Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2005 The Effects of Seductive Augmentation and Agent Role on Learning Interest, Achievement, and Attitude Sanghoon Park

Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected]

THE FLORIDA STATE UNIVERSITY

COLLEGE OF EDUCATION

THE EFFECTS OF SEDUCTIVE AUGMENTATION AND AGENT ROLE

ON LEARNING INTEREST, ACHIEVEMENT, AND ATTITUDE

BY

SANGHOON PARK

A Dissertation submitted to the Department of and Learning Systems in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Degree Awarded: Fall Semester, 2005

Copyright © 2005 Sanghoon Park All Right Reserved

i The members of the committee approved the dissertation of Sanghoon Park defended on October 13, 2005.

______John Keller Professor Directing Dissertation

______Elizabeth Jakubowski Outside Committee Member

______Amy Baylor Committee Member

______Walt Wager Committee Member

Approved:

______Frances Prevatt, Chairperson, Department of Educational Psychology and Learning Systems

The Office of Graduate Studies has verified and approved the above named committee members.

ii

ACKNOWLEDGEMENT

My time at FSU has provided wonderful memories I will keep for my entire life. Especially, my work on this dissertation has been memorable - at times challenging but overall an exciting, instructive, and fun experience. There are many people who supported and taught me, to whom I would like to show my heartfelt appreciation. First of all, I would like to give a deepest thank to my advisor, Dr. John Keller, both for his inspiring and encouraging manner in guiding me to a deeper understanding of becoming a scholar and for his invaluable comments throughout the dissertation process. Whenever I had problems or concerns, he was always there to listen and provide valuable advice. He and his wife, Cecilia has been true mentors in my life. My sincere thanks go to one member of my committee, Dr. Amy Baylor, for reading all drafts of this dissertation and providing many valuable comments that improved its contents. She was always interested in my academic life and provided me with engaging research opportunities that became a foundation of my studies. I also would like to thank to one of my committee members, Dr. Walter Wager, for his significant contribution in clarifying my study. He offered suggestions for research sources and helped me refine my ideas and perspective. I am also very grateful to my outside committee member, Dr. Elizabeth Jakubowski, for supporting my study and providing valuable feedback during insightful discussions. My special thanks go to Professor Kenneth Baldauf in the computer science department, for allowing me to conduct this study in his classes and providing a good study atmosphere. A friendship with Ryan Wilke has led to many exciting discussions relating to this research. I appreciate his willingness to use his voice for creating pedagogical agents. I am also grateful to my colleagues from RITL (Center for Research of Innovative Technologies for Learning), E Shen, Chanhee Son, and Sue Ebbers, for helping me develop ideas about creating agents. Last but not least, I thanks to Jung Lim for being always there to support and encourage me. Finally, I send my endless love to my parents and sister who has always been with me.

iii

TABLE OF CONTENTS

List of Tables ………………………………………………….…..…………………..…….…... vi List of Figures.....…………………………………………….…………..……..………..……... vii Abstract …………………………………………………….…………………………….….…... x

CHAPTER 1. INTRODUCTION ...... 1

Context of the Problem...... 1 Problem Statement...... 6 Research Questions...... 6 Significance of Study...... 7

CHAPTER II. REVIEW OF RELEVANT LITERATURE ...... 8

Introduction ...... 8 Theoretical Background of Learning Interest...... 10 The Framework of Research on Interest...... 13 Seductive Augmentations ...... 16 Pedagogical Agent Supported Learning ...... 20 Graphics and Interest...... 23 Hypotheses...... 27

CHAPTER III. METHOD ...... 35

Introduction ...... 35 Participants ...... 35 Research Design ...... 36 Independent Variables...... 37 Dependent Variables...... 40 Treatment Group Description...... 45 Material...... 47 Procedure...... 49 Data Analysis...... 50

iv CHAPTER VI. RESULTS ...... 55

Introduction ...... 55 Preliminary Data Analysis...... 56 Tests for the Assumptions for Initially Planned Tests ...... 61 Examination of Hypotheses...... 64 The effects of seductive augmentation and agent role on learning interest ...... 65 The effects of seductive augmentation and agent role on achievement...... 71 The effects of seductive augmentation and agent role on attitude toward instructional material...... 74 Summary of the Hypotheses Tests...... 78

CHAPTER V. DISCUSSION ...... 80

Introduction ...... 80 Seductive graphics, learning interest, achievement, and attitude...... 81 Seductive messages, learning interest, achievement, and attitude...... 84 Pedagogical agents, learning interest, achievement, and attitude...... 87 Limitation of the Study...... 87 Implications ...... 89 Further Research...... 90 Conclusion...... 91

APPENDIX A. Instructional Material Text ...... 93 APPENDIX B. Pre-interest Instrument ...... 97 APPENDIX C. Post-interest Instrument...... 99 APPENDIX D. Recall Test ...... 102 APPENDIX E. Achievement Test Items...... 104 APPENDIX F. Attitude Instrument...... 106 APPENDIX G. Achievement Test Item Analysis...... 109 APPENDIX H. Agent Acripts...... 112 APPENDIX I. Instructional Material Screen (Companion Agent + Seductive Graphics) ...... 118 APPENDIX J. Instructional Material Screen (Instructor Agent + Seductive Graphics) ...... 124 APPENDIX K. Instructional Material Screen (Text Messages + Seductive Graphics) ...... 130 APPENDIX L. Instructional Material Screen (Base + Seductive Graphics)...... 137 APPENDIX M. Shapiro-Wilk Normality Tests for Learning Interest and Attitude ...... 143 APPENDIX N. Box Tests for Learning Interest and Attitude ...... 146 APPENDIX O. Levene’s Tests for Learning Interest and Attitude...... 148 APPENDIX P. Shapiro-Wilk Normality Tests for Recall Test and Comprehension Test...... 150 APPENDIX Q. Levene’s Tests for Recall Test and Comprehension Test ...... 152 APPENDIX R. Human Subject Committee Approval...... 154

REFERENCES ...... 157

BIOGRAPHICAL SKETCH ...... 165

v

LIST OF TABLES

Table 2.1 The types of augmentation and instructional materials...... 19 Table 2.2 The list of the acronyms of treatment groups...... 28 Table 3.1 Achievement test items ...... 43 Table 3.2 Treatment conditions...... 46 Table 3.3 The description of experimental conditions of the study...... 49 Table 4.1 Means and standard deviations for individual interest...... 57 Table 4.2 The list of dependent variables and single variables...... 58 Table 4.3 Correlation coefficients for relations among five measures of learning interest ...... 62 Table 4.4 Correlation coefficients for relations among three measures of attitude ...... 62 Table 4.5 Descriptive statistics for the dependent variables...... 64 Table 4.6 Means and standard deviations for learning interest...... 66 Table 4.7 Correlation of predictor variables with discriminant functions (Function structure matrix) and standardized discriminant function coefficients...... 68

Table 4.8 Multivate and Univariate Analysis of Variance F rations for learning interest...... 69 Table 4.9 Multivate and Univariate Analysis of Variance F rations for learning interest between agents ...... 70

Table 4.10 Means and standard deviations for achievement ...... 72 Table 4.11 Two-way analysis of variance for recall test score and comprehension test score ..... 72 Table 4.12 One-way analysis of variance for recall test score and comprehension test score...... 74 Table 4.13 Means and standard deviations for attitude...... 75 Table 4.14 Multivate and Univariate Analysis of Variance F rations for attitude...... 76 Table 4.15 Multivate and Univariate Analysis of Variance F rations for attitude between agents ...... 77

Table 4.16 Summary of hypotheses tests...... 78

vi

LIST OF FIGURES

Figure 2.1 Three approaches to interest research...... 14 Figure 2.2 Individual interest and Situational interest...... 16 Figure 2.3 The features of pedagogical agent affecting learner’s situational interest ...... 23 Figure 2.4 The features affecting learner’s situational interest in pedagogical agent supported multimedia learning ...... 27

Figure 3.1 Research design...... 36 Figure 3.2 The example of seductive graphic...... 38 Figure 3.3 The example of seductive message ...... 38 Figure 3.4 The example of a companion role of agent and an instructor role of agent ...... 39 Figure I.1 Sample screen showing a main page...... 119 Figure I.2 Sample screen showing a companion role of agent ...... 119 Figure I.3 Sample screen showing a purpose of instructional material ...... 119 Figure I.4 Sample screen showing a concept of intellectual property ...... 120 Figure I.5 Sample screen showing a three types of intellectual property ...... 120 Figure I.6 Sample screen showing a concept of a patent (1) ...... 120 Figure I.7 Sample screen showing a concept of a patent (2) ...... 121 Figure I.8 Sample screen showing a conditions to be a patent ...... 121 Figure I.9 Sample screen showing a concept of a trademark (1)...... 121 Figure I.10 Sample screen showing a concept of a trademark (2)...... 122 Figure I.11 Sample screen showing a concept of a copyright (1)...... 122 Figure I.12 Sample screen showing a concept of a copyright (2)...... 122 Figure I.13 Sample screen showing a concept of a copyright (3)...... 123 Figure I.14 Sample screen showing a closing of material ...... 123 Figure J.1 Sample screen showing a main page ...... 125 Figure J.2 Sample screen showing a companion role of agent...... 125

vii Figure J.3 Sample screen showing a purpose of instructional material...... 125 Figure J.4 Sample screen showing a concept of intellectual property...... 126 Figure J.5 Sample screen showing a three types of intellectual property...... 126 Figure J.6 Sample screen showing a concept of a patent (1)...... 126 Figure J.7 Sample screen showing a concept of a patent (2)...... 127 Figure J.8 Sample screen showing a conditions to be a patent...... 127 Figure J.9 Sample screen showing a concept of a trademark (1)...... 127 Figure J.10 Sample screen showing a concept of a trademark (2)...... 128 Figure J.11 Sample screen showing a concept of a copyright (1)...... 128 Figure J.12 Sample screen showing a concept of a copyright (2) ...... 128 Figure J.13 Sample screen showing a concept of a copyright (3) ...... 129 Figure J.14 Sample screen showing a closing of material...... 129 Figure K.1 Sample screen showing a main page ...... 131 Figure K.2 Sample screen showing a companion role of agent...... 131 Figure K.3 Sample screen showing a purpose of instructional material...... 131 Figure K.4 Sample screen showing a concept of intellectual property...... 132 Figure K.5 Sample screen showing a three types of intellectual property...... 132 Figure K.6 Sample screen showing a concept of a patent (1)...... 133 Figure K.7 Sample screen showing a concept of a patent (2)...... 133 Figure K.8 Sample screen showing a conditions to be a patent...... 134 Figure K.9 Sample screen showing a concept of a trademark (1) ...... 134 Figure K.10 Sample screen showing a concept of a trademark (2) ...... 135 Figure K.11 Sample screen showing a concept of a copyright (1) ...... 135 Figure K.12 Sample screen showing a concept of a copyright (2) ...... 135 Figure K.13 Sample screen showing a concept of a copyright (3) ...... 136 Figure K.14 Sample screen showing a closing of material...... 136 Figure L.1 Sample screen showing a main page...... 138 Figure L.2 Sample screen showing a companion role of agent ...... 138 Figure L.3 Sample screen showing a purpose of instructional material...... 138 Figure L.4 Sample screen showing a concept of intellectual property...... 139 Figure L.5 Sample screen showing a three types of intellectual property...... 139

viii Figure L.6 Sample screen showing a concept of a patent (1) ...... 139 Figure L.7 Sample screen showing a concept of a patent (2) ...... 140 Figure L.8 Sample screen showing a conditions to be a patent...... 140 Figure L.9 Sample screen showing a concept of a trademark (1)...... 140 Figure L.10 Sample screen showing a concept of a trademark (2)...... 141 Figure L.11 Sample screen showing a concept of a copyright (1)...... 141 Figure L.12 Sample screen showing a concept of a copyright (2)...... 141 Figure L.13 Sample screen showing a concept of a copyright (3)...... 142 Figure L.14 Sample screen showing a closing of material ...... 142

ix

ABSTRACT

Learning interest plays an important role in the learning process, determining what to learn, and how to learn the chosen information. Based on the distinction between individual interest and situational interest, studies on situational interest suggest that a high level of interestingness will lead to a high degree of attention and will foster the readiness of learners to get involved in the learning process, thus will increase the probability of successful learning. The source of interestingness is found in seductive details or seductive augmentations. Studies on seductive augmentation, that incorporated the use of multimedia, have mainly focused on instances in which students interact with a traditional computer screen interface, not with human- like agents. Several previous studies employed sound as a type of seductive augmentation, but verbal presentation of instructional messages from an agent have not been implemented yet. Consequently, the effects of seductive augmentation in an agent supported learning context where instructional messages are mediated by a pedagogical agent are largely unknown. As a result of this lack of seductive augmentation research in pedagogical agent supported learning, questions remain regarding how to design a pedagogical agent supported learning environment that uses seductive augmentation to promote learner’s interest and further, achievement. Therefore the purpose of this research was to investigate the effects of seductive augmentation and agent role on learning interest, achievement, and attitude in pedagogical agent supported learning. 127 college undergraduate students enrolled in “Computer literacy” classes at a large public university in the United States participated in this study. Students were randomly assigned to one of the eight treatment groups based on the sequence of their arrival to the research lab. The two independent variables were seductive graphics and the source of seductive messages. Seductive graphics contained two levels of variables: presence of seductive graphics and absence of seductive graphics. The source of seductive messages included four levels: companion agent, instructor agent, text without agent, and no message. Two agents (an instructor role of agent and

x a companion role of agent) were implemented to examine the agent effect. The three dependent variables were post interest, achievement, and attitude toward the learning material. A pilot test was conducted to determine the level of students’ prior knowledge about “intellectual property.” The topic of the instructional material used in this study was “Introduction to the intellectual property.” In this material, three main concepts of intellectual property; patents, trademarks, and copyrights were explained in detail. A two way between-groups MANOVA was performed to find the main effect of the two independent variables for two dependent variables, learning interest and attitude. In addition, a two-way between-groups ANOVA was conducted to examine the effects of seductive graphics and the source of seductive messages on the number of recalled keywords and comprehension test. In order to find the effect of specific agent role, the mean scores of the dependent variables for the companion agent role condition and instructor agent role condition were compared. The results indicated that the use of seductive graphics and seductive messages was effective to improve learner’s attention to the learning material in terms of learning interest, and also learner’s attitude, especially relevance scores. However, no significant difference was found for the recall test and the comprehension test. The two different agent roles did not make any differences on the three dependent variables. The primary contribution of this study is twofold. First, the results of this study illuminate the concept of learning interest as it concerns seductive augmentation. Especially, unlike the previous studies, the result of this study shows how seductive augmentation could affect three sub components of situational interest: attention, arousal, and involvement. Second, this study generalized the use of seductive augmentation in a multimedia learning context where a pedagogical agent is present. One of the strategies employed in this study to promote learner’s interest was seductive messages delivered by a pedagogical agent. It was found that students tend to believe and remember more when the message is delivered by agents. Further research is needed to examine the effects of seductive augmentation in different subject fields. Also the other aspect of situational interest, cognitive interest, needs to be studied since this study only implemented the concept of emotional interest. Further studies to verify the findings are also recommended.

xi

CHAPTER 1 INTRODUCTION

Context of the Problem

Interest increases learning. Promoting interest in the classroom increases learner’s learning motivation and improves the use of learning strategies. (Keller, 1983; Pressley, El- Dinary, Marks, Brown, & Stein, 1992; Sweet, Guthrie, & Ng, 1997). Interest also plays a significant role in the learning process, determining what to learn, and how to learn the chosen information (Alexander & Jetton, 1996; Garner, 1992). Therefore, the issue of how to increase learner’s interest is directly related to initiating learner’s motivation, and further improving learner’s achievement. Since Dewey’s (1913) theoretical discussion regarding learning interest in his book “Interest and effort in education”, there has been a notable upsurge in work on the concept of “interest” in learning contexts as indicated by some researchers (Alexander, Kulikowich, & Jetton, 1994; Hidi, 2001). However, the studies on interest declined noticeably in seventies because of a lack of consensus on the concept of interest. Later researchers began to study a renewed concept of “interest” as an explanatory construct in the field of learning and development (Krapp, Hidi, & Renninger, 1992). Generally, the concept of interest is defined in three categories which play a foundational role in contemporary discussions on motivation and interest (Krapp, 1999; Schiefele; 1992). First, interest is conceptualized as a dispositional characteristic of a person. According to this definition, learning interest is interpreted as a personality trait or motivational disposition a human-being possesses such as a long-lasting preference for a certain learning area or study topic.

