Ecologic Study of Children's Use of a Computer Nutrition Education Program

R E S E A R C H A RT I C L E ...............................................................................................

Ecologic Study of Children’s Use of a Computer Nutrition Education Program D O N N A M AT H E S O N 1 A N D C H E RY L A C H T E R B E R G 2 1

Stanford Center for Research in Disease Prevention, Stanford University, School of Medicine, Palo Alto, California 94304; 2 Department of Nutrition, The Pennsylvania State University, University Park, Pennsylvania 16802

INTRODUCTION

ABSTRACT The purpose of this research was to describe the context created by students as they worked in groups on a nutrition computer-assisted instruction (CAI) program. Students worked on the program in groups of three. Observational methods were used to collect data from students in two sixthgrade classrooms that were part of an experimental program designed to restructure the educational process. Thirty-two students, from 12 groups, were observed as they completed the program. The groups were assigned by the teachers according to standard principles of cooperative learning. Students completed “Ship to Shore,” a program designed specifically for this research. The program required three to five 50-minute classroom periods to complete.The objectives of the program were to change children’s knowledge structure of basic nutrition concepts and to increase children’s critical thinking skills related to nutrition concepts.We collected observational data focused on three domains: (1) student-computer interaction, (2) studentstudent interaction, and (3) students’ thinking and learning skills. Grounded theory methods were used to analyze the data. Specifically, the constant-comparative method was used to develop open coding categories, defined by properties and described by dimensions. The open coding categories were in turn used in axial coding to differentiate students’ learning styles. Five styles of student interaction were defined. These included (1) dominant directors (n = 6; 19%), (2) passive actors (n = 5; 16%), (3) action-oriented students (n = 7; 22%), (4) content-oriented students (n = 8; 25%), and (5) problem solvers (n = 5; 16%). The “student style” groups were somewhat gender specific.The dominant directors and passive actors were girls and the action-oriented and content-oriented students were boys. The problem solvers group was mixed gender. Children’s responses to computer-based nutrition education are highly variable. Based on the results of this research, nutrition educators may recommend that nutrition CAI programs be implemented in mixed gender groups.

Human behavior is thought to be a function of the interaction between a person and his/her environment.1 Accordingly, education researchers are interested in the interaction that occurs between students and their learning environment. In recent years, computers have radically changed the learning environment, yet there is little research to document these changes.2–4 Furthermore, in a content analysis of nutrition education research, only 8 of 80 studies measured environmental variables.5 Most research in nutrition education has determined if program participants achieved a prespecified task or outcome and assessed the accompanying personal attributes that predict the success of the participants.6 There is little research to guide nutritionists in creating environments that will optimize learning. Bronfenbrenner proposed a definition of environment that differed in scope and meaning from others in that people’s perceptions of their environment, not its objective reality, are what he believed really matters in development.7 The ecologic environment, according to Bronfenbrenner, consists of a series of interconnected systems that are embedded within each other.7 The most basic system is the microsystem, which refers to “the pattern of activities, roles, and interpersonal relations experienced by the developing person in a given setting with particular physical and material characteristics.”7 Three features that determine the significance of the microsystem are (1) the tasks or operations in which a person sees himself or others engaging, (2) the perceived interconnections between the people in the setting as members of a group engaged in common, complementary, or independent undertakings, and (3) the role that the individual has in the system.7 Therefore, environment, according to this definition, is not measured by sociodemographic characteristics but by interaction between individuals within a given setting. In computer-assisted instruction (CAI) research, interaction between individuals and the computer has been humanized in accordance with Vygotsky’s belief that learning is a direct result of social interaction.8 Based on the assumption that the computer is part of one’s social environment, chil-

(JNE 33:2–9, 2001) ................................................... This research was conducted at The Penn State Nutrition Center, University Park, PA. Address for correspondence: Donna Matheson, Ph.D., Stanford Center for Research in Disease Prevention, Stanford University School of Medicine, 1000 Welch Road, Palo Alto, CA 94305; Tel: (650) 498-4765; Fax: (650) 725-6906; E-mail: [email protected]. ©2001 SOCIETY FOR NUTRITION EDUCATION

