTable 6). However, field dependent students felt confused over which options theyshould choose (see Table 7). This may be because field independent individuals usemore active approaches and are better at transferring concepts to new situations.Conversely, field dependent students are more comfortable in guided learning pro-cesses (C. Chou & Lin, 1997). These findings are consistent with Reiff's (1996) viewthat field independent individuals are self-structuring and use an internal frame of

Table 5: Students' Reactions and Their Cognitive Styles

Cognitive Style

Field independentMSD

IntermediateMSD

Field dependentMSD

Positive Perceptions

77.411812.9800

54.526310.6738

31.52636.6738

Negative Attitudes

36.73686.9269

57.722210.1911

72.722211.1911

Note. Kruskal-Wallis results: For positive perceptions, N - 61; x^ - 11-32; p •.003. For negative attitudes, N = 61; x^ = 13.990; p = .001.

Chen and Macredie

Table 6: Views on Learning Topics in Any Order

Response Selected

Strongly disagreeDisagreeNeutralAgreeStrongly agree

(like the fact tliat it allowed me to learn topics in any order.

Field Independent

n %

2 8.73 13.06 26.0

10 43.52 8.7

Intermediae

n

16680

/o

4.828.628.638.10.0

Field Dependent

n

18422

cy

5.847.123.511.811.8

Note. Significance: x^ (N - 61) = 8.54,;) - .014.

Table 7: Views on the Range of Options Provided

/ was confused which options I wanted, because it provided too many choices.

Response Selected

Strongly disagreeDisagreeNeutralAgreeStrongly agree

Field Independent

n %

2 8.710 43.56 26.15 21.70 0.0

Intermediate

n

08670

/o

0.038.128.633.3

0.0

Field Dependent

n

01394

%

0.05.9

17.653.023.5

Note. Significance: x^ {N = 61) = 15.33, p = .001.

reference to structure problems and organize information; field dependent individ-uals rely more on external frames of reference and operate best where structures andanalyses are already provided (Lyons-Lawrence, 1994). As the results in this sectionsuggest, cognitive styles have significant effects on learners' attitudes toward non-linear interaction. Therefore, Hypothesis 1 was supported.

4.2. Independent Learning

In terms of Independent learning, field independent students appreciated the factthat the Web-based instructional program allowed them to work at their own pace(see Table 8). As suggested by Ford, Wood, and Walsh (1994), field independentlearners tend to be more analytical, imposing their own structure on a situation,and to be relatively less passive in their behaviors. Conversely, field dependent stu-dents prefer to learn from human tutors rather than from a Web-based instructionalprogram (see Table 9), implying that they seem better at learning material via hu-man interaction. As indicated by Castaneda, Ramirez, and Herold (1972), field de-pendent learners have a greater social orientation than field independent learners

Cognitive Modeling in Web-Based instruction 389

Table 8: Views on Working at the Learner's Own Pace and Direction

/ like the fact that this tutorial allowed me to work at my own pace and direction.

Field Independent Intermediate Field Dependent

Response Selected

Strongly disagreeDisagreeNeutralAgreeStrongly agree

n

155

102

o/to

4.321.721.743.5

%J

n

09570

/o

0.042.923.833.3

0.0

n

18440

5.947.123.523.50.0

Note. Significance: x^ (N - 61) - 7.50, p - .023.

Table 9: Views on the Learner's Preference for Human Tutors

/ would prefer to learn from human tutors tlian from this tutorial

Field Independent Intermediate Field Defiendent

Response Selected

Strongly disagreeDisagreeNeutralAgreeStrongly agree

n

211640

%

8.747.826.117.40.0

n

010650

to

0.047.628.623.80 0

n

024

101

%

0.0

11.823.558.8

5.9

Note. Significance: x^ (N = 61) = 23.01, p = .005.

and are more ready to accept other people's opinions or those positively related tointerpersonal competencies. These findings also echo the view of Olstad, Juarez,Davenport, and Haury (1981), in that the construct of field dependence is associ-ated with certain personality characteristics that may have important instructionaland learning ramifications. These results suggest that learners with different cogni-tive styles showed different reactions to independent learning, so Hypothesis 2was supported.

4.3. Navigation Toois

The Web-based instructional program provided six types of navigation tools, in-cluding alphabetical index, back-forward buttons, a hierarchical map, hypertextlinks, a main menu, and section buttons. Students with different cognitive stylessignificantly favored three of these six tools (map, index, and hyperlinks). The fol-lowing sections describe the findings related to each tool.

Hierarchicai map. Field dependent students thought that the hierarchicalmap provided them with a meaningful framework (see Table 10). This is consistent

Chen and Macredie

Table 10: Views on the Map of the Tutorial

The map in this tutorial gives a meaningful frameivork of HTML.