1 Second, interest is considered as a characteristic emerging from a situation. In learning, this concept of interest is characterized as interestingness of a learning situation. The third concept of interest is referred to as a psychological state within a person, combining both the concepts of interest as an individual disposition and as interestingness of a learning situation. This last concept of interest is generally accepted and used in interest related research. Most of the research performed in past decades has defined interest as the third concept because it represents the actualized interest within a person affected by internal factors and external factors (Krapp, 1999; Renninger, Hidi, & Krapp, 1992; Schiefele, 1998; Schraw, Flowerday, & Lehman, 2001; Schraw & Lehman, 2001). In addition, the third concept of interest can be traced back either to an “interesting” factor of the context (situation) or to an already existing (dispositional) interest. Therefore interest, in this study, is defined as the third concept which is an actualized psychological state within a person. Given a focus on interest as a psychological state, both characteristics of the person and the learning situation are critical factors that affect the actualized learner’s interest. Research on individual interest is mainly concerned with a person’s characteristics, which is the subject side of the “person-object relationship”. On the other hand, research approaches engaged with the object side of this relationship is conceived as situational interest (Krapp, 1999). The distinction between individual interest and situational interest is useful for researchers because it helps them not only understand interest as a psychological state but also consider the distinction as a practical research framework to conduct empirical research in which interest is manipulated as a research variable. In fact, interest-related studies have been conducted from these two viewpoints of interest, individual interest and situational interest (Garner, Brown, Sanders, & Menke, 1992; Hidi, 2001; Hidi & Anderson, 1992; Krapp et al., 1992; Schiefele, 1991, 1992; Schraw, Flowerday, & Lehman, 2001; Stipek, 2002). From the view of individual interest, interest is implied as a characteristic of a person. It is specific to individuals, develops slowly, tends to be long lasting, and is triggered by an individual’s predisposition. Situational interest is generated as a result of the interestingness of a situation. It is caused primarily by situational factors such as certain conditions or any concrete objects in a specific situation. In this line of studies, it is assumed that a high level of interestingness will automatically lead to a high degree of attention and will also foster the readiness of learners to get involved in the learning process, thus finally will increase the probability of successful

2 learning. However, it has been reported that the interestingness of a learning context also shows unwanted “positive” effects referred to as seductive detail effects (Garner et al., 1992). Seductive detail refers to the types of interesting but irrelevant details that are added to a passage to make it more interesting (Harp & Mayer, 1997). Researchers who examined the effect of seductive details usually used attractive pictures or anecdotes to increase the interestingness of learning material. Garner, Gillingham, and White (1989) found that adding “unimportant but interesting” sentences to expository texts impeded the learning of the main points in the text. They labeled this finding the seductive detail effect, which is the term used to describe the effect of how added details seduce the readers away from the main idea in the text. However, the results of the “seductive detail effect” were not consistent across the research. Thalheimer (2004) summarized research on seductive details in his meta-analysis study and suggested that only those seductive details that divert the learner’s cognitive processing away from the information that should be learned hurt learning. The term “seductive detail” has been used extensively in the reading education field to refer to the interesting yet unimportant text segments added to expository texts in order to increase a learner’s situational interest. However, the concept of seductive details is not restricted to the reading educational field but has expanded to broader areas such as multimedia learning. Harp and Mayer (1997) added to the research by utilizing seductive illustrations, not just seductive phrases and sentences. This technology innovation sparked the use of a number of additional types of seductive details including context-relevant sounds, music, and video (Moreno & Mayer, 2000). Similarly, the base instructional material has been expanded beyond simple printed texts to multimedia animations and narrated audio (Mayer, Heiser, & Lonn, 2001). In order to refer to this combination of multiple seductive details in a multimedia learning environment the term “seductive augmentation” was generated as it presents a more inclusive statement of the multiple presentations of seductive details (Thalheimer, 2004). Similar to seductive details, seductive augmentation is defined as non-supporting, but vivid details embedded in and part of instructional material as a multiple forms of multimedia presentations (Thalheimer, 2004). It includes highly memorable details to the point that they disrupt learning or even the identification of important ideas. It is most seductive when they are novel, active, concrete, and personally interesting (Garner et al., 1992).

3 Regarding the effects of seductive augmentation, there have been two opposing research findings in line with the results on seductive details. One set of findings asserts that it energizes readers so that they pay more attention to learning and learn more overall, because it influences the learner’s affect by promoting his/her enjoyment of the topic. Hence, it causes the learner to pay more attention to and encode more of the information from the material (Izard & Ackerman, 2000; Kintsch, 1980). The importance of perceptual arousal to increase learner’s attention is suggested as one of the motivational design guidelines (Song, 1998). The other findings present a totally opposite position. They emphasize that it disrupts the learner’s construction of the cause- and-effect chain, so that adding seductive augmentation to the material will result in decreases on tests of retention and on solutions to transfer problems (Harp & Mayer, 1997; Wade & Adams, 1990). Advantages of using animated pedagogical agents in multimedia learning environments have been increased recently, as new technologies have made them more accessible (Craig & Gholson, 2002; Johnson, Rickel, & Lester, 2000). In a multimedia learning environment where a pedagogical agent is implemented, learning contents are delivered through the verbalized instructional messages spoken by the pedagogical agent as well as the multiple presentations on screen such as graphics, texts, sounds, and animations. Pedagogical agents are animated life-like characters designed to facilitate learning in computer-mediated learning environments (Johnson et al., 2000). Regarding the effects of pedagogical agents, researchers have indicated positive effects on learner’s attitude toward learning and performance (Baylor, 2002a, 2002b; Baylor & Ryu, 2003; Moreno, Mayer, Spires, & Lester, 2001). Also researchers have found positive effects of pedagogical agents on learner’s interest and motivation. Two explanations were suggested in regard to the positive effects of pedagogical agents on learner’s interest (Baylor & Ryu, 2003). The first reason was the agent persona represented by an agent’s image, animation, voice, and emotional expressions. Learners tend to like a pedagogical agent because they like the agent’s character or personality based on the agent’s persona. The second reason was the information provided by the agent. Learners who experienced the agent in a learning module reported that they liked the agent because they found the quality of messages delivered by the agent were informative and helpful. The research using agents with evasive emotion and no motivational messages also found that learners who experienced no motivational massages perceived no interest in learning content as well as an

4 agent used. From the post experiment interview, the learners mentioned that they would be much interested in both learning content and the agent presented to them if the agent was emotionally expressive and delivered motivational messages. (Warren, Shen, Park, & Balyor, 2004). Therefore, the factors affecting learner’s interest could be categorized into two dimensions, one is an agent’s persona and the other is the type of information delivered by agent. From the aspect of seductive augmentation, learners are exposed to learning content through two sources in a pedagogical agent supported learning environment. Learning material as information is presented on screen and messages are delivered by a pedagogical agent. In the learning environment, text information is not the only medium to deliver instruction. Together, sounds, graphics, and the agent’s persona represented by the agent’s image, animation, voice, and emotional expressions, make an integrated media used to deliver learning content to learners. Therefore, the three main factors that affect learner’s interest in a pedagogical agent supported learning environment are: instructional information presented on screen, instructional messages delivered by a pedagogical agent, and the agent’s persona represented by the agent’s image, animation, voice, and emotional expression. Agent persona, in addition, defines an agent’s role. Studies on seductive augmentation, incorporating the use of multimedia, have mainly focused on instances in which text and graphics are used to promote learner’s situational interest and in which students interact with a traditional computer screen interface, not with human-like agents. Several previous studies employed sound as a type of seductive augmentation (Mayer, Heiser, & Lonn, 2001; Moreno & Mayer, 2002; Moreno & Mayer, 2000), but verbal presentation of instructional messages from an agent have not been implemented yet. Consequently, the effects of seductive augmentation in agent supported learning contexts where communication is mediated by instructional messages from a pedagogical agent and instructional text information on screen is largely unknown. As a result of this lack of seductive augmentation research in pedagogical agent supported learning, questions remain regarding how to design a pedagogical agent supported learning environment considering seductive augmentation to promote learner’s interest and further, achievement. Specifically, how the instructional messages can be effectively designed for learning settings where students are given two instructional messages via two sources; one from an agent and the other from text-based instructional information. Many studies assert “seductive augmentation effect” in paper-based learning contexts involving texts and graphics as seductive

5 augmentation. The “seductive augmentation effect” was also found in a multimedia learning context where animations and sounds were incorporated (Garner et al., 1989; Harp & Mayer, 1997, 1998; Mayer et al., 2001; Moreno & Mayer, 2000, 2002). However, the effects of seductive augmentation in a pedagogical agent supported learning environment have not been proven. Furthermore, the effects of messages and the persona of pedagogical agents have not been studied based on interest theory. Therefore, this study attempts to examine the effects of these two factors, seductive augmentation and the role of a pedagogical agent, on learners’ interest, achievement, and attitude in a web-based pedagogical agent supported learning environment.

Problem Statement

The purpose of this research was to investigate the effects of seductive augmentation and agent role on learning interest, achievement, and attitude in web based pedagogical agent supported learning.

Research Questions

The main research question leading this study was what the effects of seductive augmentation and agent role are on learning interest, achievement, and attitude toward the instructional material. To further investigate this question, the following specific research questions were formed:

1. What are the effects of seductive augmentation and the role of agent on learning

interest?

2. What are the effects of seductive augmentation and the role of agent on achievement?

6 3. What are the effects of seductive augmentation and the role of agent on the attitude

toward learning material?

Significance of Study

Instructional designers usually design learning materials with an intention to ensure learners’ engagement with those materials. Earlier research and common practice have encouraged instructional designers to add clip art, graphics, sound, video, stories, and stimulating examples to their learning interventions to garner such engagement. Experiments in the late 1980’s and 1990’s on the “seductive details effect” called this practice into question when it was found that the addition of interesting yet unimportant augmentations can divert learners from learning the main ideas (Thalheimer, 2004). This study is significant in two ways. The primary contribution is that this study will further investigate the inconsistent findings concerning the effects of seductive augmentation. As described earlier, there have been two opposing research findings in regard to the effects of seductive augmentation. Therefore one purpose of this study is to further examine the inconsistency between facilitative effects of seductive augmentation and distracting effects of seductive augmentation in a multimedia learning environment. A second contribution of this study is to provide some implications for designing the most effective pedagogical agents. In pedagogical agent supported learning, the role of pedagogical agents is crucial because two issues need to be considered; designing instructional messages via agent voice and representing agent’s role via agent’s persona. Applying seductive augmentation principles to better design instructional messages corresponding to the role of an agent is the second significance of this study.

7

CHAPTER II REVIEW OF RELEVANT LITERATURE

Introduction

The importance of learning interest has been emphasized since Dewey (1913) first asserted the crucial role of interest in the learning process. In his book “Interest and effort in education”, he proposed that interest plays an elemental role in learning and that it is different from learner’s effort. He further argued that interest mediated the relationship between effort and learning, and concluded that “the appeal to sheer effort amounts to nothing” without interest (Krapp, 1999). Even though a formal model on learning interest was not presented, the efforts to theorize on learning interest increased dramatically with the advent of a new emphasis on the role of interest in learning (Schraw & Lehman, 2001). The conceptual framework of interest has been established based on the distinction between individual interest and situational interest (Renninger, Hidi, & Krapp, 1992). This framework became the foundation for most contemporary research on interest. Studies on interest have been conducted in two different ways. One has focused on promoting individual interest and the other on situational interest. Many strategies have been suggested and investigated to promote situational interest by modifying learning environmental factors that increase interest and by implementing seductive details in text learning (Schraw et al., 2001). Among those strategies, seductive augmentation, defined as multiple segments that are highly interesting, but unimportant to understand the overall text, has been studied as one of the effective strategies to improve situational interest in multimedia learning. Multimedia learning is designed using various media such as images, sounds, animations,

8 and text. Therefore controlling the proper amount of interestingness presented by various media is an important issue in designing an interest-promoting learning module. However, the results of previous studies are equally spilt as to the potential effect they have on text comprehension (Schraw et al., 2001). Wade et al. (1993) and Schraw (1998) reported facilitative effects for seductive details, whereas other researchers such as Garner at al. (1991; 1989), Harper and Mayer (1997; 1998) reported a negative effect. In this study, the effects of seductive augmentation will be investigated in pedagogical agent supported learning. Pedagogical agents have been highlighted as a useful technology to improve learning outcomes in multimedia learning to impart some type of information to users by interacting with them. It also is seen as extension of the development of a new learning interface (Craig & Gholson, 2002). The previous studies examined the effects of seductive details only in the form of information presented on the screen. This study, however, is designed to implement two independent seductive details from two information sources which are the information presented on screen and the messages from a pedagogical agent presented in the multimedia learning environment. In order to provide a rationale for this study, a theoretical framework is presented as five sections as follows. The first section serves to present the theoretical background of learning interest from the aspect of positive emotions and motivation. Also this section reviews the concept of interest comparing to the other comparable concepts such as curiosity and attention. The second section clarifies the research framework of this study in terms of individual interest and situational interest. How previous studies have been conducted to promote both aspects of interest is presented. The third section reviews the previous studies on seductive augmentation that have been suggested as a strategy to promote learner’s interest in terms of situational interest. The purpose of this section is to provide an overview of seductive augmentation related studies and how seductive augmentation can be applied to the pedagogical agent design in multimedia learning. The fourth section discusses the role of seductive augmentation in the pedagogical agent supported learning environment. Especially, how the learning interest can be promoted and how seductive augmentation can be implemented into pedagogical agent supported learning will be discussed.

9 The last section will serve to provide the background for using graphics in instructional material. Graphics are normally used as a type of seductive detail to promote learner’s situational interest. Therefore the use of graphics as seductive augmentation will be discussed as well.

Theoretical Background of Learning Interest

Interest and Emotion

When individuals are intrinsically motivated, they participate in activity because of learning interest and activity enjoyment (Eccles & Wigfield, 2002). Interest has been conceived as an important part in the learning process because it affects the use of specific learning strategies and the way of allocating one’s attention as well as promoting one’s emotional engagement in a task and the extent to which one engages in deeper processing (Hidi, 1990; Schiefele, 1996, 1999; Schraw, 1998; Wade et al., 1993). Thus, interest and enjoyment are considered typical positive emotions in a learning context. Positive emotions are usually considered “pleasant”, and the others are considered “unpleasant”, or negative (Gadanho & Hallam, 2001). Multiple research studies have provided evidence that positive emotions have a crucial effect on diverse cognitive processes such as information processing, the communication process, the negotiation process, the decision-making process, category sorting tasks, and even the creative problem solving process (Isen & Daubman, 1984; Isen, Daubman, & Nowicki, 1987; Isen, Johnson, Mertz, & Robinson, 1985; Isen & Means, 1983; Picard, 1997). Isen and Baron (1991) summarized, based on a program of empirical research, that “persons who are feeling happy are more cognitively flexible, more able to make associations, more able to see potential relations among stimuli than other persons in a neutral state”. Fredrickson (1998), in her Broaden-and-Build Model of positive emotions, identified four positive emotions; joy, interest, contentment, and love, and further suggested that positive emotions broaden the scope of attention, the scope of cognition, and the scope of action. The theory holds that, over time, the broadening triggered by positive emotions builds a range of personal resources, including physical resources (e.g., physical skills, health, longevity), social resources (e.g., friendships, social support networks), intellectual resources (e.g., expert knowledge, intellectual complexity), and psychological resources (e.g., resilience, optimism,

10 creativity) (Fredrickson, 1998). Forgas (1998) also found that people in a positive mood formulated action plans that were more cooperative and integrative, and achieved agreements of higher quality than did neutral or negative mood participants. As a type of positive emotion, the role of interest and its crucial role in learning were emphasized from the beginning of the 20th century. Dewey was the first who believed that interest differed from effort, and that interest led to deeper learning (Dewey, 1913). Since then, research and theorizing about interest rapidly increased until it reached its highest level in the early and mid-1990s. Many studies have shown the importance of interest in learning by conducting empirical studies. Especially, researchers who conducted a number of empirical studies on interest from 1980 to 1990 supported following three general conclusions (Schraw & Lehman, 2001). First, interest is related positively to attention and learning. Second, interest level varies from person to person. And lastly, interest is elicited by both internal factors and external factors. Renninger et al (1992) published the book “The role of Interest in Learning and Development”, which provided an integrated theoretical framework based on the distinction between individual interest and situational interest. This framework has been used as a theoretical ground for studies on learning interest. The most commonly accepted assumption regarding interest is that interest is a phenomenon that emerges from an individual’s interaction with his or her environment (Garner et al., 1992). Based on this assumption, psychologists considered the characteristics of interest as follows (Izard & Ackerman, 2000): First, interest motivates exploration and learning, and guarantees the person's engagement in the environment. Second, interest is a type of emotion that can sustain long-term constructivist or creative endeavors. These two significant features of interest imply that interest and motivation have a very strong relationship.