2

Journal of Nutrition Education Volume 33 Number 1

dren’s affective and cognitive reactions to the computer software programs have been described.2,9–12 However, the characteristics of interaction between students and computers, such as the roles and responsibilities that students adopt, or the degree to which they personify the computer, have not been studied. More effective CAI nutrition interventions may be developed with a better understanding of how computers are used in an instructional setting.Therefore, the purpose of this research is to describe the interpersonal interaction, student-computer interactions, and students’ thinking processes that occur while children complete a nutrition CAI program in their classroom. To study interaction, it must first be defined and operationalized. Interaction between individuals and the computer is determined by the instructional approach of the CAI lessons,2,13 the learning approach of the program users,9,14 and the synergism created when individuals become engaged in a CAI program.15 For example, the level of challenge or competition in a CAI program may partially determine the cooperative or competitive nature of individuals’ responses to the program.16 Therefore, attributes of both program participants and the CAI program itself are required to fully understand children’s interaction with computers.To conceptualize and measure this interaction, three qualities of the interaction must be considered: (1) its temporal and dynamic nature, (2) its ecologic validity, and (3) its variability among individuals. First, interaction is defined as dynamic, as opposed to static, because it happens in the present. Each act is linked to the previous interaction,17 and these ongoing activities have goals or endpoints that determine the individual’s perceived purpose in the activity.7 Therefore, interaction cannot be assessed retrospectively or prospectively but must be measured as it occurs. Moreover, social interaction cannot be measured by unconnected evaluations of student characteristics and their learning environment that are statistically linked together. Rather, observational methods enable the researcher to simultaneously evaluate an individual and his or her environment. Second, interaction must be measured in an ecologically valid setting to ensure that participants’ interactions are genuine and uncontrolled. For example, in schools, hardware shortages typically require that students work together on computer games or programs.18 Therefore, research conducted in experimental settings, where children individually complete CAI programs, may not be representative of how a CAI program is used in practice. Instead, ecologically valid research describing the frequency and types of interaction that students engage in while working together on CAI programs2,9 and how groups of students responded to various CAI instructional modes (i.e., tutorials, simulations, or level of user control) provides a rich understanding of how computers change classroom dynamics.2,19 This research design allows the efficacy and effectiveness of a CAI program designed for classroom use to be adequately evaluated.

January • February 2001

3

Third, the reciprocity that characterizes the interaction between the person and his or her environment makes this relationship flexible, dynamic, and different from one individual to the next. Therefore, the differences and the similarities in program users’ learning strategies should be examined. Furthermore, not all development or learning is good. Dewey coined the term “miseducation” to describe educational programs that do not positively affect development.20 To fully understand the relationship between a student and a CAI environment, both the positive and negative outcomes must be evaluated on an individual basis.This approach is radically different from most education research where the goal is to generalize, across a population, the group gains in personal attributes (e.g., knowledge) that result from an intervention.Therefore, the purpose of this research is to describe the student-computer interaction that occurred while middle-school students completed a nutrition CAI program. In accordance with Bronfenbrenner’s definition of a microsystem,7 the activities, roles, and interpersonal relations that transpired as the students simultaneously interacted with each other and with the computer were observed and analyzed.

METHODS Participants and study site. This study was conducted in two sixth-grade classrooms in a middle school (population 672 students in grades 6 through 8) in a predominantly Caucasian, rural community. This research project was part of a larger study designed to restructure the education process through the use of technology and principles of cooperative learning.We conducted this research during the first year of implementation for the larger study; only two of the six sixth-grade students were involved in the project at this point. The classrooms were equipped with numerous computers that were used for individual lessons and for group work. Teachers had assigned students to work in groups of three according to standard principles of cooperative learning.21 Students of different abilities were grouped together, and friends were not in the same group.We worked within this classroom structure, and students completed the test program as a group project. Teachers were not involved in the implementation of the program.The CAI program “Ship to Shore” used nutrition as a theme to integrate skills from the core subject areas. The objectives of the program were to change children’s knowledge structure of basic nutrition concepts and to increase children’s critical thinking skills related to nutrition concepts. It consisted of six adventures that required three to five 50-minute class periods to complete.The variability in the length of time that students spent on the computer was in part due to (1) students’ reading, writing, and typing skills; (2) the degree to which they worked as individuals or as a team; and (3) how focused they were in conducting library research. Data collection occurred daily over a period of 5 months.Various instructional strategies including library research, fill in the blank questions,