Field Independent Intermediate Field Dt'pendent

Response Selected

Strongly disagreeDisagreeNeutralAgreeStrongly agree

n

0101300

0.043.456.5

0.0U.O

n

05

1240

%

0.023.857.119.00.0

n

011

114

%

OX5»5.9

64.723.5

Note. Significance: x^ (N - 61) = 6.04, p = 0.041.

with the findings of Ford and Chen's (2000) study. Field dependent students tendto use a global, spectator, and less analytic approach to learning (Witkin et al., 1977)and to require more structure and guidance compared with field independent stu-dents. Arguably, favoring the hierarchical map could be considered as reflecting agreater need for authoritative guidance.

Alphabetical index. Field independent students found it easy to select rele-vant information for the prachcal task using the alphabetical index (see Table 11).The alphabetical index provides learners with a means to locate particular informa-tion without going through a fixed sequence of information. A possible interpreta-tion of this finding is that field independent students are strong in perceptual andconceptual tasks, actively segmenting information into relevant parts and analyz-ing the interrelationships among those parts (Goodenough, 1976). This is in accordwith Witkin et al.'s (1977) findings that field independent individuals are more ableto engage in learning requiring independent and analytical thought.

Hypertext iinks. Intermediate students found itusefultodiscover the relation-ships between different topics via hypertext links (see Table 12). In addition to this

Table 11: Views on the Index of the Tutorial

/(IS easy to find a route for a specific task with the index

Field independent Intermediate Field Dependent

Response Selected

Strongly disagreeDisagreeNeutralAgreeStrongly agree

1!

02894

%

0.08.7

34.839.117.4

n

03

1422

%

0.014.366.79.59.5

ij

04

1120

/u

0.023.564.711.80.0

Note. Significance: x^ (N - 61) = 9.42, p - .009.

Cognitive Modeling in Web-Based Instruction 391

Table 12: Views on the Links of the Tutorial

/ tiked to use the links because they can help me to discover the relationships betzveen different techniques.

Response Selected

Strongly disagreeDisagreeNeutralAgreeStrongly agree

Note. Significance

Field Independent

n

02885

:X^(N = 6i;

%

0.08.7

34.834.821,7

1 = 7.12, p = . 0 2 8

Intermediate

n

033

132

%

0.014.314.361.99.5

Field Dependent

n %

0 0,02 11,8

11 64,74 23.50 0,0

quantitative data, the qualitative data from open-ended questions also indicatedthat 10 of 21 intermediate students thought that insufficient hypertext links were in-cluded in the program (see Section 4.7). That intermediate students favored using thehypertext links could arguably be seen as indicating high levels of engagement withthe subject content, in that the hyperlinks represent interest in "follow-up" informa-tion relevant to the particular subject content being read at the time (Chen & Ford,1998). The results presented in Section 4.3 suggest that different cognitive stylegroups favored using different types of navigation tools provided by the Web-basedinstructional program. Therefore, Hypothesis 3 was also supported.

4.4. Content Presentation

Results related to the content presentation of the program showed that field inde-pendent students thought the content was too superficial (see Table 13). They echoedthis feeling in their responses to a further question where they said that they wouldlike to have more detailed information (see Table 14). These results imply that theypreferred to focus their attention on detail, suggesting that field independent learn-ers preferred to takea serialist learning approach that concentrated primarily on pro-cedural details when processing information in a learning context (Pask, 1976,1979).

Table 13: Views on the Content of the Tutorial

Response Selected

Strongly disagreeDisagreeNeutralAgreeStrongly agree

The content of this tutorial is too superficial.

Field Independent

n

064

130

%

0.026.117.456.50,0

Intermediate

n

011550

%

0.042,423.823,8

0,0

Field Dependent

n

010322

%

0.058.817.611.811.8

Note. Significance: y} (N = 61) = 27.44, p = ,001.

Chen and Macredie

Table 14: Views on the Levels of Detailed Explanations

Response Selected

Strongly disagreeDisagreeNeutralAgreeStrongly agree

I would like to have more detailed explanations.

Field Independent

n

245

111

%

8.717.421.747.8

4,3

Intermediate

n

06690

%

0,028.628.642,90.0

Field Dependent

n

19430

%

5.952.923.517.60.0

Note. Significance: x^ (^ ^ 61) - 7.63, p ^ ,022.

Conversely, the field dependent students expressed opposite opinions in their an-swers to these questions, suggesting that they favored taking a holist approach thatconcentrates on building a conceptual overview (Wilson, 1998).

The results presented in Section 4.4 suggest that field independent and field de-pendent learners show different requirements related to content presentationwithin the Web-based instructional program, offering support for Hypothesis 4.

4.5. Confidence in Understanding

The aforementioned sections suggested that field independent and field depend-ent students showed different attitudes to the features of the Web-based instruc-tional program and had different preferences with respect to the navigation toolsand content development. However, they were similarly confident in understand-ing the subject content. After taking this program, 47.8% of field independent stu-dents and 52.9% of field dependent students felt that they had developed a clearunderstanding of HTML (see Table 15). In addition, 43.4% of field independent stu-dents and 47% of field dependent students thought that they had enough knowl-edge to design home pages (see Table 16). These results suggest that cogrutive style

Table 15: Views on Confidence to Understanding HTML

At the beginning, I was not sure how HTML worked, but I Imve gained a dear under<.tandin^ by using thisprogram.