Interest and Motivation

In order to understand the concept of interest and its relationship with learning motivation, it is necessary to compare interest with another comparable motivational construct such as curiosity. Interest and curiosity are psychological constructs that are often used interchangeably without clear distinctions. Although interest shares some common features with curiosity, there are a couple of differences in terms of scope.

11 From the aspect of motivation, interest is considered one of the main categories of motivational conditions. In Keller’s motivational model developed to understand and predict human behavior in learning contexts, interest encompasses several theories of curiosity and arousal (Keller, 1983). Therefore the concept of curiosity is understood as a sub construct of interest. Furthermore, when interest is conceptualized as individual interest and situational interest, curiosity is mainly related to the situational interest. Unlike individual interest, which is interpreted as a personality trait or motivational disposition such as a long-lasting preference for a certain topic, situational interest is referred as the interestingness of a learning situation and evolved by the external factors in the learning condition. According to Kintsch (1980), interest occurs in two different ways in terms of situational interest; emotional interest and cognitive interest. Cognitive interest adjuncts such as explanative summaries, influence learner’s cognition by promoting the reader’s structural understanding of text. On the other hand, emotional interest is explained by the addition of interesting but peripherally relevant material to a textbook lesson, with the intention of energizing learners so that they pay more attention and learn more overall (Harp & Mayer, 1997). Therefore, both cognitive interest and emotional interest can be counted as aspects of situational interest. Curiosity is conceptualized into perceptual curiosity and epistemic curiosity. According to Berlyne’s distinction (1965), perceptual curiosity relates to attention. It refers to a sensory- level reaction and selective attention in response to particular objects in the environment. Epistemic curiosity refers to information seeking and problem-solving behavior that occurs as a result of the stimulation of curiosity (Keller, 1983). Given the difference between perceptual curiosity and epistemic curiosity, both are still aroused by some characteristics of the learning situation. Perceptual curiosity is explained as a selective attention, which is derived from stimulus of objects in a specific situation. Epistemic curiosity occurs as a result of the stimulation of curiosity, and well observed cases include when a child works at a jigsaw puzzle in a science problem solving process (Keller, 1983). Therefore the distinction between perceptual curiosity and epistemic curiosity is also similar to the distinction between emotional interest and cognitive interest.

12 The Framework of Research on Interest

The distinction between individual interest and situational interest has been drawn as a framework in many empirical studies on interest (Garner et al., 1989; Garner et al., 1992; Harp & Mayer, 1997, 1998; Hidi, 1990, 2001; Hidi & Anderson, 1992; Hidi & Baird, 1986; Krapp, 1999, 2002; Krapp et al., 1992; Schiefele, 1991, 1992, 1996, 1998, 1999; Schiefele & Krapp, 1996; Schraw, 1998; Schraw et al., 2001; Schraw & Lehman, 2001; Wade et al., 1993). From the view of individual interest, interest is implied as a characteristic of person. It is specific to individuals, developed slowly, tends to be long lasting, and is triggered by an individual’s predisposition (Renninger et al., 1992; Schiefele, 1998; Schraw & Lehman, 2001; Silvia, 2001). For example, learners who are interested in a topic or an activity pay more attention and acquire more knowledge than participants without such interest. Individual interest can be assessed through a learner analysis process by asking several background questions or administering a simple questionnaire. However, it is not easy to consider individual interest in designing learning material because individual interest refers to a student’s relatively enduring preference for different topics, tasks, or contexts (Krapp, 1999; Tobias, 1994). Krapp (1992) insisted that situational interest be a foundation of individual interest. According to him, when the content of the learning material is not a subject area in which the learner has established individual interest, the interesting factors in the subject learning situation is necessary to awake the interest for a short or longer period of time. This interest is defined as situational interest and the central psychological process “Internalization” supports the transformation process of situational interest into long-lasting individual interest as described in Figure 2.1. Situational interest is generated as a result of interestingness of situation. It is caused primarily by certain conditions and concrete objects in the environment, triggered by environmental factors, elicited by certain aspects of a situation, and it is assumed to contribute to the interestingness of the situation (Harp & Mayer, 1997; Hidi & Anderson, 1992; Krapp, 1999, 2002; Renninger et al., 1992; Schiefele, 1992, 1996; Schraw & Lehman, 2001; Silvia, 2001). While individual interest is a relatively stable evaluative orientation towards certain domains, situational interest is formed if an emotional state aroused by specific features of an activity or a task. The following Figure 2.1 illustrates the relationship between individual interest and

13 situational interest. This study grounded its framework of inquiry in this relationship.

Individual interest as a disposition (Characteristics of the person)

Actualized individual interest & situational interest Internalization (Psychological state within the person)

Interestingness (Characteristics of the learning context)

Figure 2.1 Three approaches to interest research (Krapp et al., 1992)

As illustrated in Figure 2.1, the relationship between individual interest and situational interest is not obviously defined. Both individual interest and situational interest are combined together in certain learning circumstances to represent the interest as a psychological state within a person.

Individual Interest

Individual interest is topic specific and persists over time (Hidi, 1990; Krapp et al., 1992). Individual interest is defined as people’s relatively enduring preferences for different topics, tasks, or contexts (Krapp, 1999; Tobias, 1994). Schiefele (1991, 1999) further distinguished between two subcomponents of individual interest; a feeling-related component and a value-related component as illustrated in Figure 2.2. Feeling-related valences refer to the feelings that are associated with an object or an activity, for instance, involvement, stimulation, and flow. It occurs when an individual experiences positive affect and emotions in conjunction with a particular topic or activity. It is presumed that these positive feelings provide a strong motivational incentive to engage in the activity (Schraw & Lehman, 2001). Value related valences refer to the attribution of personal

14 significance of importance to an object and activity. It is presumed that value related interest increases engagement because an activity or body of knowledge is judged to be salient to one’s goal (Sansone, Weir, Harpster, & Morgan, 1992; Schraw & Lehman, 2001). Since those two aspects are highly correlated each other, it is hard to determine what aspect of components affect individual interest (Schiefele, 1991, 1999). However, it is obvious that some learners prefer to learn specific topics or participate in specific activities because they like it primarily based on feeling, and others also do because they think the activity is important for some future purpose.

Situational Interest

Situational interest is elicited by aspects of a situation, such as novelty or intensity, and by the presence of a variety of human interest factors contributing to the attractiveness of different types of contents (Krapp, 1999; Tobias, 1994). Most of the research on situational interest has focused on the characteristics of academic tasks that create interest (Hidi & Baird, 1986). Features that have been found to arouse situational interest and promote text comprehension and recall are: personal relevance, novelty, activity level, and comprehensibility. Situational interest is often considered to precede and facilitate the development of individual interest (Krapp et al., 1992). Most interest research has focused on various aspects of situational interest. As Schraw and Lehman (2001) indicated, one of the reasons is that situational interest is more amendable to change compared to individual interest. They further categorized three general aspects of situational interest as text-based, task-based, and knowledge based. Text-based interest refers to the feature of the text information that affects interest. Task-based interest involves task manipulations or encoding instructions that increase interest. Finally, knowledge-based interest refers to the aspects of learner’s knowledge-base that promote interest. Seductive detail, which refers to the highly interesting but unimportant text segments, is considered one of the factors in text-based aspects of situational interest. Regarding situational interest, the concept of cognitive interest and emotional interest was proposed by Kintsch (1980). Cognitive interest adjuncts such as explanative summaries, influence learner’s cognition by promoting the reader’s structural understanding of the explanation. On the other hand, emotional interest is explained by that the addition of interesting but peripherally relevant material to a textbook lesson that energizes the learner so that they pay

15 more attention and learn more overall (Harp & Mayer, 1997). According to the definitions of cognitive interest and emotional interest, these are interest state derived by specific feature of a learning material. Therefore, both of cognitive interest and emotional interest need to be counted as two aspects of situational interest as shown in Figure 2.2.

Feeling related component Individual interest as a disposition (Characteristics of the person)

Value related component

Cognitive interest Interestingness (Characteristics of the learning context)

Emotional interest

Figure 2.2 Individual interest and Situational interest

Seductive Augmentations

According to Gagne (1988), gaining attention of the learner is suggested to be the “first event” of instruction. Many methods to engage learners by promoting interest have been suggested such as surprise-ending stories, interesting sentences in text, and seductive details (Harp & Mayer, 1997). As a matter of fact, many methods to engage learners in learning have been used, including telling stories, using case studies, walking around the classroom, using flipcharts, prompting group discussions, and asking questions. Similarly, instructional designers have augmented text-based learning materials by adding interesting elements such as stimulating stories, biographical details, clip art, photographs, sounds, and video (Talhaemer, 2004).

16 Seductive augmentation uses interesting segments that contain highly interesting, but unimportant information to understand the main topic. It has been suggested as a strategy to grasp learner’s attention and promote interest (Schraw & Lehman, 2001). The term “seductive augmentation” is derived originally from “seductive detail” which has been a focus of studies on reading and interest (Renninger et al., 1992). Studies have used the term seductive detail to refer to interesting but unimportant details that are added to a passage to make it more interesting. However, seductive detail is not irrelevant but peripherally relevant to main topic. This term has been used extensively in reading education to refer to the interesting yet unimportant text segments added to expository texts in order to increase learner’s interest. Recently, the original focus of the research has evolved into broader learning contexts. In multimedia learning, seductive detail usually is called seductive augmentation because the term “seductive augmentation” refers to not only text but also graphics, narratives, voice, animation, and text accompanied in multimedia learning environments with the purpose of increasing learner’s situational interest (Thalheimer, 2004). Harp and Mayer (1997) added to the research by utilizing seductive illustrations, not just seductive texts. This series of research inspired the use of a number of additional types of seductive details, including context-relevant sounds, narrative voice, music, and video. Similarly, the format of instructional material has been expanded beyond simple printed texts to multimedia animations and narrated audio. Seductive augmentations are non-supporting, but vivid details embedded in and part of instructional material. It includes highly memorable details to the point that they disrupt learning or even the identification of important ideas. Regarding the usefulness of seductive augmentation, there have been two opposite research suggestions (Harp & Mayer, 1997; 1998). The first suggestion asserts that it energizes readers so that they pay more attention to learning and learn more overall, because it influences the learner’s affect by promoting his/her enjoyment of the topic. Hence, it causes the learner to pay more attention to and encode more of the information from material. The importance of perceptual arousal to increase learner’s attention is also suggested as a motivational design guideline (Song, 1998). The second suggestion takes a totally opposite position. They emphasize that it disrupts the learner’s construction of the cause-and- effect chain so that adding seductive augmentation to the material will result in decreases on tests of retention and on solutions to transfer problems. This is called the “seductive augmentation

17 effect”. Seductive detail effect is a term used to describe the cognitive consequences of adding seductive details to text. This effect is typified by proficient recall of seductive details within a passage but poor recall of structurally important ideas (Garner, et al. 1989). Summarizing previous studies in regard to the effects of seductive augmentation in learning material, Thalheimer (2004) found that the seductive augmentation effect was found to be different depending on the material learners were involved with and the type of augmentation utilized. In other words, the types of material content such as expository information or bibliography information, and the types of seductive augmentation such as text, picture, illustration, sound, voice, animation accompanied with narratives showed sometimes seductive augmentation effect and sometimes not. He further suggested that more research is needed to clarify the situations in which adding interesting instructional items hurts learning or helps learning. In order to provide the rationale of this study, the results of previous research on seductive augmentation is presented in Table 2.1. Even though seductive augmentation has been suggested as a method for promoting learner’s interest, the research has reported mixed results on its effectiveness (Garner et al., 1991; Garner et al., 1989; Harp & Mayer, 1997, 1998; Mayer et al., 2001; Moreno & Mayer, 2000, 2002; Schraw, 1998). Thalheimer (2004) summarized the results of previous research by examining 24 comparisons from nine published articles. He categorized the types of augmentation into five areas and types of instructional materials into four as shown in Table 2.1. Of the 24 comparisons, 14 showed a seductive-augmentation effect for both recall of main idea or transfer, two showed a seductive-augmentation effect for transfer nut not for recall, and eight showed no seductive-augmentation effect. The cases when seductive augmentation effect was not shown were when printed seductive details are added to narrative biographical text and when context-appropriate sounds were added to multimedia narrated animation & CBI (Computer Based Instruction).

18 Table 2.1 The types of augmentation and instructional materials (Thalheimer, 2004)

Type of Augmentation Type of instructional materials Type Description Type Description Sentences added to text Text passages passages where sentence focused on a single Seductive details content is unimportant but Text passage person (often a interesting historical figure). Photos/Illustrations that Text passages are designed to promote focused on interest explaining a topic, Photos/Illustrations Text passage principle, or describing how something works. Sounds that were Expository text contextually appropriate passages presented to the visuals being Expository on a computer. Sounds displayed (e.g., wind for multimedia visual depicting downdraft). Music not necessarily Expository narrated relevant or associated to animations the visuals being Multimedia presented on a Music displayed (e.g., bland Narrated computer. synthesized music for a Animation visual depicting downdraft). Short video clips that were intended to be Video interesting, and relevant, but not explanatory.

The use of animated pedagogical agents in multimedia learning environments has increased as new technologies have made them more accessible (Craig & Gholson, 2002; Johnson, Rickel, & Lester, 2000). One of the advantages to implementing pedagogical agents in a computer interface is that the agent can increase the computer’s ability to engage learners and motivate them. In multimedia learning environments when pedagogical agent is implemented, learning content is delivered through instructional messages verbalized by the pedagogical agent as well as texts and graphics on the computer screen. According to the framework of interest

19 study, especially from the aspect of situational interest, the type of message is one of the factors that can affect learner’s interest. Additionally, the type of onscreen information such as graphics is also a factor. Therefore it is necessary to investigate the effect of the graphics delivered through instructional material and the effect of the messages delivered by pedagogical agents in terms of seductive augmentation.

Pedagogical Agent Supported Learning

Pedagogical Agent and Interest

Pedagogical agents are animated life-like characters designed to facilitate learning in computer-mediated learning environments (Johnson et al., 2000). Regarding the effects of pedagogical agents, researchers using lifelike pedagogical agents have indicated positive effects on learner’s attitude toward learning and performance (Baylor, 2002a, 2002b; Baylor & Ryu, 2003; Moreno et al., 2001). Also researchers have found positive effects of pedagogical agents on learner’s interest and motivation. Two explanations were suggested in regard to the positive effects of pedagogical agents on learner’s interest (Baylor & Ryu, 2003). The first reason was agent persona. Learners tend to like a pedagogical agent because of the agent’s character or personality based on agent’s persona composed of image, animation, voice, and emotional expressions. The second reason was the information provided by agent. Learners reported that they liked an agent because they found the quality of the message delivered by the agent was informative and helpful. The research using agents with evasive emotion and no motivational support also found that learner’s interest was aroused from the agent’s persona and information delivered by agent (Warren et al., 2004). Therefore, the variables affecting learner’s interest could be categorized into two dimensions, one is an agent’s persona and the other is information delivered by the agent. In this paper, the variables affecting learner’s interest are described in detail in terms of the persona of pedagogical agents and the message delivered by the pedagogical agent.