4

Matheson and Achterberg/CHILDREN’S USE OF A COMPUTER NUTRITION EDUCATION PROGRAM

written assignments, and computer drawing were used in the program. An overview of the lessons in the program is provided in Table 1. Data collection methods. Observational methods were used to collect the data for this research.The computer used for the test program was in the back corner of one of the classrooms.As students worked together on the program, field notes of nonverbal behaviors were taken on a screen-byscreen basis. All sessions were audiorecorded and students’ words were subsequently transcribed into the field notes.An 89-page coding booklet with pictures of two screens from the computer program on the top of each page was used to collect the data. The form included demographics of the group, the session during which the students completed the activities, the time spent on the screens, and the type of interaction observed for each student.As the field notes were collected, they were categorized into three domains of interaction: (1) student-student interaction, (2) student-computer interaction, and (3) learning and thinking skills. In addition, semistructured, in-depth interviews that were conducted for another aspect of this evaluation were used to supplement the observational data.22 One observer did all of the observations and coding. On two occasions, a second observer accompanied the primary observer and collected data from the same children on the same day. The observation notes from both observers were compared for completeness and reliability. The data were analyzed using grounded theory methodology.17,23 The purpose of this analysis was to develop a theory to explain how students used the CAI nutrition program. The observational data were analyzed by comparing and contrasting each student’s behaviors, reactions, responses, and comments in each data collection domain (student-student interactions, student-computer interactions, and learning and thinking skills). The process of comparing data, between and within students, generated open coding categories that described the variation in the student responses Table 1.

Figure 1. sions.

Example coding of categories, properties, and dimen-

to the CAI program.These categories were defined by properties. Properties were specific qualities or actions. In turn, properties were described by dimensions. Dimensions were differences in manifestation, degree, or level of a property.An example of a category and the properties and dimensions that described and defined this category are illustrated in Figure 1. All open coding categories that emerged from this analysis were then applied to data from each student, and the patterns between these categories, within an individual, were examined. This process is called axial coding.17 A detailed description of the methods used in this study has been previously published.23

RESULTS Sample. Eight groups of three students and four groups of two students were included in this analysis (n = 32). In two of the pairs, the third member of the group refused to par-

“Ship to Shore” lessons, topic areas, and example activities.

Lesson Boarding the Ship

Stormy Seas

Nutrition Topic

Example Activities

Number of servings and serving sizes from

Calculate the food supplies and number of animals to

each of the food groups

bring on board the ship

Starvation

Propose solutions to starvation resulting from a loss of food due to a storm

The Stowaway

Vitamin D deficiency

Library research to discover why a stowaway has deformed bones

The Heat Wave

Dehydration, osmosis

Discovery of the New World

Vitamin C deficiency

Foods Refugees on the Raft

Simulation of osmosis experiment Library research to determine vitamin content of New World foods

Vegetarianism

Art activity promoting decreased meat consumption to fellow passengers

Journal of Nutrition Education Volume 33 Number 1

ticipate or did not return the consent form.The sample was 56% male. The majority of the groups were mixed gender, but three of the groups were all male and one group was a pair of girls.The groups were drawn from both sixth-grade classrooms; 15 students were from one class and 17 students were from the other class.The project was conducted in the second half of the school year, so the students were familiar with computers and comfortable working in groups. Description of categories. The grounded theory analysis produced categories or variables in each data collection domain that were fundamental in describing how students used a CAI program. Together, these categories were interrelated to describe the similarities and differences in students’ social, behavioral, and cognitive responses to the CAI program.The central theme, or core category, that emerged from this analysis was that students, working within a group, had different styles of using the nutrition program. Using grounded theory analysis, the “student styles” were described through the development of properties and dimensions.The five student styles were (1) dominant directors, (2) passive actors, (3) action-oriented students, (4) content-oriented students, and (5) problem solvers. The categories that were used to differentiate the student styles are defined in Table 2. The specific properties and dimensions of the five student styles are described below.