Field Independent intermediate Field Dependent

Response Selected

Strongly disagreeDisagreeNeutralAgreeStrongly agree

Note. Significance x

n

05774

;2{N = 61)

%

0.021.730.430.417.4

- 5,16, P - . 0 7 6 ,

n

07509

u/

0.033323.80.0

42.9

n

04472

%

0.023.523.541.111.8

Cognitive Modeling in Web-Based Instruction 393

Table 16: Views on Confidence to Design the Homepages

After using this program, 1 can easily use my knowledge to design home pages.

Response Selected

Strongly disagreeDisagreeNeutralAgreeStrongly agree

Note. Significance; x

Field Independent

n

42773

^ {N - 61)

%

17.48.7

30.430.413.0

- 3.95, p - .139

Intermediate

n

47118

%

19.033.34.84.8

38-1

Field Dqjendent

n

43217

%

23.517.711.85.9

41.1

had no effect on students' confidence in their understanding of the subject contentprovided by the Web-based instructional program, and that Web-based instruc-tional programs have the potential to be an effective learning medium for differentcognitive style groups. Therefore, Hypothesis 5 was rejected.

4.6. Learning by Doing

In addition to the aforementioned interesting quantitative results, the qualitativedata reveal some unexpected findings. The data from the open-ended questions re-vealed that whether "learning by doing" could assist some students in setting ef-fective learning strategies was dependent on students' cognitive styles. A signifi-cant number of field independent students (n - 13) reported that doing the task wasa useful way of helping them to set a focus in the Web-based instructional program.Typical feedback included "Doing the task gives me a practical opportunity to ap-ply what I have leamt immediately," "If I did not do the tasks at the same time,probably 1 would be confused with how to start," and "Doing the task let me have agoal to learn." Those opinions implied that learning by doing could help studentsset effective learning strategies. This finding is in line with the view of Smith andParks (1997), in which tasks serve to simulate "goal directed" browsing in such away that learning performance can be enhanced. On the other hand, quite a few ofthe field dependent students (?j = 10) showed a more negative response to doing thepractical task because it was difficult for them to find relevant information. Typicalcomments included "I feel [it is] difficult to find relevant information for eachtask," "I am not sure how to include large amounts of data within one rule," and"The task sheet is not clear enough for me to find relevant information."

In addition, field independent students obtained better task scores than field de-pendent students (see Table 17). These differences echo the findings of Stanney andSalvendy's (1995) study in that field independent students should be able to con-struct a mental hierarchy of the system to guide their search. Conversely, field de-pendent individuals should tend to perceive information as is, without construct-ing an integrated model of the levels w ithin the embedded hierarchy. The lack of

394

Task Score

MSD

Table 17: Task Score

Field Independent

8.591.88

Chen and Macredie

Field Dependent

4.00L83

this mental model may lead field dependent students to have poorer search perfor-mance when compared with field independent students. As a result, it seems clearthat field independent students benefited more from doing practical tasks than didfield dependent students.

4.7. Difficulties and Problems

The other interesting qualitative results gained from the open-ended questionssuggested that three different cognitive style groups met different problems anddifficulties when they interacted with the Web-based instructional program. Anumber of field dependent students {n = 12) met more problems in the initial explo-ration; quite a few intermediate students {n - 10) felt that insufficient links wereprovided in the program, and several field independent students {n = 14) were con-cerned with how to find detailed information. Table 18 presents representativecomments.

Table 18: Students' Problems and Difficulties Associated with CognitiveStyles

Cognitive Style Problem Difficulty

Field dependent Problems to make an initial exploration:(N - 12) I needed to have a clear path through starting at the basics.

I would have liked more personal support/instruction in the beginning.It may be quite easy to get lost when starting this learning program.It is difficult to know where to start in this program.I really did not know how to order information at the beginning.

Intermediate Insufficient links provided in the program:{N = 10) I would have preferred to have links between examples and detailed description.

There are no links to different subsections.I cannot skip some pages that I am already familiar with.I could not compare different HTML techniques with hyperlinks.This structure made it difficult to jump to a particular page with hyperlinks.

Field independent Difficult to find detailed information:(N = 14) 1 would have preferred to have more explanations in the maps.

Index could be a bit simpler. Some information could not be found in the index,e.g., title, head, etc.

It would be useful to have a search facility to find particular information.Examples are too simple, and more complicated cases are needed.This program did not give in-depth analysis to compare different techniques,

e.g., the differences between Head and Heading.