20 Features of Pedagogical Agent Affecting Learner’s Interest

Pedagogical agent’s persona: The key characteristics that constitute a pedagogical agent’s persona include its propensity to be engaging, person-like, credible, and instructor-like (Baylor & Ryu, 2003). Engaging personas of agents facilitate the learner-agent relationship and motivates the learner to be involved in the learning task. Person-like personas of agents form a viable relationship with the learner. Credible personas of agents make learners confident in the agent and help them recognize the agent as trustworthy, competent, and consistent in behavior. Finally, instructor-like personas of agents serve as a pedagogical mentor to effectively represent the content and pedagogy (Baylor, 2000). Media features such as voice, emotional expression, gesture, image, and animation are integrated to create such personas. The effect of a pedagogical agent on interest is explained also by social presence theory. According to Short, Williams, and Christie (1976), social presence is a subjective quality of the communication medium and can be a function of both verbal cues (e.g. tone of voice) and nonverbal cues (e.g. facial expression, direction of gaze, posture, and dress). Hence, a pedagogical agent is considered to have higher social presence than other computer mediated media, because pedagogical agents provide nonverbal cues as well as verbal cues, which can lead to promoted learner interest. Therefore media features are essential in terms of constructing agent personas and indicating social presence of an agent to promote learning interest. Voice has been suggested as a key aspect for enhancing agent persona. In fact, the voice feature has been found to be a critical element for agent-based learning environment. Prior research has indicated that voice has a superiority effect to visual appearance for communication in computer-based media (Mayer & Moreno, 1998; Moreno & Mayer, 1999), and with pedagogical agents in particular. While there are consistent results that voice (in conjunction with text) is a key aspect for enhancing agent persona, the role of agent image and animation is not proven (Baylor & Ryu, 2003). However, given the importance of emotional expression by an agent, animation would be necessary for an agent to demonstrate facial expressions. Animation with emotional expression increases the agent’s persona of credibility and also improves social presence of the agent by increasing non- verbal behavior. Baylor (2003) found that animation provides the most positive impact for an agent to be perceived as engaging, because it contributes to the agent expression of a personality

21 through non-verbal behavior, leading to be more likable, and thus more enjoyable to learn with and also more emotionally expressive.

Message delivered by Pedagogical agent: Agents also facilitate interactive learning with learners by delivering messages to convey information. The type of information delivered by a pedagogical agent through messages is also a variable that affects learning interest. Information delivered by agent could be directly related to the learning content by signaling the structure of the learning content so that learners can promote the understanding of the content. Also, information could be interesting but irrelevant to understanding the learning content. According to Kintch (1980), information has an effect on the learner’s interest in two different ways depending on the features of information and he suggested the concept of cognitive interest and emotional interest. The cognitive interest influences learner’s cognition by promoting the reader’s structural understanding of the content. On the other hand, emotional interest is explained by the addition of interesting and irrelevant material to understand the objectives of instructional material. However, it energizes a learner’s arousal so that they pay more attention and learn more overall (Harp & Mayer, 1997, 1998). Researchers have used the term seductive detail to refer to interesting but irrelevant details that are added to a passage to make it more interesting in the reading education field. In multimedia learning, seductive detail usually is called seductive augmentation because the term “seductive augmentation” refers to not only text but also graphics, narratives, voice, animation, and text accompanied in a multimedia learning environment with the purpose of increasing learner’s situational interest (Thalheimer, 2004). As being illustrated in the previous sections, two opposing suggestions have been indicated regarding the usefulness of seductive detail. The first suggestion asserted that it energizes readers so that they pay more attention to learning and learn more overall, because it influences the learner’s affect by promoting his/her enjoyment of the topic. Hence, it causes the learner to pay more attention to and encode more of the information from the material (Izard & Ackerman, 2000; Kintsch, 1980). The second suggestion contends the opposite position. It emphasizes that seductive detail disrupts the learner’s construction of the material comprehension so that adding it to the material will result in decreases on tests of retention and on solutions to transfer problems (Wade & Adams, 1990). In summary, both agent itself including agent persona and social presence, and the

22 information delivered by the agent plays a critical role in promoting learning interest from the aspect of situational interest. In particular, media features are important factors for constructing an effective agent persona. Information type is also a critical factor because it determines the quality of information delivered by agent. Figure 2.3 illustrates the detailed relationship describing the features of a pedagogical agent affecting learner’s situational interest.

Pedagogical Agent (1)

Auditory message (Seductive narrative)

Internalize Situational interest in IP Individual interest in IP (2)

Agent persona (Role) Learning Interest in IP

(1) Auditory messages delivered by pedagogical agent (2) The persona of pedagogical agent

Figure 2.3 The features of pedagogical agent affecting learner’s situational interest

Graphics and Interest

The effective use of graphical illustrations in designing instructional material has been suggested as an important facet of instructional message design (Anglin, Towers, & Levis, 1996). Using illustrations in instructional material is an effective method to promote interest and support learning because it can be used as an interest-getting device and it also helps learners interpret and remember the context of illustrated text. Since Spaulding (1955) reviewed sixteen studies conducted between 1930 and 1953 on using illustrations in instruction, many researchers have investigated the effects of illustrations on knowledge acquisition in instructional settings

23 (Brody, 1984; Concannon, 1975; Harp & Mayer, 1997; Holliday, 1973; Levi & Lentz, 1982; Levin & Mayer, 1993; Mayer, 1989; Mayer, Steinhoff, Bower, & Mars, 1995; Samuels, 1970; Schallert, 1980). Reviewing those studies, Anglin et al. (1996) concluded that visual illustrations can facilitate the acquisition of knowledge when they are presented with text material concurrently. However, it also has been pointed out that the results of illustration related research can not be integrated across all studies because of the lack of connections in terms of the function of the illustration in the instructional treatment (Anglin et al., 1996). In other words, each illustration related study was focusing on the one function of illustration that might be different from the function of other illustration related research. In order to avoid the generalization of the results of illustration related research, the functional framework was suggested. This functional framework provides assistance in classifying visual illustrations into meaningful functional categories. Using this framework, the research results can be combined and generalized differently depending on the function of illustration. Therefore it is critical to determine the specific function of the illustration before conducting actual research on the effect of a visual illustration. Regarding the function of illustration, Levie and Lantz (1982) suggested a functional framework that includes classifying illustrations in text based on how they impact a learner. According to them, framework contains four major functions which are attentional, affective, cognitive, and compensatory. The attentional function attracts or directs attention to the material. The affective function enhances enjoyment or affects emotion and attitude. The cognitive function serves to facilitate learning textual content through improving comprehension, increasing retention, or providing additional information. The last function, the compensatory function, is used to accommodate poor readers. Among those four functions, the previous research usually was focusing on only the cognitive function of illustration. It has not been answered clearly what effect the visual illustrations have on learner’s emotion and attitude and how the illustrations need to be designed to improve the affective function of illustration. In terms of the affective function of illustration, Kintch (1980) insisted that the visual illustration has an effect on the learner’s affective status in two different ways depending on the features of the illustration. According to his cognitive interest and emotional interest theory, visual illustration improves learner’s cognitive interest and it promotes learner’s emotional

24 interest as well depending on the characteristic of it. The cognitive interest influences learner’s cognition by promoting the reader’s structural understanding of the explanation. On the other hand, emotional interest is explained by the addition of interesting but irrelevant material to a textbook lesson. It energizes learner’s arousal so that they pay more attention and learn more overall. Therefore, two types of graphics are possibly designed; expository graphics to promote cognitive interest and seductive graphics to raise emotional interest. An expository graphic is defined as the illustration that signals the structure of the explanation. A seductive graphic is defined as the illustration which refers to interesting but irrelevant illustration to understand the structure of text. However, it plays an important role from the motivational aspect, because it increases emotional arousal and further influences the leader’s cognitive process. In a same vein, Levie and Lantz (1982) reviewed previous research comparing three separate research areas concerning the role of illustration in learning. The first research area concerned learning illustrated text information, the second was about learning non-illustrated text information, and the last was about learning using a combination of illustrated and non- illustrated text information. According to the cognitive interest theory, the illustrated text information is similar with the cognitive interest illustration. Also, the non-illustrated text information is comparable to the emotional interest illustration. They concluded that learning would be facilitated when the information in the written text is depicted in the illustrations and learning of text material would not be helped nor necessarily hindered with illustrations that are not related to the text. However, they didn’t consider the affective function of illustration separately. Harp and Mayer (1997) examined the effects of emotional interest adjunct and cognitive interest adjunct on information retention, learning transfer, and learning interest. They reported that learners in the base text group recalled the most, whereas learners who read passages containing the base text along with emotional interest text and illustrations recalled the least. This result was consistent with the prediction of cognitive interest theory and inconsistent with the prediction of emotional interest theory. The present study is in a line of learning interest related studies, especially from the aspect of situational interest using seductive details/augmentation. In this study, interest is defined as a psychological state of the person, comprised of individual interest and situational interest.

25 Reviewing previous research revealed a lack of studies investigating the effects of seductive augmentation in multimedia learning, especially with the assistance of a pedagogical agent. In this study, three major issues are raised. The first issue is that it is doubtful if the seductive detail effect would be still remaining when the instructional material is delivered through multimedia presentation. In the multimedia learning, courseware screens consist of text, graphics such as still illustration and/or animation, and video clip. Therefore different types of seductive details/augmentation can be implemented simultaneously. In multimedia learning with the assistance of a pedagogical agent, seductive augmentation can be expanded to the agent’s messages as well. Thus two sources of seductive augmentation, one from agent’s message and the other from graphics included in the instructional material, are presented concurrently. Previous research has applied only one source of seductive augmentation in a multimedia learning environment. Therefore the question, whether the seductive detail effect would be still valid when learners are given two sources of seductive augmentation stimulus still needs to be answered. The second issue is the nature of the instructional material. Previous studies usually employed paper-based instruction materials. Learners were not allowed to read the entire text multiple times, but forced to study only one time for short moment. In multimedia learning, learners are allowed to read the information on screen more than once. Learners have a more control over the process of learning in multimedia material than learners do in paper-based material setting. The third issue is that the results of achievement tests can be different depending on the type of text. If it contains factual information rather than procedural information usually used to explain scientific principles, the learners who are given the instructional material containing seductive graphics may recall the main ideas differently from the learners who are given the instructional material without seductive graphics. In order to answer above three issues, this study was designed with learner controlled multimedia learning material containing factual information and supported by a pedagogical agent. As illustrated in Figure 2.4, the type of messages delivered by the agent, the role of the agent represented by agent’s persona, and the type of graphics used in the instructional material are established as features affecting learner’s situational interest in this study.

26 Therefore the purpose of this study is to investigate the effects of seductive augmentations and the role of an agent on learning interest, achievement, and attitude in a pedagogical agent supported learning environment.

Pedagogical Agent Instructional Material

Auditory messages (Seductive narratives)

Visual messages (Seductive graphics)

Internalize Situational interest in IP Individual interest in IP

Agent persona (Role) Learning Interest in IP

Figure 2.4 The features affecting learner’s situational interest in pedagogical agent supported multimedia learning

Hypotheses

This study was designed to investigate three main research questions. For each research question, a theoretical hypothesis and operational hypotheses are presented in this section. Eight conditions were implemented in this study to investigate the treatment effects on learning interest, achievement, and attitude toward instructional material: 1) seductive graphics with seductive messages delivered by the companion role of agent (Sg-Ca), 2) seductive graphics with seductive messages delivered by the instructor role of agent (Sg-Ia), 3) seductive graphics with seductive messages in text format (Sg-Tx), 4) seductive graphics with no seductive messages (Sg-No), 5) no seductive graphics with seductive messages delivered by the companion role of agent (Ng-

27 Ca), 6) no seductive graphics with seductive messages delivered by the instructor role of agent (Ng-Ia), 7) no seductive graphics with seductive messages in text format (Ng-Tx), and 8) no seductive graphics with no seductive messages (Ng-No). Table 2.2 shows the list of the acronyms used to refer the treatment groups in this study.

Table 2.2 The list of the acronyms of treatment groups

Treatment Group Acronym The source of seductive of Seductive graphic message condition Condition 1 seductive graphics (Sg) the companion role of agent (Ca) Sg-Ca Condition 2 seductive graphics (Sg) the instructor role of agent (Ia) Sg-Ia

Condition 3 seductive graphics (Sg) text format (Tx) Sg-Tx

Condition 4 seductive graphics (Sg) No (No) Sg-No

Condition 5 no seductive graphics (Ng) the companion role of agent (Ca) Ng-Ca

Condition 6 no seductive graphics (Ng) the instructor role of agent (Ia) Ng-Ia

Condition 7 no seductive graphics (Ng) text format (Tx) Ng-Tx

Condition 8 no seductive graphics (Ng) No (No) Ng-No

Theoretical hypothesis 1: Because seductive augmentation affects situational interest, there will be significant differences among the eight conditions in terms of learning interest.

Operational hypothesis 1.1: The mean score for learning interest for the Sg condition will be significantly higher than that for the Ng condition.

Seductive augmentation, which refers to the highly interesting but unimportant text segments, is considered a factor affecting emotional interest in terms of situational interest

28 (Thalheimer, 2004). Thus it increases interestingness of situation while students are studying instructional material given to them. In this study, seductive graphics were used to improve students’ situational interest. Therefore, it was predicted that the Sg condition will show significantly higher learning interest score than the Ng condition.

Operational hypothesis 1.2: The mean score for learning interest for the Ca and Ia conditions that contain an agent as a source of seductive messages will be significantly higher than that for the Tx condition that contains seductive messages in text format and the No condition that contains no seductive messages.

According to social presence theory, social presence is a subjective quality of the communication medium and can be a function of both verbal cues (e.g. tone of voice) and nonverbal cues (e.g. facial expression, direction of gaze, posture, and dress) (Short, Williams, and Christie, 1976). Hence, a pedagogical agent is considered to have higher social presence than other computer mediated media because the pedagogical agent provides nonverbal cues as well as verbal cues, which can lead to promoted learner interest. Since the Ca and Ia conditions included verbal cues containing seductive messages delivered by a pedagogical agent, whereas the Tx condition containing seductive messages delivered by text box only, the learning interest scores of the Ca and Ia conditions were predicted to be significantly higher than the Tx condition. In the same vein, it was also expected that the learning interest scores of the Ca and Ia conditions would be significantly higher than the No condition because the No condition did not contain seductive messages nor a pedagogical agent that is necessary to promote situational interest.

Operational hypothesis 1.3: The mean score for learning interest for the Ca condition will be significantly higher than the Ia condition, when a pedagogical agent is presented.

In terms of the role of agent, the key characteristics that constitute a pedagogical agent persona include its propensity to be engaging, person-like, credible, and instructor-like (Baylor & Ryu, 2003). Media features such as voice, emotional expression, gesture, image and

29 animation are integrated to create such a persona. In order to improve the agent’s persona, it was found that voice has a superiority effect to visual appearance for communication in computer- based media, particularly with pedagogical agents (Mayer & Moreno, 1998; Moreno & Mayer, 1999). Information to be delivered through the agent’s voice also needs to be consistent with the appearance of the pedagogical agent to improve the agent’s persona as a companion. Therefore, in this study, seductive messages were predicted to be well delivered when the companion role of agent is present by contributing to improving the persona of the companion role of agent. Therefore, it was predicted that the mean score of learning interest for the Ca will be significantly higher than for the Ia condition when a pedagogical agent is presented.

Theoretical hypothesis 2: Because seductive augmentation affects situational interest, there will be significant differences among the eight conditions in terms of achievement.

Operational hypothesis 2.1: The mean score for the achievement test between the Sg condition and the Ng condition will be significantly different.