January • February 2001 Table 2.

5

Categories that emerged from the open coding of the three

interaction domains. Student-student interactions Control/passive—the degree to which students controlled the computer hardware or speed with which the group moved through the program. Supportive behaviors/conflict initiation—the degree to which students encouraged other group members or criticized and argued with others Cooperative/individual learning—the degree to which groups worked on the program together or to which they worked as individuals Authority/submissive—the degree to which students provided directions or guidance to other group members or followed directions from other group members Competitiveness—the degree to which students challenged each other in the program or designated winners or losers Student-computer interactions Affect toward computer program—the degree to which students immortalized or talked to the computer Fantasy—the degree to which students embellished or became involved in the fantasy in the program Technical interest—the degree to which students were interested in the computer hardware or how special effects on the computer were developed Difficulty of activity—the degree to which students had cognitive

Dominant directors. In this research, the group of dominant directors was comprised of six girls (19% of the sample).The student-student interaction domain predominated in the analysis of this group. Specifically, the controlling dimension in the category “control/passive” and the authority dimension in “authority/submissive” described their behavior.The focus of these students was on trying to manage the group and control the computer. Their primary means of control was speed, and one of the most common strategies they used was to read all of the screens in the program out loud as quickly as possible. An alternative strategy to control the computer was to provide all of the answers. Therefore, these students often guessed using random answers, just to maintain control of the keyboard or mouse. All of these girls became impatient with others who did not work as quickly as them or who questioned their authority. They cooperated with other group members by taking turns typing or suggesting that other members do certain tasks, but they insisted on making the final judgment on the answers for most questions. Furthermore, they often took credit for correct answers that they did not originate.They were very competitive and concerned about how they would be graded on the program. Traditional education often rewards speed and quantity over quality; therefore, their focus was on getting as much done as quickly as possible rather than on learning per se. Four of the six dominant directors were established leaders in the classroom. The learning strategies that they used made them very efficient and thorough workers, but they

or conceptual difficulties with the computer activities Interpretation of screens—students’ misinterpretation of graphics or special effects in the computer program Thinking and learning skills Creativity—the ability of students to propose novel or imaginative ideas to solve the problems posed in the CAI activities Flexibility—the ability of students to apply different learning strategies to the activities Evaluation of performance—the ability of students to evaluate their performance and plan alternative strategies to correct their mistakes Integrates viewpoints—the ability of students to listen to and incorporate ideas presented by other group members Proposes learning strategies—the ability of students to develop an overall plan or strategy to approach a problem or lesson

tended to lack creativity and flexibility. For example, most of the girls used the strategy of going immediately to the end of the lesson to determine the final questions so that they could focus on finding the correct answer as they worked through the lesson. In doing so, they missed details and did not integrate the new knowledge into the broader context of the lesson. Furthermore, four of six girls became outwardly frustrated with any activity that was not easy for them to understand or complete.These students preferred to learn by rote and were very good at it.