Cognitive Modeling in Web-Based Instruction 395

5. DEVELOPMENT OF A DESiGN MODEL

As discussed in the previous section, cognitive style has a significant effect onlearners' perceptions of Web-based instructional programs. Field independent andfield dependent learners have different characteristics, which influence their pref-erences in relation to Web-based instructional programs. The findings have the po-tential to feed into the development of a design model for the implementation ofWeb-based instructional programs that can meet with the preferences of each indi-vidual. Figure 2 presents such a design model drawn from the preceding analysis,which illustrates how a Web-based instructional program is perceived by studentswith different cognitive styles.

The value of the design model lies within the ability of analyses of learners' re-quirements, which can help designers to develop an understanding of the prefer-ences of each cognitive style group and can provide an overall picture of learners'needs in taking Web-based instructional programs. With the proposed designmodel, the designers can recognize more easily what learners need and why and,hence, design Web-based instructional programs that more effectively addresslearners' unique needs and concerns. In analyses of their requiren:ients, the designmodel can especially be used to decide the approaches to instructional design andidentify the framework of system development. For instructional design, the char-acteristics of each cognitive style are important issues in the selection of teachingstrategies and the preparation of individual support. For system development, de-cisions about navigation support and content presentation provided by Web-basedinstructional programs should take into account the characteristics of learners' cog-nitive styles. The next section discusses detailed implications.

Field Independent

• Active Approach

• Indivitlualislic Style

• Serialist Patterns

• Internally Directed

Instructional Design

Teaching SCrdtegies

• Pracircal Task

• indefKtuknl Learning

Inilividual Support

• fcjitra Guidance

• Human Support

System Development

Navigntion Support

• Effective Shortciils

• Flexible Paths

Content Presenlalion

• Glnbiil Oriental ion

• Suggested Paihs

Field Dependent

Pa-SMVe Approach

Sixialisiit Style

HolisliL- Patterns

Ei^temally Directed

FIGURE 2 Design model of Web-based instruction.

Chen and Macredie

5.1. Instructional Design

Teaching strategies. The dimensions of field dependence and field inde-pendence reflect how well a learner is able to restructure information based on theuse of salient cues and the field arrangement (Welleretal., 1994). Field independentstudents tend to take an active approach and are able to extract the relevant cuesthat are necessary for completion of a task (Goodenough, 1976). Conversely, fielddependent students tend to adopt a passive, spectator role in learning and are moredominated by salient cues in learning (Anastasi, 1988). These differences may ex-plain the results of this study that field independent students enjoyed learningHTML by doing a practical task, but field dependent students appeared to havemore difficulties in searching for relevant information to help them undertake thetask.

Therefore, learning by doing can be a suitable teaching strategy for field inde-pendent students, because they tend to be focused and goal-oriented, and theirlearning motivation can be raised by practical exercises, such as factual recall, com-parative analyses, and hands-on demonstration. On the other hand, field depend-ent learners are "influenced by salient features" (Jonassen & Grabowski, 1993, p.88). It is still possible for them to learn the subject by doing a practical task, but ad-ditional guidance may be needed. For example, it would be useful for them totransfer knowledge into an activity by receiving instruction in task procedures us-ing down-to-earth visual material, such as graphic instruction or concrete exam-ples (Ford & Chen, 2000). Along with visual material, providing clear labels forpages would also be an appropriate approach to assist field dependent learners.Because field dependent learners tend to rely on vivid features, labels that clearlyindicate the role of a particular page may help them successfully decide the correctcoherent path (Lewis & Poison, 1990). An alternative is to provide field dependentstudents with a suggested route that can appear in a separate area or window(Hedberg, Harper, & Corrent-Agostinho, 1998).

Individual support. Web-based instructional programs provide studentswith an opportunity to work independently, requiring students to function in aself-directed way. The results of this study showed that field independent studentsconsidered independent learning a congruent and congenial approach becausethey tend to be more individualistic (Jonassen & Grabowski, 1993). Web-based in-structional programs are widely applied in distance learning, which requires stu-dent autonomy. Therefore, distance learning would be an appropriate educationalmode for field independent students (Richardson, 1998).

Conversely, the results showed that field dependent students preferred to learnfrom human tutors, instead of Web-based instructional programs. This is probablybecause field dependent students tend to have personalities with a greater socialorientation (Saracho, 1998). Video conferencing is becoming a powerful learningtool (Tuttle, 1997), allowing students and tutors to communicate in many wayssuch as through visuals, signs, spoken words, written text, and body language.Such remote communication might better support the special needs of field de-

Cognitive Modeiing in Web-Based Instruction 397

pendent students by providing additional content-related discussion withWeb-based applications. This may allow field dependent students to feel more con-fident in interaction with Web-based instructional programs.

However, Web-based instructional programs should provide another tool to al-low field independent learners to switch off such human support because they mayfeel disturbed. It is important to note that designers of Web-based instructionalprograms or other coniputer-based learning systems should consider providingoptions that will allow both field dependent and field independent learners to feelcomfortable in their interaction with instructional programs.