Seductive graphics are expected to play a role to improve students’ situational interest by promoting their emotional interest. Using graphics in instructional material has been suggested as an effective method to promote interest and support learning because graphics can be used as an interest-getting device and they also help learners interpret and remember the context of illustrated text. Theoretically, seductive graphics are defined as the illustration which refers to interesting but irrelevant information to understand the structure of text. And seductive graphics play an important role from the motivational aspect, because it can increase emotional arousal and further influence the reader’s cognitive process. However, it also has been pointed out that the use of seductive graphics in instructional material can show unexpected effects on cognitive process in the opposite direction. “Seductive detail/augmentation effect” is a term used to describe the negative cognitive consequences of adding seductive details to text. According to the seductive augmentation effect, adding seductive augmentation disrupts the learner’s construction of the main idea so that adding seductive augmentation to the material will result in decreases on tests of retention and on

30 solutions to transfer problems. This effect is typified by proficient recall of seductive details within a passage but poor recall of structurally important ideas (Garner, et al. 1989). Seductive augmentation effect occurs when only seductive augmentation is presented as the Sg condition does in this study. Previous research showed that seductive augmentation effect hurts both recall of main ideas and problem solving transfer of critical information when seductive illustrations are added to expository text (Harp & Mayer, 1997; 1998). The text used in this study was about “intellectual property” that is mostly related to the concepts, examples, and cases of three laws: patent, trademark, and copyright. In this study, the Sg condition was given seductive graphics and the Ng condition was given no seductive graphics. Given the two opposing perspectives described above, emotional interest perspective and seductive detail effect perspective, it was predicted that there will be a difference on achievement scores between the condition with seductive graphics and the condition without seductive graphics.

Operational hypothesis 2.2: The mean score for achievement test for the Ca and Ia conditions that contain an agent as a source of seductive messages will be significantly higher than that for the Tx condition that contains seductive messages in text format and the No condition that contains no seductive messages.

Previous studies showed that seductive augmentation effect was not found in two studies out of three when context-appropriate sound was added to multimedia narrative animation and CBI (Moreno & Mayer, 2000; 2002). Although the messages from pedagogical agents were implemented as narratives in this study, as Thalheimer (2004) pointed out, the context- appropriate augmentation can be easily processed together with the visual information garnered by reading the onscreen text. Therefore, it was predicted that the achievement score for the Ca and Ia conditions will be significantly higher than that for the Tx and the No conditions.

31

Operational hypothesis 2.3: The mean score for achievement test for the Ca condition will be significantly higher than the Ia condition, when a pedagogical agent is presented.

As explained in hypothesis 1.3, voice performs an important role for communication in computer-based media (Mayer & Moreno, 1998; Moreno & Mayer, 1999), and with pedagogical agents in particular. Information delivered and voice used by agents should be consistent to improve learner’s perception of agent persona. Thus, seductive messages are expected to be delivered well when the companion role of agent is embedded with that message. Therefore, it was predicted that achievement test score for the Ca condition will be significantly higher than that for the Ia condition when a pedagogical agent is presented.

Theoretical hypothesis 3: Because seductive augmentation affects situational interest, there will be significant differences among the eight conditions in terms of attitude toward to the instructional material.

Operational hypothesis 3.1: The mean score for attitude for the Sg condition will be significantly higher than that for the Ng condition.

In this study, students’ attitude toward instructional material was measured using IMMS (Instructional Material Motivational Survey) developed by Keller (1993). IMMS is designed to measure students’ attitude toward instructional material based on four motivational components, attention, relevance, confidence, and satisfaction. In this study, the three components of relevance, confidence, and satisfaction were used to measure students’ attitude. The attention component was used to measure the dependent variable of situational interest. It should be noted first that there have been no empirical data to support the role of seductive augmentation on students’ attitude toward instructional material. However the prediction can be made based on a previous study conducted by Wade et al. (1993). Wade et al. (1993) categorized text information into four different types and asked readers how they would describe the text excerpts. Types of text information were important

32 factual details, main ideas, boring trivia, and seductive details. Students described the seductive details as most interesting and important factual details as most uninteresting. However, both important factual details and main ideas were rated as most important. Seductive details were conceived as most unimportant. In terms of easiness to understanding and remembrance, seductive details were ranked first. Boring trivia and main idea ranked next. There was only 4% difference between boring trivia and main ideas. Based on the results of this study, seductive details of information were considered interesting but not important. On the contrary, factual details of information were conceived as uninteresting but important. Based on this study, it was anticipated that the attitude survey score for the Sg condition will be significantly higher than that for the Ng condition because the Sg condition contains factual information as well as seductive augmentations.

Operational hypothesis 3.2: The mean score for attitude for the Ca and Ia conditions that contain an agent as a source of seductive messages will be significantly higher than that for the Tx condition that contains seductive messages in text format and the No condition that contains no seductive messages.

In this study, all conditions contained factual information. Since the Ca and Ia conditions contained more seductive augmentation than the Tx condition, the attitude score for the Ca and Ia conditions was predicted to be significantly higher than the Tx condition. In a same vein, it was also expected that the attitude scores for the Ca and Ia conditions will be significantly higher than the No condition because the No condition didn’t include seductive augmentation which is necessary to promote situational interest.

Operational hypothesis 3.3: The mean score for attitude for the Ca condition will be significantly higher than the Ia condition, when a pedagogical agent is presented.

Information delivered and the voice used by agent should be consistent to improve learner’s perception of agent persona, as in the hypothesis 1.3 and 2.3. Therefore, it was

33 predicted that the score for attitude measurement for the Ca condition will be significantly higher than that for the Ia condition when a pedagogical agent is presented.

34

CHAPTER III METHOD

Introduction

This study investigated the effects of seductive augmentation and the role of a pedagogical agent on learning interest, achievement, and attitude toward instructional material in a pedagogical-agent supported learning environment. This chapter describes the research methodology including participants, research design, materials, variables, measurement instruments, experiment procedure, and data analysis methods.

Participants

The participants were 127 college undergraduate students enrolled in “Computer literacy” classes in a large public university located in the southeastern United States. This course was one of the required courses for the undergraduate students. All participants were recruited from ten sections of the course and offered extra credits as compensation. Only participants who agreed to participate in the study by signing the consent form were included in the final data analyses. A total of 136 students voluntarily participated at the beginning of the study. Of these, 127 participants were included in the final data analyses because nine students didn’t complete the post questionnaire. All of the 127 participants were undergraduate students. The average age of the sample was 19.72 years (SD = 1.96). Among those 127 participants, 63.0% were Caucasian, 19.7% were

35 African-American, 11.0% were Hispanic/Latino, 2.4% were Asian/Asian American, 1.6% were bi-racial, and 2.3% were other ethnicity groups. There were 60.6% of male students and 39.4% of female students. The majority of the participants were sophomores (43.3%) with 21.3% freshmen, 20.5% juniors, and 15.0% seniors. The required sample size was estimated by using pre-determined alpha, effect size, and power. The sample size was determined using Cohen’s guideline (1988), with a power level of 0.9, and large effect size of .50 at α-level .05. Using G*Power program, which is a general power analysis program, the minimum required sample size was twelve in each treatment. Thus, this research required at least ninety-six subjects (pq = 12 x 8=96) total for the eight treatment conditions. In this study, total 127 subjects participated ranging fifteen to seventeen subjects in each treatment condition. Therefore, it was verified that the number of participants were enough to conduct the study.

Research Design

The study was designed as a 2 × 4 factorial design. The variables included seductive graphics (present vs. absent), and the source of seductive messages (companion role of agent, instructor role of agent, text without agent, and no message).

Random Group Group Treatment Posttest Assignment a b A X 1 Y 1 O1 B X1 Y2 O2 C X1 Y3 O3 Computer literacy D X Y O R 1 4 4 class students E X2 Y1 O5 F X2 Y2 O6 G X2 Y3 O7 H X2 Y 4 O8 a First independent variable: the use of seductive graphic (1: presence, 2: absence) b Second independent variable: the source of seductive messages (1: companion role of agent, 2: instructor role of agent, 3: text message without agent, 4: no message)

Figure 3.1 Research design

36 This study employed a randomized group post-test design. In order to test hypotheses and explore research questions, participants were randomly assigned into the one of the eight conditions (A through H) based on the sequence of their entry to the research lab as shown in Figure 3.1.

Independent Variables

The independent variables in this study included the seductive graphics and the source of seductive messages. The first independent variable, seductive graphics were implemented with two levels: 1) presence of seductive graphics and 2) absence of seductive graphics. The second independent variable, the source of seductive messages was implemented with four levels: 1) companion agent delivered message, 2) instructor agent delivered message, 3) text message without agent, and 4) no message. The eighth condition that contained neither seductive graphics nor seductive messages was considered the control group.

The presence / absence of seductive graphics

The first independent variable was the use of seductive graphics. Seductive graphic is defined as graphics/illustrations that are interestingly relevant to learning content but irrelevant to the learning objectives. As the topic of instructional material for this study was “intellectual property”, the seductive graphics/illustrations related to the topic were used as shown in Figure 3.2. In this study, two levels of seductive graphics were implemented; one was the presence of seductive graphics and the other was the absence of seductive graphics. Seductive graphics included in this study were collected from the website titled CartoonStock®. (http://www.cartoonstock.com/default.asp). CartoonStock® is a searchable database of over 60,000 quality gag cartoons, political cartoons, cartoon pictures, and illustrations by over 290 of the world's best cartoonists. The permission of using graphics in this study was obtained from CartoonStock® via email prior to the main experiment.

37

Figure 3.2 The example of seductive graphic

The source of seductive messages

The second independent variable was the source of seductive messages. Seductive message is defined as script that is interesting but irrelevant to learning contents. In this study, seductive narratives were designed to provide learners stories regarding the usage of intellectual property as shown in Figure 3.3.

Seductive “The Annual Consumer Electronic Show in Las Vegas is an internationally known message: show. Just last January over 120,000 people from over 110 countries attended the show to see the latest product innovations from the world's leading companies! That’s amazing…Don’t you think? It must be important. Well, without intellectual property protection, many of these companies would not invest billions every year to launch new

waves of products.”

Figure 3.3 The example of seductive message

Four levels of the source of seductive messages were implemented in this study. The first source of seductive messages was the companion role of agent. The second source of

38 seductive messages was the instructor role of agent. Both roles of agent were employed using two pedagogical agents as shown in Figure 3.4. The third source of seductive messages was a text message without an agent. The last condition employed no seductive message and was considered as a control condition. The contents and length of the seductive messages were identical across the four different conditions as shown in Appendix H except that agent delivered messages utilized agent personal name such as Mike for a companion role of agent and Dr. Handriks for an instructor role of agent.

Agent representing a companion role Agent representing an instructor role

Figure 3.4 The example of a companion role of agent and an instructor role of agent

In regard to the use of agent as a source of seductive message, two agents (a companion role of agent and an instructor role of agent) were implemented in this study. These two pedagogical agents were developed and validated by RITL (Center for Research of Innovative technologies for Learning) at Florida State University (Baylor, 2003). Agent roles were implemented through the general persona represented by image, voice, animation, and affect. The companion agent had expressive animations in terms of facial expressions while the instructor agent showed the most subdued expressions. In addition, the companion role of agent named “Mike” was designed with a friendly and energetic voice, youthful appearance, and

39 expressive emotions. Yet, the instructor role of agent named “Dr. Handricks” was designed with a dry and straight-forward voice with little inflection, older appearance, little animation, and informative and directive characteristics.

Dependent Variables

Dependent variables for the study included level of interest, achievement, and attitude toward the instructional material in this study.

Learning interest

The learning interest in this study is defined as the psychological status within a person as a combination of individual interest and situational interest. Therefore, learning interest was measured in terms of individual interest and situational interest as shown in Appendix B and Appendix C. Prior to beginning the instructional material, participants were informed that the topic of the instructional material is “Intellectual property”. Then, the level of individual interest was measured from two aspects: 1) feeling-related interest and 2) value-related interest. To measure feeling-related interest, the participants indicated the feeling they expected to have while studying the instructional material. A five-point Likert scale was provided with response choices ranging from “not at all true” to “very true”. The participants were asked to read the following five adjectives in estimating their expected feelings (“While studying the instructional material on “Intellectual Property”/ I expect to feel…”) “bored”, “stimulated”, “interested”, “indifferent”, and “engaged”. The reliability of the feeling-related interest survey for this study was .74. In order to estimate value-related interest aspect of individual interest, participants were asked to read the terms “meaningful”, “unimportant”, “useful”, and “worthless” to describe the value of the topic “Intellectual Property” to them personally. A five-point Likert scale was provided with response choices ranging from “not at all true” to “very true”. For each participant, a score of individual interest was computed by adding feeling related interest scales and value related interest scales. Previous studies have shown that this measure of interest is

40 unidimentional and highly reliable (Schiefele, 1996, 1998; Schiefele & Krapp, 1996). The reliability of the value-related interest survey for this study was .75. Situational interest was measured from three aspects: 1) arousal, 2) involvement, and 3) attention as included in Appendix C. In order to measure the arousal level, five items from the AD-ACL (Activation- Deactivation Adjective Check List) were used (Thayer, 1985, 1986). The AD-ACL is a multidimensional test of various transitory arousal states, including 20 items within four main categories; energy (general Activation), tiredness (deactivation-Sleep), tension (high activation), and calmness (general deactivation). According to Thayer (1986), self-reports of arousal are more valid than physiological indicators and it has been used widely in many psychological contexts. Since this study focused on the arousal state of participants, only the energetic dimension of AD-ACL was selected in this study because only the energetic dimension is involved in the motivational process and interest. The original items of AD-ACL used four-point Likert scale, but the original scales were modified to the five-point Likert scale to fulfill the purpose of this study. The reliability of the arousal measurement for this study was .91. In order to measure participants’ involvement, two dimensions (intensity and persistence) were considered based on Reynolds’ distinction (Reynolds, 1992). The intensity dimension is measured by self-report items, and the persistence dimension is measured by means of recording the subjects' participating times (Schiefele & Krapp, 1996). In this study, only the intensity dimension was employed because it was assumed that the persistence dimension would be very different due to different amount of seductive details in each condition. The intensity dimension was assessed by two items “I was completely caught up in what I was studying”, and “When learning from the material, I was concentrated”. A five-point Likert scale was provided with response choices ranging from “not at all true” to “very true”. The reliability of the involvement measurement for this study was .73. In order to measure participants’ attention level, attention sub scale from Keller’s IMMS (Instructional Material Motivation Survey) was employed (Keller, 1993). The IMMS is intended to be a situational measure of students’ motivational reaction to instructional materials and designed with the theoretical foundation represented by the ARCS model (Keller, 1987). ARCS is an acronym of four motivational components; attention, relevance, confidence, and satisfaction. The original statements of attention measurement were changed based on the context. For

41 example, an original item, “There was something interesting at the beginning of this lesson that got my attention” was revised to “I found something interesting at the beginning of this instructional material that got my attention” The reliability of the original attention measurement is .89 (Keller, 1993), but the response reliability of the survey for this study was .83.

Achievement

Student achievement was measured on two levels: 1) recall test and 2) comprehension test.

Recall test: A recall test was designed to assess students’ ability to recall as many keywords as possible from the instructional material on intellectual property included in Appendix D. The recall test was conducted on the computer screen where the following instruction was typed: "Please type in every keyword you can remember from this material" Participants were allowed to type in as many keywords as they could recall from the instructional material. Participants were not allowed to reproduce the original wording. The number of keywords recalled correctly was calculated. Typed answers of the learners’ recalls were scored based on the list of candidate keywords created by the researcher. The list was created before grading the recall test. Then, keywords in the answering box were matched with those keywords in the list of candidate keywords that conveyed the same idea. For each keyword in the answering box, if the keyword was contained in the list of candidate keywords, one point was assigned. For all other keywords in the answering box that were not found in the list of candidate keywords, zero points were assigned. The purpose of the recall test was to examine how many main keywords participants were able to remember from the main text, not from the seductive text or messages.

Comprehension test: The comprehension of the instructional material was measured by a comprehension test as included in Appendix E. The comprehension test was designed to assess students’ ability to select the correct information by applying what they learned from the instructional material, without further access to the material. To determine the validity of test items, the pilot test was conducted a month prior to the main study. The purpose of the pilot test was to determine the level of prior knowledge

42 participants already had on “intellectual property” and to find a size of variance among participants’ test scores. The total number of items on the initial achievement test was eleven including seven true/false items, three multiple-choice items, and one open-ended question as shown in Table 3.1. There was no time limit imposed for answering the questions in the pilot study.