6

Matheson and Achterberg/CHILDREN’S USE OF A COMPUTER NUTRITION EDUCATION PROGRAM

Passive actors. The second group, called passive actors, was another group of girls (n = 5; 16%) who were opposite the dominant directors.These girls were submissive and did not strive to control or dominate the group or the computer. They were slow readers who had difficulty working as fast as the other students. Accordingly, they adopted one of two strategies to accommodate their inability to keep up with the rest of the group.These girls either sat back and did nothing or took responsibility for one task (e.g., typing or reading) and focused their attention on this aspect of the program. Regardless of their approach, they quickly became confused and uncertain about the content and storyline in the program.Therefore, they needed to check all of their actions and thoughts with the other members.They were vulnerable to making mistakes and sometimes deferred their turn to answer questions to another group member. If they were working with unsupportive students, they often criticized the answers of others so that they would not be alone in making mistakes. To appear involved in the program, two of these girls took the role of typist, and one read many of the screens. Despite their physical involvement, they did not generate many original ideas or try to answer questions on their own.They typed what the other group members dictated to them or asked others when to move forward in the program. In general, they wanted to complete the program with as little mental effort as possible. The most striking similarity among these girls was that they looked for easy answers and avoided creating original solutions. Whenever possible, they depended on other group members to do the work.They did not like the library research and would rather guess, using random and unrelated words from the program. Moreover, rather than generating new ideas, they would try variations in spelling and capitalization in hope that the computer would accept an answer. Most of the passive actors disliked writing journal entries and counted the number of pages the group wrote. They based completion of a journal assignment on writing an arbitrary number of pages. The passive actors enjoyed the fantasies and humor that other group members interjected, but this group of girls did not become actively or emotionally involved in the storyline unless they were able to personalize the content. For example, one became interested in the food spoilage activity because she had heard of someone with food poisoning.Two of the girls became interested in the journal entry that asked them to describe their favorite vegetable. However, there were few opportunities in the program for the students to relate the content to their personal circumstances; therefore, this finding could not be substantiated. Action-oriented students. The third group, which was 22% (n = 7) of the sample, was a group of boys who primarily wanted to be entertained. Most of their interaction was with the computer.They were fascinated with technology and the “special effects” (i.e., sounds, graphics, and animations) in the program. In general, these boys were labeled as poor class-

room performers and had limited attention spans for work that was abstract or repetitive. However, their attention could be held by action, especially violence.They approached the nutrition CAI program like an arcade game. For example, they pretended that their hands were a gun to shoot characters on the screen and randomly clicked all over the screen in search of buttons that would cause audiovisual effects. These boys personified the computer and scolded it for being “too slow” or “too picky.”As well, they imitated the bells and whistles that signaled correct and incorrect answers. But, it made no difference whether they were right or wrong; they just liked the sounds made by the computer. Like the dominant director girls, these boys wanted to control the mouse and tried to move quickly through the program, but their reasons for control were different. They simply wanted to be surprised by the “special effects” and wanted to skip through sections of the program that were not action oriented.They liked to be taken off guard by animations or to be “grossed out.”These boys were not motivated to learn from the program and were unconcerned whether they could answer as many questions as other group members. Like the passive actors, they tried to avoid library research by suggesting random answers and disliked writing in the journal. They contributed absurd ideas that were ignored by other members in the group or became bored and easily distracted as the other students completed the journal entries. Content-oriented students. The fourth student style group was comprised of eight boys (25%).The ability of these students varied, but their enthusiasm for the program was consistent. These students became emotionally involved in the story. They identified with characters in the program and adopted their role as apprentice on the ship. They talked directly to the characters in the program and expressed their personal opinions to them. Like the action-oriented students, the content-oriented students approached the program as a game, but their fantasy revolved around the plot and characters.Their focus was not on controlling the computer or getting the right answer; instead, they were motivated by the goal of saving or rescuing one of the characters. For these students, feedback on whether their answers were correct was insufficient. They wanted closure on the adventures and “happily-ever-after” endings that described what happened to the characters. Furthermore, through the in-depth interviews, it was evident that the knowledge these students acquired from the program was tightly woven into the details of the stories. If the plot of a story was not cohesive, they made inferences or rationalizations that often led them to develop inaccurate knowledge. As a group, these students liked writing in the journal where they elaborated on the details of the characters or the setting. Moreover, the ideas that they generated became a real part of the program for these students. They originated causes and solutions, which varied in degree of complexity or accuracy. Often, they were inefficient workers and got

Journal of Nutrition Education Volume 33 Number 1

January • February 2001

7

sidetracked by their vivid imaginations.When working with students who rushed through the program, they became frustrated because they could not analyze or comment on the details of the program. At times, content-oriented students complained that the others were moving too fast, but they did not try to take control of the computer.Two girls whose involvement in the fantasy was similar to the content-oriented students were classified as dominant directors because, unlike the content-oriented students, they took control of the program, at the expense of integrating and interpreting details.