5.2. System Development

Navigation support. In this study, field dependent and field independentlearners showed different preferences with respect to navigation tools. A major is-sue in identifying differences between these two cognitive styles is their separatetendencies to adopt holistic and serialistic patterns. Field dependent learners typi-cally see the global picture, ignore the details, and approach a task more holisti-cally. Field independent learners tend to discern figures as being discrete from theirbackground, to focus on details, and to be more serialistic in their learning strate-gies (Ash, 1986). Therefore, cognitive style is an important factor for designers inidentifying the navigation support offered within Web-based instructional pro-grams. Field dependent learners prefer to build the structure of the context througha conceptual overview, so they may be able to get benefits from a global orientationthat helps them to understand the structure of the overall hyperspace and their ab-solute position I n the Web-based instructional programs (Schwarz, Brusilovsky, &Weber, 1996). This can be achieved by providing overview diagrams, fish-eyeviews, and hierarchical maps that show the whole picture of the context (Linard &Zeilger, 1995). In contrast, field independent learners prefer to focus on each indi-vidual element by understanding local detail. Thus, they may find it useful thatWeb-based instructional programs can be designed with effective shortcuts thatcan help them reach targeted information faster. An alphabetical index, keywordsearching, or other tools that can help them find specific information and allowmore effective engagement may support field independent students in learning ef-ficiently (Chen & Macredie, 2002).

Content presentation. Nonlinear presentation is another feature ofWeb-based instructional programs. The results of this study showed that students'cognitive styles had significant effects on their attitudes toward nonlinear learning.Field independent learners enjoyed determining their own paths, but field de-pendent learners felt confused about which options to choose. This may be attrib-utable to the fact that field independent learners operate more from internally de-fined goals; conversely, field dependent learners rely on externally provided cues(C. Chou & Lin 1997). Therefore, flexible paths are beneficial for field independentlearners who engage in learning with analytical approaches, and Web-based in-

Chen and Macredie

structional programs should provide them with multiple routes, free choice, andvisual control so that they can decide learning strategies by themselves.

On the other hand, suggested paths are helpful for field dependent learners, whoprefer to he guided in their learning process (C. Chou & Lin, 1997). Providing guidedroutes can help them gradually learn the hyperspace. For example. Web-based in-structional programs can provide them with direct guidance, which is an example ofadaptive navigation support. It can decide what is the next "best" page for the stu-dents to visit according to the their goal and other factors (Brusilovsky, 1998). To pro-vide direct guidance, the instructional program can outline visually the link to thebest page or present an additional dynamic link (called next or continue), which isconnected to the best page. Field dependent students seemed overwhelmed by thenumber of choices offered by a nonlinear structure, and direct guidance is a possiblesolution, allowing them to find their way through the hyperspace.

6. CONCLUSIONS

In response to the research question "What are the effects of students' cognitivestyles on their perceptions and attitudes toward learning within a Web-based in-structional program?" the answer seems to be that cognitive style plays an influen-tial role in students' reactions to a Web-based instructional program. As resultsfrom this study show, some learners may need greater support and guidance frominstructors, whereas others may be able to follow Web-based instructional pro-grams relatively independently. The results also suggest that students with differ-ent cognitive styles show different preferences with respect to the features ofWeb-based instructional programs, indicating the need for awareness of cognitivestyles when one is planning for Web-based instructional programs in educationalsettings. Thus, instructors should not assume that every student would benefitequally from Web-based instructional programs in educational settings. There re-mains the need for guidance to ensure that all students can meet the learning out-comes of such programs. Therefore, it is important to consider versatility in pro-gram design to allow for use by a variety of individuals, rather than a particularuser group. To achieve this aim, this study developed a support mechanism for thedesign of Web-based instructional programs in the form of a design model. Thismechanism can be used to support the development of Web-based instructionalprograms, with the final goal being to create programs that can be tailored to thepreferences associated with each cognitive style.

This study has shown the importance of understanding cognitive styles in thedevelopment of Web-based instructional programs, but it was only a small-scalestudy. Further studies have to be undertaken with a larger sample to provide addi-tional evidence. The other limitation is that this study adopted a self-developedquestionnaire, so the reliability and validity of the questionnaire are questionable.Therefore, testing and modification of the questionnaire are needed in the future.There is also a need to conduct further research to examine how cognitive styles in-fluence students' learning preferences in different types of tasks, such as the skillsof social interaction and critical thinking, because the nature of tasks lend them-selves to different techniques of instruction and a variety of learning approaches.

Cognitive Modeling in Web-Based Instruction 399

Such research should also be conducted within a more sophisticated multimediaWeb-based instructional program, including the presentation of animation andvideo. It would be interesting to see how different cognitive style groups perceivemultimedia interface features. The findings of such studies could be integrated tobuild robust user models for the development of personalized Web-based instruc-tional programs that can accommodate the preferences associated with differentcognitive styles.