Table 3.1 Achievement test items

Contents Number Item type Item number of items Intellectual property 1 Open-ended Item 11 Patent 4 Three True/False Item 2, 3, 4 One multiple choice Item 9 Trademark 3 Two True/False Item 5, 6 One multiple choice Item 10 Copyright 3 Two True/False Item 1, 7 One multiple choice Item 8

Three items among eleven items referred to a concept and conditions of a patent, one type of intellectual property. These questions required knowledge about the features of a patent and three conditions to be a patent. Another three items among eleven items referred to a feature of a trademark and three advantages of registering a trademark. The remaining three items referred to the concept of a copyright. True/false items were constructed to test students’ ability to remember the factual knowledge about a patent, a trademark, and a copyright correctly. True-false questions require the students to select a response (true or false) that shows recognition of correct or incorrect information that is presented to them. True-false questions are well suited for testing student recall or comprehension. Multiple choice items were constructed to test students’ ability to compare and evaluate related ideas and concepts of a patent, a trademark, and a copyright. One of the items was an open-ended question that referred to the relationship among intellectual property, a patent, a trademark, and a copyright. Students were requested to explain in writing how the intellectual property and three sub concepts are related. The purpose of the open-ended

43 question was to assess students’ general understanding of the main topic “Intellectual property.” As a foundation of constructing eleven comprehension items, instructional objectives and unit content major concepts to be covered were utilized. First, instructional objectives were listed. Then, researcher and a research assistant identified most main keywords from the list of main keywords candidate list created for the keywords recall test. All items were constructed based on the information used to create the instructional material as presented in Appendix A. It was assumed that, if the general level of prior knowledge is low, identical items would be used in the main study. If the general level of prior knowledge is high, eleven items would be re-examined and revised so that participants would receive more challenging items. Besides, if the size of variance among students is low, then it proves that there is not much difference among participants’ prior knowledge level. Then, current test items would be used in the main study. If the size of variance among students is high, it proves that there is much difference among participants’ prior knowledge level. Therefore, prior knowledge test would be used as covariate to control the effect of prior knowledge on achievement. The score of achievement items in pilot study was collected a month prior to the main study. Students participating in the pilot study were given eleven test items on the first day of computer literacy class and asked to answer the questions without any further information on “intellectual property.” The result of data analysis was used to determine whether to use prior knowledge test or not in the main experiment. The average score of the pilot comprehension test was 3.4 (SD = .92) out of total score 11. The second true/false item with item difficulty of .94 was problematic because 94% of students selected a correct answer. Therefore, the second item was dropped from the comprehension test. In the main study, the comprehension test without the second true/false item was given to students at the end of the material to assess their knowledge on “Intellectual property.” The comprehension test was graded based on a pre-determined answer sheet. Since there was one correct answer for each item, the answer sheet was prepared based on the information in the instructional material on intellectual property. Total possible score range was from 0 through 11. The reliability of the comprehension test was .27. The low reliability was occurred because each item was designed to measure students’ ability to compare and evaluate related ideas and concepts of a patent, a trademark, and a copyright. It was found that the average score of the comprehension test in the main study was 5.50

44 (SD=1.50) out of total score 10. Comparing with the average score from the pilot study, the score was increased 24% from 31% of correct answers to 55% of correct answers. Therefore it was concluded that the instructional material was effective for students to learn about intellectual property.

Attitude

The student’s attitude toward to the instructional material was measured in terms of three categories: 1) Relevance, 2) Confidence, and 3) Satisfaction. The IMMS developed by Keller (1993) was used because it allowed the researcher to measure the situational components of the ARCS model. Since the attention component was used to measure learning interest, only the remaining three components were used as included in Appendix F. The IMMS measurement included 36 items intended to be a situational measure of learners’ motivational reactions to the instructional material. Only 24 items except for the attention items was used in this study. The responses ranged from one to five on a Likert scale with nine relevance component items, nine confidence component items, and six satisfaction component items. The reliability of IMMS based on Cronbach’s alpha for each subscale was Relevance: .81, Confidence: .90, and Satisfaction: .92.

Treatment Group Description

Table 3.2 describes each treatment group and control groups. Each group in this study was different across eight treatment conditions. Participants were randomly assigned into one of the eight treatment conditions.

45 Table 3.2 Treatment conditions

Seductive graphics Presence Absence Condition 1 Condition 5 Companion • Seductive graphics • No seductive graphics role of agent • Seductive verbal messages • Seductive verbal messages delivered by companion agent delivered by companion agent Condition 2 Condition 6 Instructor role • Seductive graphics • No seductive graphics Source of of agent • Seductive verbal messages • Seductive verbal messages seductive delivered by instructor agent delivered by instructor agent message Condition 3 Condition 7 Text without • Seductive graphics • No seductive graphics agent • Seductive messages in text • Seductive messages in text Condition 4 Condition 8 No message • Seductive graphics • No seductive graphics • No seductive messages • No seductive messages

The first condition: The participants in condition 1 were presented the instructional material containing seductive graphics. Also, seductive verbal messages were delivered with the assistance of the companion role of the pedagogical agent. The instructional material consisted of text information explaining “intellectual property” as shown in Appendix I. The second condition: The participants in condition 2 were presented the instructional material containing seductive graphics. Also, seductive verbal messages were delivered with the assistance of the instructor role of the pedagogical agent. The instructional material consisted of text information explaining “intellectual property” as shown in Appendix J. The third condition: The participants in condition 3 were presented the instructional

46 material containing seductive graphics. Seductive messages were delivered as texts mode combined with existing text information on screen as shown in Appendix K. The fourth condition: The participants in condition 4 were presented the instructional material containing seductive graphics. Seductive messages were not used in this condition as shown in Appendix L. The conditions from fifth condition through the eighth condition were implemented with the same sources of seductive messages as used in the condition 1 through the condition 4. However, seductive graphics were not implemented in these conditions.

Material

Instructional material

The topic of instructional material used in this study was “Introduction to intellectual property.” In this material, three main sub-concepts of intellectual property including patents, trademarks, and copyrights were explained in detail as included in Appendix A. The following three textbooks were used as references in order to add detailed information to the instructional material.

1. Intellectual Property Examples & Explanations: Examples and Explanations (The Examples & Explanations Series) by Stephen M. McJohn, 2003 2. Intellectual Property: Patents, Trademarks, and Copyright by Arthur R. Miller, Michael H. Davis; Paperback., 2000 3. Essentials of Intellectual Property by Alexander I. Poltorak, Paul J. Lerner, 2002

The goal of this material was to introduce three basic sub-concepts and several basic rules of intellectual property to computer literacy class students. The topic covered three important areas in intellectual property, which are patents, trademarks, and copyrights. The topic of intellectual property was considered as appropriate for this study for three reasons. First, the topic “Intellectual property” has been included as one of the topics the class should cover.

47 Therefore, students were aware of the topic, but not familiar with detailed information, specifically regarding the areas of patents, trademarks, and copyrights. Second, as the primary goal of this study was to investigate the relationship between learning interest promoted by seductive augmentations and achievement, instructional text for which students presumably had low levels of prior knowledge needed to be used. Third, this topic was related to the students’ attitude in everyday life regarding how to use computer applications without violating any legal and ethical issues. Therefore, this topic was considered as appropriate for computer literacy course to examine students’ comprehension and attitude.

Instructional material on Intellectual property consisted of three learning phases. • Introduction phase: Students were given a brief introduction about intellectual property and basic information containing history, related regulations, and examples. • Learning phase: Students were given a detailed explanation in regard to the three sub- concepts such as patents, trademarks, and copyrights. • Test phase: After completing instructional material, students were given an opportunity to actually test what they had read in the instructional material.

In order to check the difficulty of the main contents, a readability index was calculated using the SMOG-Grading. The SMOG-Grading English texts was developed by McLaughlin in 1969 and it shows an appropriated level of school grade. The SMOG-Grading for the main text in instructional material in this study was 13.4 (grade). Therefore, the difficulty of text was confirmed. The material was developed as a web-based format. Students were forced to go through all screens in order from the introduction phase to the test phase. Only navigation moving onto the next screen was allowed in order to verify that students were not given a second chance to refer to the text information. When students completed the material, a [Next] button on the last material page led them to the post survey questionnaire.

48 Procedure

The purpose of this study was introduced at the beginning of the study and a short orientation was demonstrated. The instructional material was presented in a computer laboratory equipped with 24 individual personal computers. Participants participated voluntarily in the study as a part of class activity and randomly assigned into the one of the experimental conditions. As a reward, all participants were given 20 extra points in compensation. The treatment conditions were structured as shown in Table 3.3.

Table 3.3 The description of experimental conditions of the study

Group Group description Treatment group 1 Seductive graphics + Seductive verbal messages delivered by companion agent Treatment group 2 Seductive graphics + Seductive verbal messages delivered by instructor agent Treatment group 3 Seductive graphics + Seductive messages in text Treatment group 4 Seductive graphics + No seductive messages Treatment group 5 No seductive graphics + Seductive verbal messages delivered by companion agent Treatment group 6 No seductive graphics + Seductive verbal messages delivered by instructor agent Treatment group 7 No seductive graphics + Seductive messages in text Treatment group 8 No seductive graphics + No seductive messages

When participants log on the computer and connect to the web based study material, they were guided the on-line consent form describing that they would be studying multimedia instructional material on the topic of “Intellectual Property” and that, after they finish studying, they would be asked a series of questions about what they had studied. Since the instructional material was designed as a user controlled program, students were instructed to read the material carefully in their normal studying speed. Researcher and research assistant were present at all times to ensure that they were not distracted by other computer activities such as instant

49 messenger programs, playing on-line games, and email checking. Each participant was presented the instructional material corresponding to his/her treatment group and told to begin the material. They were not allowed to take notes or refer to other resources. When individual participants finished the instructional material, they were guided to the post measure instruments on screen. First, the learning interest items were asked and a recall test was given on screen. The database was used to collect the student’s answers automatically. Completing a recall test, participants were presented a comprehension test and an attitude survey in order. When the attitude survey was completed, participants were thanked for their participation and given extra credit points as compensation.

Data Analysis

The study was designed as a 2 × 4 factorial design. The variables included the use of seductive graphics (present vs. absent) and the source of seductive message (companion agent, instructor agent, text without agent, no message). First, preliminary data analyses were conducted to detect problematic observations and to assess violations of the assumptions for statistical procedures. In a primary data analysis, the main effect of two independent variables was conducted for three dependent variables: 1) learning interest score, 2) achievement score, and 3) attitude score. The significance level for all the analyses was set at α < .05.

Learning interest

The learning interest data was analyzed according to the two groups of dependent measures: 1) individual interest and 2) situational interest. Individual interest data was collected prior to the main experiment so as to be compared among conditions as a prior interest. The scores for individual interest were added to the scores for situational interest that were collected after the main study and together constructed a learning interest score, the first main dependent

50 variable in this study.

Operational hypothesis 1.1: The mean score for learning interest for the Sg condition will be significantly higher than that for the Ng condition.

Operational hypothesis 1.2: The mean score for learning interest for the Ca and

Ia conditions that contain agents as a source of seductive messages will be

significantly higher than that for the Tx condition that contains seductive messages

in text format and the No condition that contains no seductive messages.

To investigate the effects of seductive graphics augmentation on learning interest (Hypothesis 1.1 and 1.2), a two way between-groups multivariate analysis of variance (MANOVA) was performed. Learning interest was defined as a combination of individual interest and situational interest. Therefore, the learning interest score was collected by combining individual interest scores and situational interest scores. The hypothesis 1.1 was tested using five dependent variables: feeling-related interest, value-related interest, arousal, involvement, and attention. The independent variable was the use of seductive graphics and the type of seductive messages. The use of a two way MANOVA in this study was justified because it controls the increased risk of Type I error. Also, correlation among dependent variables was examined to confirm the issues regarding multicollinearity and singularity among dependent variables. It was found that five dependent variables were moderately correlated and thus, MANOVA was verified.

Operational hypothesis 1.3: The mean score for learning interest for the Ca condition will be significantly higher than the Ia condition, when a pedagogical agent is presented.

To investigate the role of pedagogical agent on learning interest as planned (Hypothesis

51 1.3), a one way between-groups multivariate analysis of variance was performed. Five dependent variables were used in this analysis: feeling-related interest, value-related interest, arousal, involvement, and attention. The use of one way between-groups MANOVA in this study was justified because of the same reason described in hypothesis 1.1 and 1.2. A one way between- groups MANOVA was performed by comparing the companion agent conditions (condition 1 and condition 5) and the instructor agent conditions (condition 2 and condition 6).

Achievement

The achievement was measured in two levels: 1) recall test and 2) comprehension test.

Operational hypothesis 2.1: The mean score for achievement test between the Sg condition and the Ng condition will be significantly different.

Operational hypothesis 2.2: The mean score for achievement test for the Ca and

Ia conditions that contain agents as a source of seductive messages will be

significantly higher than that for the Tx condition that contains seductive messages

in text format and the No condition that contains no seductive messages.

A two-way between-groups analysis of variance was conducted to explore the impact of seductive augmentation on achievement. Achievement was measured in two levels: 1) recall test and 2) comprehension test.

Recall test: A two-way between-groups ANOVA was conducted to examine the effects of seductive graphics and the source of seductive messages on the number of recalled keywords. The two-way ANOVA was selected because the correlation between recall test score and comprehension test score was low. Therefore, the use of a two way ANOVA was justified in this study. Comprehension test: A two-way between-groups ANOVA was conducted to examine the effects of seductive graphics and the source of seductive messages on the comprehension of

52 instructional material. The two-way ANOVA was selected because the correlation between recall test score and comprehension test score was low. Therefore, the use of a two way ANOVA was justified in this study.

Operational hypothesis 2.3: The mean score for achievement test for the Ca condition will be significantly higher than the Ia condition, when a pedagogical agent is presented.

Two one way ANOVA were also conducted to test hypothesis 2.3. Among students who were presented with a pedagogical agent, the recall test score was compared between students who were given an instructional material containing a companion role of agent and students who were given an instructional material containing an instructor role of agent. Also, the comprehension test score was compared between these students.

Attitude

Operational hypothesis 3.1: The mean score for attitude for the Sg condition will be significantly higher than that for the Ng condition.

Operational hypothesis 3.2: The mean score for attitude for the Ca and Ia

conditions that contain agents as a source of seductive messages will be

significantly higher than that for the Tx condition that contains seductive messages

in text format and the No condition that contains no seductive messages.

To investigate the effects of seductive graphics augmentation on students’ attitude towards the instructional material (Hypothesis 3.1 and 3.2), a two way between-groups multivariate analysis of variance (MANOVA) was performed. Attitude was defined as a score of students’ perception toward instructional material in terms of relevance, confidence, and satisfaction. The hypothesis 3.1 was tested using three dependent variables: relevance, confidence,

53 and satisfaction. The independent variables were the use of seductive graphics and the type of seductive messages. The use of a two way MANOVA in this study was justified because it controls the increased risk of Type I error. Also correlation among dependent variables was examined to confirm the issues regarding multicollinearity and singularity among dependent variables. It was found that five dependent variables were moderately correlated and thus, MANOVA was verified.

Operational hypothesis 3.3: The mean score for attitude for the Ca condition will be significantly higher than the Ia condition, when a pedagogical agent is presented.

To investigate the role of pedagogical agent on students’ attitude toward instructional material, (Hypothesis 3.3), a one way between-groups multivariate analysis of variance was performed. Three dependent variables were used in this analysis: relevance, confidence, and satisfaction. The use of one way between-groups MANOVA in this study was justified because of the same reason described in hypothesis 3.1 and 3.2. A one way between-groups MANOVA was performed by comparing companion agent condition (condition 1 and condition 5) and instructor agent condition (condition 2 and condition 6).

54

CHAPTER VI RESULTS

Introduction

This study investigated the effects of seductive augmentation and agent role on learning interest, achievement, and attitude toward the instructional material. Nine research hypotheses were formulated to answer the following three main research concerns: (a) the effects of seductive augmentation and agent role on learning interest; (b) the effects of seductive augmentation and agent role on achievement; and (c) the effects of seductive augmentation and agent role on attitude toward the instructional material. For the purposes of convenience, this chapter will designate the same acronyms described in Table 2.2 to each treatment condition.

In order to test the research hypotheses, the statistical data analyses conducted included factorial ANOVA and factorial MANOVA. Statistical analyses were conducted with the use of SPSS version 12.0 software.