Unclassified students. One boy remained unclassified after several reviews of the grounded theory coding sheets. He was in a group with two action-oriented students, who also were his friends. Typical of action-oriented students, these boys did not care whether they answered questions in the program. Rather, they were interested only in playing with the computer.To complete the program, this boy became a dominant director, controlling the keyboard and reading all of the screens. He was clearly uncomfortable in this role and often asked the other boys to be more serious about the program or to help him.

Problem solvers. The fifth group was comprised of five students (16%), three boys and two girls. Like the passive actors, these students did not control or dominate the program, nor did they become extremely involved in the fantasy of the storyline. Instead, they watched the computer and what the other students were doing, and they thought about the concepts and the conditions of the problems that they were required to solve. They interacted with either the computer or the other students only when they needed clarification of the problem or if they wanted to propose a solution. At times, they directed the actions of other group members, but they did not try to become physically involved in the program or to control the group. Usually, these students ignored other group members who tried random answers or who were slow readers. However, one of the students who worked with a dominant director stated that she would rather have worked alone because the dominant director questioned her answers and would not try them in the program. The primary mode of involvement for the problem solvers was intellectual, and they were engrossed in learning. These students were adept at defining the problem and had the previous experience and knowledge required to draw accurate and feasible conclusions. Compared to other students, their solutions and thinking strategies were more complex, logical, and thorough. In an anecdotal observation, Blissett and Atkins described a young boy whom we would have classified as a problem solver:9

Composition of student groups. The composition of the student styles within groups did not reveal any consistent patterns.That is, the presence of one student style did not cause another student to adapt a particular student style. For example, the presence of a dominant director did not require another student to adopt the role of a passive actor. Because the composition of groups was so diverse, it was difficult to assess the most effective combination of student styles. However, compared to groups that consisted of one type of student style, groups comprised of various student styles were more effective and efficient at completing the computer program than less diverse groups. Two action-oriented boys or two dominant director girls in the same group were unproductive teams because these students were focused on vying for control and not on learning. Content-oriented boys in the same group often spent an excessive amount of time on the program because, together, these boys developed extensive fantasies around the program. They needed a problem solver or dominant director to focus their effort on the content and challenges in the program.

He did not participate in his group’s discussions and did not seek control of the keyboard at any time. Rather, he sat a little back from the rest of the group with his arms folded, apparently disengaged from what they were doing. However, when he did make a contribution it was on the basis of deductive reasoning and was the only example of highly systematic, logical and correct thinking recorded.

In this research, one of the activities in the program involved an abstract simulation of an osmosis experiment. The content of the lesson was difficult for the sixth-grade children to understand, but they could easily guess at the answers to complete the lesson. The problem solvers distinguished themselves from other students because only they tried to understand osmosis rather than just guess at answers.

DISCUSSION The results of this implementation study indicated that children’s responses to computer-based nutrition education are highly variable, and their response to CAI may affect what they learn from the program.The grounded theory developed in this research provided a framework to explicate the differences and the similarities between children as they worked together on a CAI nutrition program. Actual behaviors and the variations, range, and consistency of these behaviors were the basis for the five student styles described in this research. Evaluations of learning styles have previously been used in the development of nutrition education programs for adults.24 A cancer risk reduction program used Kolb’s learning styles,25 which include (1) accommodators, (2) divergers, (3) convergers, and (4) assimilators. Program participants rated the program as informative, useful, relevant, and well organized. However, this classification of learning styles may be inappropriate for CAI and for young adolescents.The taxonomy was originally developed to advise college students on career paths and therefore focused only on higher order thinking skills.25 Specifically, it was based on two bipolar con-