REFERENCES

Anastasi, A. (1988). Psychological testing. New York: Macmillan.Andrew, M. (2001) Web-based strategies for improving undergraduate commitment to

learning. Proceedings of ED-MEDIA 2001 World Conference 2001, Tampere, Finland (June25-32), 53-58.

Ash, B. (1986). Identifying learning styles and matching strategies for teaching and learning. MA:(ERIC Document Reproduction Service No. ED270142)

Boyce, K. E. (1999). Delivering continuing professional education at a distance: The correlation offield dependence/independence and learning using the World Wide Web (field dependence, dis-tance education). Unpublished doctoral dissertation. University of Oklahoma.

Brusilovsky, P. (1998). Adaptive educational systems on the World-Wide-Web: A review of avail-able technologies. In Proceedings of workshop "WWW-Based Tutoring" at 4th International Con-ference on Intelligent Tutoring Systems (ITS '98), San Antoruo, TX. Retrieved August 1, 2004from http://www-aml.cs.umass.edu/-stem/webits/itsworkshop/brusilovsky.html

Carter, E.W. (2002). "Doing thebestyou can v/ith what you have": Lessons learned from out-comes assessment, journal of Academic Eibrarianship. 28, 36-41.

Castaneda, A., Ramirez, M., Ill, & Herold, P. (1972). Culturally democratic learning environ-ments: A cognitive styles approach. Riverside, CA: Systems and Evaluation in Education.

Chang, H. H., Henriquez, A , Honey, M , Light, D., Moeller, B., & Ross, N. (1998). The UnionCity story: Education reform and technology studetits' performance on standardized tests. NewYork: Center for Children and Technology.

Chapelle, C , & Jamieson, J. (1986). Computer-assisted language learning as a predictor ofsuccess in acquiring English as a second language. TESOL Quarterly, 20, 27-46.

Chen, S. Y, & Ford, N. I- (1998). Modelling user navigation behaviours in ahypermedia-based learning system: An individual differences approach, internationaljournal of Knowledge Organization, 25(3), 67-78.

Chen, S. Y, & Macredie, R. D. (2002). Cognitive styles and hypermedia navigation: Develop-ment of a learning model. Journal of the American Society for Information Science and Technol-ogy. 53(1), 3-\5.

Chou, C , & Lin, H. (1997, February 12-16). Navigation maps in a computer-networked hypertextlearning system. Paper presented at the Annual Meeting of the Association for EducationalCommunications and Technology, Albuquerque, NM.

Chou, H., & Wang, T. (2000). The influence of learning style and training method on self-effi-cacy and learning performance in WWW homepage design training. International Journalof Information Management, 20, 455-472.

Chuang, Y-R. (1999). Teaching in a multimedia computer environment: A study of effects oflearning style, gender, and math achievement. Retrieved from February 8, 2004http://imei.wfu.edu./arHcles/1999/l/10/

Davis, I. K. & Cochran, K. E. (1989). An information processing view of field dependence-in-dependence. Early Child Development and Care, 51, 31-47.

400 Chen and Macredie

Durfresne, A., & Turcotte, S. (1997). Cognitive style and its implications for navigation strat-egies. In B. Boulay & R. Mizoguchi (Eds.), Artifical inteiiigence in education knowledge andmedia learning system (pp. 287-293). Kobe, Japan: Amsterdam IOS Press.

Ellis, R. D,, & Kumiawan, S. H. (2000). Increasing the usability of online information forolder users: A case study in participatory design. International journal of Human-ComputerInteraction, 12, 263-276.

Even, M. J. (1982). Adapting cognitive style theory in practice. Lifelong Learning: The AdultYears, 5(5), 14-17, 27.

Federico, P. (2000). Learning styles and student attitudes toward various aspects of net-work-based instruction. Computers in Human Behavior, 16, 359-379.

Federico, P.-A. (1999). Hypermedia environments and adaptive instruction. Computers inHuman Behavior, 15, 653-692.

Ford, N., & Chen, S. Y. (2000). Individual differences, hypermedia navigation and learning:An empirical study, journal of Educational Multimedia and Hypermedia, 9(4), 281-312.

Ford, N., & Chen, S. Y. (2001). Matching/mismatching revisited: An empirical study oflearning and teaching styles. British journal of Educational Technology, 32, 5-22.

Ford, N., & Miller, D. (1996). Gender differences in Internet perceptions and use. AsUb Pro-ceedings, 48,183-192.

Ford, N., Wood, F., & Walsh, C. (1994). Cognitive styles and online searching. Online &CD-ROM Review, 18(2), 79-86.

Frank, B. M., & Keane, D. (1993). The effect of learner's field independence, cognitive strat-egy instruction, and inherent word-list organisation on free-recall memory and strategyuse. journal of Experimental Education, 62(1), 14-25.

Goodenough, D. (1976). The role of individual differences in field dependence as a factor inlearning and memory. Psychological Bulletin, 83, 675-694.