For the purpose of presenting the research results, this chapter is divided into three sections. In the first section, the results of preliminary data analyses prior to the statistical analysis of the dependent measures are reported. Several tests were also conducted to determine whether the assumptions for parametric statistics were upheld. An alpha level of .05 was used for all statistical tests. In the second section, the results of primary data analyses of the dependent variables are reported. Hypotheses 1.1 through 1.3 and 2.1 through 2.3 regarding main effects and interaction effects of the use of seductive graphics and the use of seductive messages were

55 tested based on dependent variables and the results are reported. Hypotheses 3.1 through 3.3 regarding the effect of the role of agent were tested based on dependent variables and the results are reported. In the last section, all results are summarized.

Preliminary Data Analysis

In this section, the equivalence of treatment groups in terms of pre-interest and prior knowledge was verified. A missing value analysis, a case analysis, and a detection of violations of assumptions for the main dependent variables were conducted as the foundation for the primary data analyses.

Group Equivalence Test

This study was designed as a randomized 2 x 4 factorial design. The participants were not randomly selected from the entire undergraduate student population, but were self-selected by being enrolled in the “Computer literacy” course. Also, all participants were randomly assigned into one of the eight treatment conditions based on the sequence of class entry. Random assignment enhances the equivalency of groups prior to administering experimental treatments (Wiersma, 2000). A list of students with a unique number for each student was prepared on paper before they participated in the class. The researcher randomly assigned the first walk-in student into the first condition and second walk-in student into the second condition until the list was exhausted. Therefore, in this randomized 2 x 4 factorial design, it was assumed that the level of prior-knowledge and the level of pre-interest would be spread evenly among the eight conditions, and the effect should be the same for all eight conditions. Internal validity of the study was also supported by the random assignment process.

In order to verify the group equivalence statistically, two tests were conducted. First, a prior knowledge test was conducted to confirm the group equivalence. Then, students’ pre- interest was confirmed based on the results of the individual interest analysis.

56

Prior knowledge: Prior knowledge data was collected from the pilot study. 18students randomly selected participated in the pilot study. The mean score was 3.4 and standard deviation was .92, which was less than one. Therefore it was assumed that the prior knowledge of participants were not different from each other.

Pre interest: The level of learning interest was measured from two aspects of interest in this study. First, participants’ individual interest was measured before the instructional material was implemented. Later, participants’ situational interest was measured again in order to examine the effect of seductive augmentation and pedagogical agents embedded in the instructional material. Therefore, individual interest was considered as a pre-interest that students already possessed prior to the experiment. A one way between-groups multivariate analysis of variance was performed to investigate the difference among the eight conditions regarding individual interest. Two dependent variables were used: feeling related interest and value related interest. The independent variable was eight treatment conditions. There was no significant differences among the eight conditions on the combined dependent variables, Wilks’ Lambda = .926, F (14, 236) = .661, p = .81, η 2 = .04. It indicated that pre-interest level was not significantly different across treatment groups.

Table 4.1 Means and standard deviations for individual interest

The source of seductive messages Companion role of Instructor role of Text only No messages Total agent agent Learning interest measure M SD n M SD n M SD n M SD n M SD n Feeling-related interest Seductive graphic 2.99 .87 15 2.96 .79 16 2.73 .74 16 2.71 .68 17 2.84 .76 64 No seductive graphic 2.93 .48 15 2.96 .63 16 2.92 .61 15 2.80 .91 17 2.90 .67 63 Total 2.96 .69 30 2.96 .70 32 2.82 .68 31 2.75 .79 34 2.87 .72 127 Value-related interest Seductive graphic 3.83 .70 15 3.88 .72 16 3.78 .89 16 4.03 .80 17 3.88 .77 64 No seductive graphic 3.97 .60 15 4.05 .78 16 4.08 .60 15 4.12 .48 17 4.06 .61 63 Total 3.90 .64 30 3.96 .74 32 3.93 .77 31 4.07 .65 34 3.97 .70 127 Note. The possible score range for learning interest is 1-5.

57

Missing data analysis

Missing data analysis was not conducted because there was no missing data collected from the final 127 participants. The online survey system was employed to avoid missing data problem. The online survey didn’t allow participants miss a question item, unless they closed the instrument windows to exit the questionnaire. If participants missed any item and tried to move onto the next screen, a small warning text box appeared on the screen with the message “You have not answered all the questions on this page. Questions requiring a response are marked with an asterisk (*)”

Case analysis

Two main statistical methods for data analyses were used in this study. First MANOVA was used for two dependent variables, learning interest and attitude. Second, ANOVA was used for one dependent variable, achievement. In order to conduct MANOVA, it is necessary to check for univariate outliers for each of the dependent variables separately. Univariate outliers are cases with extreme values with respect to a single variable. It is common to define outliers as cases which are more than plus or minus three standard deviations from the mean of the variable. In this study, ten single variables were measured to construct three main dependent variables as shown in Table 4.2. Therefore, for learning interest and attitude, case analyses were performed to detect both univariate outliers and multivariate outliers. For achievement, case analyses were performed to detect only univariate outliers.

Table 4.2 The list of dependent variables and single variables

Main dependent variable Single variables Learning interest Individual interest Feeling-related interest, value-related interest Situational interest Arousal, involvement, attention Achievement Recall test, Comprehension test Attitude Relevance, Confidence, Satisfaction

58

Learning interest: Identification of univariate outliers: In this study, learning interest was measured by combining two aspects of interest, individual interest and situational interest. Two dependent variables, feeling-related interest and value-related interest, were used to measure individual interest. Three dependent variables, arousal, involvement, and attention, were used to measure situational interest. Therefore case analysis was conducted for a total of five dependent variables in regard to learning interest. First, an initial screening process by inspecting a histogram and boxplot was conducted to examine the extreme outliers graphically. There were no extreme outliers found. Second, z-score transformation was performed to identify univariate outliers as cases with extreme values. When a z score for a specific case was appreciably greater than 3 or less than -3, the case was considered as an outlier. There were no extreme values found in variables such as feeling-related interest, value-related interest, involvement, and attention. However, one univariate outlier was identified in arousal variable based on the z-score inspection. The original average score was 4 (z score = 3.17). The case was included in Ng-Ia (No seductive graphic and Instructor agent) condition. To overcome the negative effect of this case on learning interest outcome, the value was adjusted by truncation process. Truncation of outliers is a process used when values that are too high or low need to be adjusted to a reasonable level to ensure the z value actually equals 3. Therefore, the original value of the case with outlier was adjusted from 4 to 3.89. Identification of multivariate outliers: In this study, five dependent variables were combined to measure the learning interest level. Multivariate outliers as cases with extreme values with respect to five multiple variables were detected by Mahalanobis distance. This analysis was used to identify any cases that have a strange pattern of scores across the five dependent variables. The maximum value for Mahalanobis distance was 14.68 and the critical value with five dependent variables was 20.52. As the maximum value for Mahalanobis distance was less than the critical value, it was safely assumed that there were no substantial multivariate outliers in the dataset for learning interest.

Achievement: Identification of univariate outliers: In this study, achievement was measured separately by conducting two ANOVAs with two dependent variables: recall test and comprehension test. Thus, only univariate outlier inspection was performed for each dependent

59 variable. First, for both recall test and comprehension test, an initial screening process by inspecting the histograms and boxplots was conducted to examine the extreme outliers graphically. There were no extreme outliers found. Second, for both recall test and comprehension test, z-score transformation was performed to identify univariate outliers as cases with extreme values. When a z score for a specific case was appreciably greater than 3 or less than -3, the case was considered as an outlier. There were no extreme values found for both dependent variables

Attitude: Identification of univariate outliers: In this study, three dependent variables, relevance, confidence, and satisfaction, were used to measure students’ attitude toward the instructional material. Therefore, a case analysis was conducted for three variables in regard to attitude. First, an initial screening process by inspecting the histograms and boxplots was conducted to examine the extreme outliers graphically. There were no extreme outliers found. Second, z-score transformation was performed to identify univariate outliers as cases with extreme values. When a z score for a specific case was appreciably greater than 3 or less than -3, the case was considered as an outlier. There were no extreme values found in variables such as relevance and confidence. However, two univariate outliers were identified in the satisfaction variable based on the z-score inspection. The original average scores were 4.83 (z score = 3.55) and 5.0 (z score = 3.78) each. Both cases were included in Sg-Ia (Seductive graphic and Instructor agent) condition. To overcome the negative effect of these cases on the attitude outcome, the values were adjusted by truncation process. Therefore, the original values of both cases with outlier were adjusted to 4.44. Identification of multivariate outliers: Multivariate outliers as cases with extreme values with respect to three multiple variables were detected by Mahalanobis distance. This analysis was used to identify any cases that have a strange pattern of scores across the three dependent variables. The maximum value for Mahalanobis distance was 10.83 and the critical value with three dependent variables was 16.27. As the maximum value for Mahalanobis distance was less than the critical value, it was safely assumed that there were no substantial multivariate outliers in the dataset for learning interest.

60 Tests for the Assumptions for Initially Planned Tests

As the major statistical methods for data analyses of this study were MANOVA and ANOVA, each of the assumptions that should be satisfied for both MANOVA and ANOVA were examined. Tests for MANOVA assumptions were conducted for learning interest and attitude, and tests for ANOVA assumptions were conducted for achievement.

Tests for MANOVA assumptions

Assumption 1 (Sample size): The minimum required number of cases in each cell was five that is same as the number of dependent variables. In this study, the number of cases in each cell was fifteen to seventeen. Therefore the assumption of sample size was satisfied.

Assumption 2 (Independence of observation): The independence of observation assumption refers that each observation or measurement must not be influenced by any other observation. A logical design for the study was conducted to ensure that all observations are independent. This study used non-repeated measures and individual task for treatment groups. There were no interactions allowed among students. Therefore, the assumption of independence was satisfied.

Assumption 3 (Normality): MANOVA test requires that all the dependent variables are multivariate normal (Stevens, 1996). First, visual inspection of the graphical representations of the data and an examination of formal statistical analyses were conducted for each dependent variable. Second, as formal statistical analyses, the Shapiro-Wilk tests that are applicable to the normality test for small sample sizes were used. Results indicated that a few dependent measures of each group were not normally distributed as indicated in Appendix M. However, it was suggested that with large enough sample sizes, the violation of this assumption should not cause any major problem. (Gravetter & Wallnau, 2000; Stevenson, 1996). In addition, the degree of violation was trivial and MANOVA is robust enough to moderate violations of this assumption.

Assumption 4 (Linearity): This assumption refers to the presence of a straight-line relationship between each pair of dependent variables in this study. The linearity assumption was tested by

61 inspecting scatterplots between each pair of variables. The visual inspection of scatterplots did not show any evidence of non-linearity, therefore the linearity assumption was satisfied.

Assumption 5 (Multicollinearity): Multicollinearity occurs when the dependent variables are highly correlated. Correlation and the strength of the correlations among dependent variables were inspected. As all dependent variables were moderately correlated as shown in Table 4.3 and Table 4.4, less than .8, the multicollinearity among dependent variables did not occur. Therefore the multicollinearity assumption was satisfied.

Table 4.3 Correlation coefficients for relations among five measures of learning interest

Value-related Arousal Involvement Attention Measure interest Feeling-related interest .55** .38** .55** .67** Value-related interest -- .25** .51** .37** Arousal -- -- .32** .43** Involvement ------.59** ** p < .01.

Table 4.4 Correlation coefficients for relations among three measures of attitude

Measure Confidence Satisfaction Relevance .28** .63** Confidence -- .20* Satisfaction ** p < .01. * p < .05.

Assumption 6 (homogeneity of variance-covariance matrix): Box’s test was used to test the null hypothesis that the observed covariance matrices of the dependent variables were equal across groups (see Appendix N). As the Box’s tests were not significant, there was no evidence of a violation of the constant covariance matrix assumption.

62 For the specific investigation on any variance or correlation differences across the groups, Levene’s test was conducted. Levene’s test of the assumption of homogeneity of variance for each of the dependent variables resulted in failing to reject decisions for all variables, a result consistent with the assumption that the variances were equal over the groups (see Appendix O). The consistency between these results and the box’s test results supported that there is no evidence of a violation of constant variance-covariance matrix. One of the dependent variables of attitude, satisfaction resulted in not fail to reject and this result is not consistent with the homogeneity assumption that the variances were equal over the groups. Therefore, a more conservative alpha level (alpha = .01) was set for the satisfaction variable rather than the conventional .05 level.

Tests for ANOVA assumptions

For achievement, ANOVA tests were conducted with two dependent variables each: 1) Recall test and 2) Comprehension test.

Assumption 1 (Independence of observation): A logical design of the study was conducted to ensure that all observations are independent using non-repeated measures and individual task for treatment groups. There were no interactions allowed among students. Therefore, the assumption of independence was satisfied.

Assumption 2 (Normal distribution): It was assumed that the populations from which the samples were taken are normally distributed (see Appendix P). Recall test scores of four conditions were found to not be normally distributed. However, it was suggested that with large enough sample sizes, the violation of this assumption should not cause any major problem. (Gravetter & Wallnau, 2000; Stevenson, 1996). In addition, the distribution of scores for each group was inspected using histograms and no serious problem was suspected.

Assumption 3 (Homogeneity of variance): This assumption refers that samples are obtained from populations of equal variances. Levene’s test of the assumption of homogeneity of variance for each of the dependent variables resulted in failing to reject decisions for all variables, a result

63 consistent with the assumption that the variances were equal over the groups as shown in Appendix Q.

Examination of Hypotheses

Descriptive data

The descriptive statistics for all dependent variables are presented in Table 4.5 according to the eight conditions.

Table 4.5 Descriptive statistics for the dependent variables

Conditions Seductive graphic No Seductive graphic Dependent Measures Ca Ia Tx No Ca Ia Tx No variables (n=15) (n=16) (n=16) (n=17) (n=15) (n=16) (n=15) (n=17) Learning Individual interest interest a Feeling-related M 2.99 2.96 2.73 2.71 2.93 2.96 2.92 2.80 interest SD .87 .79 .74 .68 .48 .63 .61 .91 Value-related M 3.83 3.88 3.78 4.03 3.97 4.05 4.08 4.12 interest SD .70 .72 .89 .80 .60 .78 .60 .48 Situational interest Arousal M 2.25 2.16 1.56 1.79 2.12 2.22 1.97 1.74 SD .60 .62 .57 .55 .54 .74 .60 .67 Involvement M 2.83 2.34 2.09 2.44 2.50 2.66 2.27 2.26 SD 1.01 1.03 .84 .66 .87 .81 .53 .94 Attention M 3.49 3.44 2.94 2.84 3.10 3.16 2.86 2.67 SD .66 .60 .46 .56 .59 .57 .49 .76 Achievement Recall test b M 5.47 5.31 4.63 5.06 5.53 5.06 5.00 5.06 SD 1.55 .87 1.26 1.64 1.30 1.12 1.07 1.71 Comprehension M 5.60 5.81 5.13 5.94 5.53 4.81 5.66 5.53 test c SD 1.45 1.68 1.50 1.56 1.06 1.05 1.23 2.07 Attitude d Relevance M 3.93 3.82 3.66 3.28 3.67 3.53 3.71 3.13 SD .63 .76 .35 .52 .87 .54 .56 .60 Confidence M 3.33 3.42 2.93 2.86 3.28 2.99 2.94 2.78 SD .52 .55 .50 .72 .63 .62 .71 .63 satisfaction M 2.79 2.43 2.10 2.18 2.18 2.40 2.30 2.12 SD .88 1.12 .58 .55 .70 .54 .50 .50 Note. a: Possible range for learning interest (1-5)

64 b: Possible score range for recall test: Minimum: 2 , Maximum: 9 c: Possible range for comprehension test (0-10) d: Possible range for attitude (1-5)

The effects of seductive augmentation and agent role on learning interest

In order to determine the effects of seductive graphics and the source of seductive messages on learning interest (Hypothesis 1.1 and 1.2), a factorial multivariate analysis of variance (MANOVA) was performed. The results obtained for the two research hypotheses are reported below.