8

Matheson and Achterberg/CHILDREN’S USE OF A COMPUTER NUTRITION EDUCATION PROGRAM

tinua: abstract or concrete learning and action or reflection. These constructs are somewhat similar to the thinking and learning skills domain that was used to describe the student styles in this research. Consequently, Kolb’s learning styles are similar, to some degree, to the student styles proposed in this research. For example, divergers, like the content-oriented students, were categorized as imaginative, and convergers, like the problem solvers, applied abstract knowledge to solve problems. However, neither student interaction nor contextual factors were considered in Kolb’s learning styles. By reducing learning to two continua, the internal validity of Kolb’s learning styles, especially for CAI, may be limited. Moreover, Kolb’s Learning Styles Inventory relies on 12 self-report statements that are often collected without any consideration for the learning task, the format of the instruction, the content, or the learning context. Hence, the ecologic validity of Kolb’s learning styles may also be questioned.25 The interaction between students and between students and computers in classroom environments has previously been evaluated. Lists of the frequency and types of interactions in which program users engage have been compiled,2,9,10,26 and interaction has been correlated with learning outcomes.26,27 However, interaction was not interpreted as part of an ongoing and reciprocal relationship between students and the computer. Consequently, the dynamic quality of the data was lost, and the meaning of the interaction was difficult to discern. For example, an observational study of 12- to 13-year-old children reported that random trial and error guessing and systematic trial and error guessing were the strategies commonly used to complete a math CAI program.9 Guessing was hypothesized to be used by students who preferred learning tasks that are clearly structured and who were unable to cope with open-ended activities. However, we observed the same behavior with both closed- and open-ended questions and documented that the primary reason students randomly guessed at answers was to maintain control of the computer. Thus, the process of analyzing action, interaction, and outcome sequences revealed important themes that interrelated and connected the computerstudent interaction to the student-student interaction. Similar to this study, researchers have previously reported gender differences in middle-school children’s responses to CAI simulations.27 An observational study of 60 9- to 10year-old boys and girls indicated that the cognitive performance of all male groups, who completed a CAI program on farm life in a third-world country, was significantly better than female groups.The observations indicated that the girls spent more time talking about the task and the boys spent more time using the computer. Likewise, in our research, both groups of boys, the content-oriented students and the action-oriented students, became very involved in the storyline or action in the CAI program, respectively. Girls, on the other hand, who were the dominant directors and passive actors, were more concerned with interaction between group members and were less engaged with the computer.

Limitations. The categorization of student styles proposed in this research was consistent with the findings from previously reported evaluations of children’s interaction with computer software, indicating that this research has theoretical validity. However, a number of limitations should be addressed in future research to substantiate the reliability of the results of this qualitative inquiry.Although the reliability of the observations was examined, the reliability of the grounded theory coding was not replicated. Furthermore, a small, ethnically homogeneous sample from a single school site participated in this research.The student styles categories should be replicated in larger samples, with diverse populations, different CAI programs, and new observers, and using alternative methods. For example, quantitative instruments may be constructed based on the descriptions of the student styles and used to establish concurrent validity. Likewise, the generalizability of the research must be confirmed by replicating the results in other contexts. In addition to schools, the consistency of student styles should be tested using CAI programs implemented in community centers and private homes. Moreover, the stability or reliability of student style classification should be established within individuals.That is, the same child should demonstrate the same student style characteristics while working with different children and using other nutrition CAI programs.

IMPLICATIONS FOR RESEARCH AND PRACTICE This study approaches the evaluation of student interaction from an ecologic perspective.This approach generated novel findings and numerous interesting research hypotheses. In addition to methodologic validation, the student style categories need further theoretical validation. In particular, the relationship between student styles, attitudes toward learning nutrition, and change in nutrition-related outcomes should be investigated. Student style may in part determine the information a program participant attends to, how new information is incorporated into their knowledge structures, and how students develop misconceptions. These research questions would require a large sample to statistically compare student style across gender or age and within experimental groups. The practical implications of the variability in students’ approach to using CAI must be recognized by nutrition researchers and educators. Nutrition CAI programs should be developed with consideration for the differences in students’ perceptions, goals, and reactions to CAI programs. For example, an element of fantasy may be critical to engage boys who are content-oriented students. Girls who are dominant directors may respond best to highly structured organized activities in which they may monitor their progress. Moreover, differences in student style should be acknowledged and recognized in program implementation. Based on the results of this research and other reports of student interaction using computer software, nutrition educators may recom-

Journal of Nutrition Education Volume 33 Number 1

mend that nutrition CAI programs be implemented in mixed gender groups. Like the group of action-oriented boys in this study, groups consisting only of dominant directors or passive actors may not be highly productive. Finally, program evaluation should address the implementation process, which is characterized by interaction. Nutrition educators need to become equally concerned with the differences and the similarities between students to design evaluations that reveal the true range of students’ reactions to CAI and the true potential for changes in nutrition knowledge, attitudes, and practices.