Hatch, E., & Lazaraton, A. (1991). The research manual: Design and statistics for applied linguis-tics. New York: Newbury House.

Hedberg, J., Harper, B., & Corrent-Agostinho, S. (1998). Creating a postgraduate virtual com-munity: Issues for authors and students as authors. Paper presented at Apple UniversityConsortium Academic Conference 98. Retrieved February 8, 2004 fromhttp://www.uow.edu.au/auc/Conf98/papers/harpberg.htmi

Jonassen, D. H., & Grabowski, B. (1993). individual differences and instruction. New York:Allyn & Bacon.

Ke, H., Kwakkelaarb, R.,Taic, Y., &Chenc, L. (2002). Exploring behavior of e-joumal users inscience and technology: Transaction log analysis of Elsevier's ScienceDirect OnSite in Tai-wan. Library & Information Science Research, 24(3), 265-291.

Kim, K. S. (2001). Implications of user characteristics in information seeking on the WorldWide Web. International journal of Human-Computer Interaction, 13, 323-340.

Kinshuk. (1996). Effectiveness of intelligent tutoring tools interfaces in relation to student, teaming topicand curriculum cfmracteristics. Unpublished doctoral dissertation. De Montfort University.

Large, A. (1996). H3q3ertext instructional programs and learner control: A research review..Education for Information, 4, 95- 106.

Lawless, K. A., & Kulikowich, J. M. (1998). Domain knowledge, interest and hypertext naviga-tion: A study of individual differences, journal of Educational Multimedia and Hypermedia,7(1), 51-69.

Lee, Y. B. (1992). Effects of learning styles in a hypermedia instructional system. In Proceedingof the Association for Educational Communications and Technology (pp. 505-509). Iowa: Edu-cation Resources Information Center.

Leong, S., & Al-Hawamdeh, S. (1999). Gender and learning attitudes in using Web-based sci-ence lessons. Information Research, 5(1). Retrieved February 8, 2004 fromhttp://www.shef.ac.uk/-is/publications/infres/paper66.html

Cognitive Modeling in Web-Based Instruction 401

Lewis, C , & Poison, P. G. (1990). Theory-based design for easily learned interfaces. Hu-man-Computer Interaction, 5,191-220.

Linard, M., & Zeilger, G. (1995). Designing navigational support for educational software. InB. Blumenthal et al. (Eds.), Human-computer interaction: Lecture notes in computer science1015 (pp. 63-78).

Lyons-Lawrence, C. L. (1994). Effects of learning styles on performance in using com-puter-based instruction in office systems. The Delta Pi Epsilon Journal, 36(3), 166-175.

MacArthur, D., & Lewis, M. (1996). Untangling the Web: Applications of the Internet and other in-formation technologies to higher education (Report No. DRU-1401-IET). Santa Monica, CA:RAND.

Marchionini, G. (1995). Information seeking in electronic environments. Cambridge, UK: Cam-bridge University Press.

Messick, S. (1976). Individuality in learning. San Francisco: Jossey-Bass.Miller, A. (1987). Cognitive styles: An integrated model. Educational Psychology, 7(4),

251-268.Ng, W., & Gunstone, R. (2002). Students' perceptions of the effectiveness of the World Wide

Web as a research and teaching tool in science learning. Research in Science Education,32(4), 489-510.

Olstad, R. G., Juarez, J. R., Davenport, L. J , & Haury, D. L. (1981). Inhibitors to achievement inscience and mathematics by ethnic minorities. (ERIC Document Reproduction Service No.ED223404).

Pask, G. (1976). Styles and strategies of learning. British Journal of Educational Psychology, 46,128-148.

Pask, G. (1979). Final report ofS.S.R.C. Research Programme HR 2708. Richmond, England: Sys-tem Research.

Pask, G. (1988). Learning strategies, teaching strategies and conceptual or learning style. InR. Schmeck (Ed.), Perspectives on individual differences, learning strategies and learning styles(pp. 83-100). New York: Plenum.

Power, D. J., & Roth, R. M. (1999). Issues in designing and using Web-based teaching cases.Proceedings of the 1999 Americas Conference on Information System, Milwaukee, Wl, 936-938.

Purdue University. (1997). Purdue Usability Testing Questionnaire. Retrieved August 1, 2004from http://www.acm.org/~perlman/question.cgi?form=PUTQ

Quintana, C , Krajdk, J., & Soloway, E. (2000). Exploring a structured definition forlearner-centered design. In B. Fishman & S. O'Conner-Divilbiss (Eds), Fourth InternationalConference of the Learning Sciences (pp. 256-263). Mahwah, NJ: Lawrence Erlbaum Associ-ates, Inc.

Quintana, Y. (1996, June). Evaluating the value and effectiveness of Internet-based learning. Paperpresented at the sixth annual conference of the Internet Society. Retrieved February 8,2004 from http://www.isoc.org/inet96/proceedings/cl/cl_4.htm

Reed, W. M., & Oughton, J. M. (1997). Computer experience and interval-based hypermedianavigation, journal of Research on Computing in Education, 30, 38-52.

Reiff. (1996). At-risk middle level or field dependent learners. Claring House, 69(4), 231-234.Richardson, J. T. E. (1998). Field independence in higher education and the case of distance

learning. International journal of Educational Research, 29, 241-250.Riding, R., & Rayner, S. G. (1998). Cognitive styles and learning strategies. London: Fulton.Riding, R. J. (1991). Cognitive styles analysis. Birmingham, UK: Learning and Training Tech-

nology.Riding, R. J., & Cheema, L (1991). Cognitive styles—An overview and integration. Educa-

tional Psychology, 11(3/4), 193-215.Riding, R. J., & Grimley, M. (1999). Cognitive style, gender, and learning from multimedia

materials in 11-year-old children. British journal of Education Technology, 30, 43-56.

402 Chen and Macredie

Riding R. J., & Sadler Smith, E. (1992). Type of instructional material, cognitive style andlearning performance. Educational Studies, 18, 323-340.

Rossman, G. B., & Wilson, A. (1985). Numbers and words: Combining quanitative andqualitiative methods in a single large scale evaluation study. Evaluation Review, 9,627-643.

Saracho, O. N. (1998). Research directions for cognitive style and education. Internationaljournal of Educational Research, 29, 287-290.

Scanlon, E. (2000). How gender influences learners working collaboratively with sciencesimulations. Learning and Instruction, 10, 463-481.

Schwarz, H., Brusilovsky, P., & Weber, G. (1996). World-wide intelligent textbooks. In P.Carlson & F. Makedon (Eds.), Proceedings ofEDTELEKOM 96—World Conference on Educa-tional Telecommunications (pp. 302-307). Charlottesville, VA: AACE.

Shih, C, &. Gamon, j . (1999). Student learning styles, motivation, learning strategies, and achieve-ment in \Ncb-basedcourses. Re\.nc\ed February 8,2004 fromhttp://iccel.wfu.edu/publica-tions/journals/jcel/jcel990305/ccshih htm

Smith, P. A., & Parks, L. M. (1997). Virtual hierarchies and virtual networks: Some lessons fromhypermedia usability research applied to the World Wide Web. Retrieved February 8,2004 fromhttp://ijhcs.open.ac.uk/smith/smith-nf.html

Soloway, E., Guzdial, M., & Hay, K. E. (1994). Learner-centered design: The challenge forHCI in the 21st century Interactions, 1 (2), 36-48.

Soloway, E., Jackson, S. L., Klein, J., Quintana, C, Reed, J., Spitulnik, J., et al. (1996). Learningtheory in practice: Case studies of learner-centered design. Proceedings of CHI '96, pp.189-196.

Soloway, E., & Pryor, A. (1996). The next generation in human-computer interaction. Com-munication of ACM, 39(4), 16-18.

Stanney, K. M., & Salvendy, G. (1995). Information visualization: Assisting low spatial indi-viduals with information access tasks through the use of visual mediators. Ergonomics. 38,1184-1198.

Su, L. T. (1991). An investigation to find appropriate measures for evaluating interactive informationretrieval. Unpublished doctoral dissertation. State University of New Jersey.

Tuttle, H. (1997). Video conferencing for learning. Retrieved January 4, 2003 from http://peo-ple.clarityconnect.com/webpages2/htuttle/videoconf.html

Umar, I. N. (1999). A study of the effects of cognitive styles and learning strategies among Malaysianpre-college students in a hypermedia environment. Unpublished doctoral dissertation. Uni-versity of Pittsburgh, Pittsburgh, PA.

University of Maryland. (1988). The Questionnaire for User Interaction Satisfaction. RetrievedAugust 1, 2004 from http://www.cs.umd.edu/hcil/quis

Weller, H. G., Repman, J., & Rooze, G. E. (1994). The relationship of learning, behavior, andcognitive styles in hypermedia-based instruction: Implications for design of HBI. Com-puters in the Schools, 10, 401^20.

Wilson, T. D. (1998). Exploring models of information behaviour: The Uncertainty Project. Paperpresented at Information Seeking in Context 1998, Sheffield, England.

Witkin, H. A., & Goodenough, D. R. (1981). Cognitive styles: Essence and origins: Field depend-ence and field independence. New York: International Universities Press.

Witkin, H. A., Moore, C. A., Goodenough, D. R., & Cox, R W. (1977). Field-dependent andfield independent cognitive styles and their educational implications. Review of Educa-tional Research. 47(1), 1-64.

Witkin, H. A., Oltman, P. K., Raskin, E., & Karp, S. A. (1971). A manual for the Group EmbeddedFigures Test. Palo Alto, CA: Consulting Psychologists Press.

Zoe, L. R., & DiMartino, D. (2000). Cultural diversity and end-user searching: An analysis bygender and language background. Research Strategies, 17(4), 291-305.