REFERENCES 1. Bandura,A. Social foundations of thought and action. Englewood Cliffs, NJ: Prentice Hall, 1986. 2. Anderson A,Tolmie A, McAteer E, Demissie A. Software style and interaction around the microcomputer. Comput Educ 1993;20:235–50. 3. Crook C. Computers in the zone of proximal development: implications for evaluation. Comput Educ 1991;17:81–91. 4. Edwards D, Mercer N. Common knowledge. London: Methune, 1987. 5. Achterberg C, Clark K. A retrospective examination of theory use in nutrition education. J Nutr Educ 1992;24:227–33. 6. Travers KD. Nutrition education for social change: critical perspective. J Nutr Educ 1997;29:57–62. 7. Bronfenbrenner U.The ecology of human development. Boston: Harvard University Press, 1979. 8. Vygotsky LS. Mind in society: the development of higher psychological processes. Cambridge, MA: MIT Press, 1978. 9. Blissett G, Atkins M. Are they thinking? Are they learning? A study of the use of interactive video. Comput Educ 1993;21:31–9. 10. McAteer E, Anderson A, Orr M, Demissie A,Woherem E. Computerassisted learning and group work: the design of an evaluation. Comput Educ 1991;17:41–47. 11. Simsek A,Tsai B.The impact of cooperative group composition on student performance and attitudes during interactive videodisc instruction. J Comput Based Instruc 1992;3:86–91. 12. Cordell B.A study of learning styles and computer-assisted instruction. Comput Educ 1991;17:175–83.

January • February 2001

9

13. Malone TW. Toward a theory of intrinsically motivating instruction. Cogn Sci 1981;4:333–69. 14. Bianchini JA. Where knowledge construction, equity and context intersect: student learning of science in small groups. J Res Sci Teach 1997; 34:1039–65. 15. Brown SJ, Lieberman DA, Gemeny BA, Fan YC,Wilson DM, Pasta DJ. Educational video game for juvenile diabetes: results of a controlled trial. Med Inform 1997;22:77–89. 16. Appleton K. Analysis and description of students’ learning during science class using a constructivist-based model. J Res Sci Teach 1997;34:303–18. 17. Strauss AL, Corbin J. Basics of qualitative research. Newbury Park, CA: Sage, 1990. 18. Becker HJ. How computers are used in schools: basic data from the 1989 I.E.A. computers in education survey. J Educ Computing Res 1991;7:385–406. 19. Crooks SM, Klein JD, Jones EEK, Dwyer H. Effects of cooperative

20. 21.

22.

23. 24.

25. 26. 27.

learning and learner-control modes in computer-based instruction. J Res Comput Educ 1996;29:109–23. Dewey J. Experience and education. New York: Macmillan, 1938. Johnson DW, Johnson R. Learning together and alone. Cooperative, competitive and individualistic learning. 4th Ed. Needham Heights, MA: Allyn and Bacon, 1994. Matheson D. Changes in context and learning: an evaluation of computer nutrition education program for middle school students. Doctoral dissertation, The Pennsylvania State University, University Park, PA, 1994. Matheson DM, Achterberg CL. A process evaluation model for computer assisted instruction programs. J Nutr Educ 1999;31:105–13. Beffa-Negrini P, Cohen NL. Use of learning style theory in development of a nutrition education program to reduce cancer risk. J Nutr Educ 1990;22:106A–B. Kolb D. Learning style inventory: technical manual. Boston: McBer and Company, 1976. Rowland P, Stuessy CL. Matching mode of CAI to cognitive style: an exploratory study. J Comput Math Sci Teach 1988;7(4):36–40. Kutnick R. Computer-based problem-solving: the effects of group composition and social skills on a cognitive joint action task. Educ Res 1997;39:135–47.