The effect of seductive augmentation on learning interest: The hypothesis concerning seductive augmentation predicted that students who were presented with seductive graphics would have significantly higher learning interest scores than students who were not presented with seductive graphics. Also, students who were given a pedagogical agent as a source of seductive messages would haves significantly higher learning interest scores than students who were not given a pedagogical agent as a source of seductive messages. Learning interest was measured by combining the scores of individual interest and situational interest. The individual interest score consisted of feeling-related interest and value-related interest. The situational interest score consisted of arousal, involvement, and attention. Therefore, learning interest was analyzed through the factorial MANOVA design, with feeling-related interest, value-related interest, arousal, involvement, and attention as the dependent measures. The descriptive statistics for dependent variables are presented in Table 4.6.

65

Operational hypothesis 1.1: The mean score for learning interest for the Sg condition will be significantly higher than that for the Ng condition.

Operational hypothesis 1.2: The mean score for learning interest for the Ca and

Ia conditions that contain agents as a source of seductive messages will be

significantly higher than that for the Tx condition that contains seductive messages

in text format and the No condition that contains no seductive messages.

Table 4.6 Means and standard deviations for learning interest

The source of seductive messages Companion role of Instructor role of Text only No messages Total agent agent Learning interest measure M SD n M SD n M SD n M SD n M SD n Feeling-related interest Seductive graphic 2.99 .87 15 2.96 .79 16 2.73 .74 16 2.71 .68 17 2.84 .76 64 No seductive graphic 2.93 .48 15 2.96 .63 16 2.92 .61 15 2.80 .91 17 2.90 .67 63 Total 2.96 .69 30 2.96 .70 32 2.82 .68 31 2.75 .79 34 2.87 .72 127 Value-related interest Seductive graphic 3.83 .70 15 3.88 .72 16 3.78 .89 16 4.03 .80 17 3.88 .77 64 No seductive graphic 3.97 .60 15 4.05 .78 16 4.08 .60 15 4.11 .48 17 4.06 .61 63 Total 3.90 .64 30 3.96 .75 32 3.93 .77 31 4.07 .65 34 3.97 .70 127 Arousal Seductive graphic 2.25 .60 15 2.16 .62 16 1.56 .57 16 1.79 .55 17 1.93 .64 64 No seductive graphic 2.12 .54 15 2.22 .74 16 1.97 .60 15 1.74 .67 17 2.01 .65 63 Total 2.19 .57 30 2.19 .68 32 1.76 .61 31 1.76 .60 34 1.97 .64 127 Involvement Seductive graphic 2.83 1.01 15 2.34 1.03 16 2.09 .84 16 2.44 .66 17 2.42 .91 64 No seductive graphic 2.50 .87 15 2.66 .81 16 2.27 .53 15 2.26 .94 17 2.42 .80 63 Total 2.67 .94 30 2.50 .92 32 2.18 .70 31 2.35 .80 34 2.42 .86 127 Attention Seductive graphic 3.49 .66 15 3.44 .60 16 2.94 .46 16 2.84 .56 17 3.17 .63 64 No seductive graphic 3.10 .59 15 3.16 .57 16 2.86 .49 15 2.67 .76 17 2.94 .63 63 Total 3.30 .65 30 3.30 .59 32 2.90 .47 31 2.76 .66 34 3.06 .63 127 Note. The possible score range for learning interest is 1-5.

66 A factorial MANOVA indicated that there was a statistically significant difference of learning interest score between students who studied the instructional material containing seductive graphics and students who studied the instructional material that doesn’t contain seductive graphics, Wilks’ Lambda=.874, F(5,115)=3.33, p<.01, η 2 =.13. Follow-up ANOVA indicated that significant differences occurred in the attention score of learning interest, F(1,119)=4.91, p=.029, η 2 =.04. The attention score for the seductive graphic condition (M = 3.17, SD = .63) was higher significantly than the no seductive graphic condition (M = 2.94, SD = .63). However, this significant difference failed to reach statistical significance using a Bonferroni adjusted alpha level of .01. Bonferroni adjusted alpha level was calculated by dividing the alpha level (.05) by the number of comparisons. In this analysis, the number of comparisons was five as five dependent measures were included in the MANOVA analysis. Thus, Bonferroni adjusted alpha level was set to .01. Therefore, it was concluded that the overall learning interest was significantly different between the seductive graphic condition and non- seductive graphic condition, yet the univariate results indicate that the use of seductive graphics has not been successful. Field (2004) explained the reason of this result is because the multivariate test takes account of the correlation between dependent variables and further suggested using the discriminant analysis to investigate the nature of the relationship among dependent variables. The result of discriminant analysis indicated that all three variables have a large contribution to the first variate as shown in Table 4.7. However, the highest correlation coefficients between the attention variable and the first variate showed that attention is the most important in differentiating the two groups (seductive graphic condition vs. non-seductive graphic condition). Therefore the group separation can be best explained in terms of one underlying dimension. In this study, the dimension was the learning interest itself. Therefore, it was concluded that the use of seductive graphics does influence the underlying dimension of learning interest, especially the attention variable because attention was found to be more important in terms of learning interest than other two variables. Therefore the hypothesis 1.1 was supported by the findings.

67 Table 4.7 Correlation of predictor variables with discriminant functions (Function structure matrix) and standardized discriminant function coefficients

Correlation with Standardized discriminant discriminant functions b function coefficient c

Predictor variable a Function 1 Function 1 Attention .676 1.291 Arousal -.212 -.608 Involvement .003 -.575 Note. a. The feeling-related interest and the value-related interest were not included in the discriminant analysis because those two individual interest components were measured before the experiment. b. The relative contribution of each dependent variable to group separation. c. The relative contribution of each variable to the first variate.

A factorial MANOVA also revealed that there was a statistically significant difference of learning interest score among the four seductive message conditions, Wilks’ Lambda=.709, F(15,318)=2.81, p<.001, η 2 =.11. Follow-up ANOVA indicated that significant differences occurred in arousal, F(3,119)=4.98, p< .01, η 2 =.11, and attention, F(3,119)=7.01, p< .001,

η 2 =.15. To asses pairwise differences among the four levels for the main effect for arousal, the Tukey HSD (Honestly Significant Difference) follow-up procedure (alpha=.05) was used to provide additional family-wise protection. The result indicated that arousal score for the companion agent (M = 2.19, SD = .57) was significantly higher than both the text only condition (M = 1.76, SD = .60) and the no message condition (M = 1.77, SD = .60). Also, arousal score for the instructor agent (M = 2.19, SD = .68) was significantly higher than both the text only condition (M = 1.76, SD = .60) and the no message condition (M = 1.77, SD = .60). For attention, pairwise differences among the four levels for the main effect was assessed using the Tukey HSD follow-up procedure (alpha=.05). The result indicated that the attention score for the companion agent (M = 3.30, SD = .65) was significantly higher than both the text only condition (M = 2.90, SD = .47) and the no message condition (M = 2.76, SD = .66). Also, attention score for the instructor agent (M = 3.30, SD = .59) was significantly higher than both the text only condition (M = 2.90, SD = .47) and the no message condition (M = 2.76, SD = .66). The attention

68 score for text only condition (M = 2.90, SD = .47) was higher than no message condition (M = 2.76, SD = .66) but the difference was not statistically significant. The results are summarized in Table 4.8.

Table 4.8 Multivate and Univariate Analysis of Variance F rations for learning interest

ANOVA a Feeling- Value-

a related related Arousal Involvement Attention MANOVA interest interest Variable F b F c F F F F Seductive graphic (G) 3.33** .21 1.92 .43 .002 4.92 Seductive message 2.81*** .66 .37 4.98** 1.83 7.01*** source (M) G x M .55 .17 .14 1.15 .97 .39 Note. a. MANOVA = multivariate analysis of variance; ANOVA = univariate analysis of variance. b. Multivariate F ratios were generated from Wilks’ Lambda statistic; Multivariate df for seductive graphic = 5,115. Multivariate df for seductive message source = 15,318. c. Univariate df for seductive graphic = 1,119. Univariate df for seductive message source = 3,119. A Bonferroni adjusted alpha level = .01 was used. ** p < .01 *** p < .001.

In order to examine the difference between both agent conditions and the text only condition, a contrast analysis was performed. A contrast of .5 was assigned to the companion agent condition and also to the instructor agent condition. And -1 was assigned to the text only condition. Consistent with the MANOVA analysis, students in the companion agent (M = 3.30, SD = .65) and the instructor agent (M = 3.30, SD = .59) conditions rated the learning interest score as significantly higher than did students in the text only condition (M = 2.90, SD = .47) in terms of arousal score, t(90) = 3.13 , p < .01, and attention score, t(90) = 3.15 , p < .01. Therefore, hypothesis 1.2 was supported by the findings.

69 The effect of the role of agent on learning interest: A one way MANOVA was also conducted to test the hypothesis 1.3.

Operational hypothesis 1.3: The mean score for learning interest for the Ca condition will be significantly higher than the Ia condition, when a pedagogical agent is presented.

Among students who were presented with a pedagogical agent, the learning interest score was compared between students who were given instructional material containing a companion role of agent and students who were given instructional material containing an instructor role of agent. The result of analysis revealed no significant difference between the two conditions, Wilks’ Lambda=.974, F(5,56)=.30, p= .91. The result is presented in Table 4.9.

Table 4.9 Multivate and Univariate Analysis of Variance F rations for learning interest (Comparison between the companion role of agent and the instructor role of agent)

ANOVA a Feeling- Value-

a related related Arousal Involvement Attention MANOVA interest interest Variable F b F c F F F F The role of agent .302 .00 .12 .00 .50 .00 (companion vs. instructor) Note. a. MANOVA = multivariate analysis of variance; ANOVA = univariate analysis of variance. b. Multivariate F ratios were generated from Wilks’ Lambda statistic; Multivariate df = 5,56. c. Univariate df = 1,60.

70 The effects of seductive augmentation and agent role on achievement

In order to determine the effects of seductive graphics and the source of seductive messages on students’ achievement (Hypothesis 2.1 and 2.2), a factorial analysis of variance (ANOVA) was performed. The results obtained for the two research hypotheses are reported below.

The effect of seductive augmentation on achievement: The hypothesis concerning seductive augmentation predicted that students who were presented with seductive graphics would have significantly higher achievement scores than students who were not presented with seductive graphics. Also, students who were given a pedagogical agent as a source of seductive messages would have significantly higher achievement scores than students who were not given a pedagogical agent as a source of seductive messages. Achievement was measured by two separate tests, the recall test and the comprehension test. Therefore, achievement was analyzed through the two separated factorial ANOVA design, with recall test score as a dependent variable for the first ANOVA and the comprehension test score as dependent variable for the second ANOVA. The descriptive statistics for dependent variables are presented in Table 4.10

Operational hypothesis 2.1: The mean score for achievement test between the Sg condition and the Ng condition will be significantly different. . Operational hypothesis 2.2: The mean score for achievement test for the Ca and Ia conditions that contain an agent as a source of seductive messages will be

significantly higher than that for the Tx condition that contains seductive messages in text format and the No condition that contains no seductive messages.

71 Table 4.10 Means and standard deviations for achievement

The source of seductive messages Companion role of Instructor role of Text only No messages Total agent agent Achievement measure M SD n M SD n M SD n M SD n M SD n Recall test a Seductive graphic 5.47 1.55 15 5.31 .87 16 4.63 1.26 16 5.06 1.64 17 5.11 1.37 64 No seductive graphic 5.53 1.30 15 5.06 1.12 16 5.00 1.07 15 5.06 1.71 17 5.16 1.32 63 Total 5.50 1.41 30 5.19 1.00 32 4.81 1.17 31 5.06 1.65 34 5.13 1.34 127 Comprehension test b Seductive graphic 5.60 1.45 15 5.81 1.68 16 5.13 1.50 16 5.94 1.56 17 5.63 1.55 64 No seductive graphic 5.53 1.06 15 4.81 1.05 16 5.67 1.23 15 5.53 2.07 17 5.38 1.44 63 Total 5.57 1.25 30 5.31 1.47 32 5.39 1.38 31 5.74 1.81 34 5.50 1.50 127 Note. a. The possible score range for recall test: Minimum: 2 , Maximum: 9 b. The possible score range for comprehension test is 1-10

Recall test: A 2 (seductive graphic presence vs. seductive graphic absence) x 4 (seductive message from a companion role of agent, seductive message from an instructor role of agent, seductive message as a text, no seductive message) between-groups ANOVA on recall test scores revealed that there was no significant effect for seductive graphics, F(1,119)=.04, p = .84, and no significant effect for the source of seductive messages, F(3,119)=1.37, p = .27. The result is summarized in Table 4.11.

Table 4.11 Two-way analysis of variance for recall test score and comprehension test score

Source df MS F Recall test score Seductive graphic (two levels) 1 .07 .04 Seductive message (four levels) 3 2.50 1.37 Seductive graphic x Seductive message 3 .52 .29 Error 119 1.83 Comprehension test score Seductive graphic (two levels) 1 1.74 .78 Seductive message (four levels) 3 1.16 .52 Seductive graphic x Seductive message 3 3.28 1.46 Error 119 2.24

72

In order to examine the difference of total recalled number of keywords between both agent conditions and text only condition, a contrast analysis was performed. A contrast of .5 was assigned to the companion agent condition and also to the instructor agent condition. And -1 was assigned to the text only condition. The results indicated that students in the agent conditions recalled significantly higher number of total keywords than did students in the text only condition, t(90) = 1.99, p < .05.

Comprehension test: A 2 (seductive graphic presence vs. seductive graphic absence) x 4 (seductive message from a companion role of agent, seductive message from an instructor role of agent, seductive message as a text, no seductive message) between-groups ANOVA on comprehension test scores revealed that there was no significant effect for seductive graphics, F(1,119)=.78, p = .38, and no significant effect for the source of seductive messages, F(3,119)=.52, p = .67. Therefore, hypothesis 2.1 and hypothesis 2.2 were not supported by the findings.

The effect of the role of agent on achievement: A one way ANOVA was also conducted to test the hypothesis 2.3. Among students who were presented with a pedagogical agent, the recall test score was compared between students who were given instructional material containing a companion role of agent and students who were given instructional material containing an instructor role of agent. Also, the comprehension test score was compared between these students

Operational hypothesis 2.3: The mean score for the achievement test for the Ca condition will be significantly higher than the Ia condition, when a pedagogical

agent is presented.

73 The result of analysis revealed no significant difference between the two conditions for the recall test F(1,60)=1.03, p= .32, and comprehension test F(1,60)=.54, p= .47. Therefore, hypothesis 2.3 was not supported by the finding. The result is summarized in Table 4.12.

Table 4.12 One-way analysis of variance for recall test score and comprehension test score (Comparison between the companion role of agent and the instructor role of agent)

Source df MS F Recall test score The role of agent 1 1.51 1.03 (companion vs. instructor) Error 60 1.47 Comprehension test score The role of agent 1 1.00 .54 (companion vs. instructor) Error 60 1.87

The effects of seductive augmentation and agent role on attitude toward instructional material

In order to determine the effects of seductive graphics and the source of seductive messages on students’ attitude toward the instructional material (Hypothesis 3.1 and 3.2), a factorial multivariate analysis of variance (MANOVA) was performed. The results obtained for the two research hypotheses are reported below.

The effect of seductive augmentation on attitude toward instructional material: The hypothesis concerning seductive augmentation predicted that students who were presented with seductive graphics would have significantly higher attitude scores than students who were not presented with seductive graphics. Also students who were given a pedagogical agent as a source of seductive messages would have significantly higher attitude scores than students who were not given a pedagogical agent as a source of seductive messages. Students’ attitude was analyzed through the factorial MANOVA design, with relevance, confidence, and satisfaction as the

74 dependent measures. The descriptive statistics for dependent variables are presented in Table 4.13.

Operational hypothesis 3.1: The mean score for attitude for the Sg condition will be significantly higher than that for the Ng condition.

Operational hypothesis 3.2: The mean score for attitude for the Ca and Ia conditions that contain an agent as a source of seductive messages will be significantly higher than that for the Tx condition that contains seductive messages in text format and the No condition that contains no seductive messages.

Table 4.13 Means and standard deviations for attitude

The source of seductive messages Companion role of Instructor role of Text only No messages Total agent agent Attitude measure M SD n M SD n M SD n M SD n M SD n Relevance Seductive graphic 3.93 .63 15 3.82 .76 16 3.66 .35 16 3.28 .52 17 3.66 .62 64 No seductive graphic 3.67 .87 15 3.53 .54 16 3.71 .56 15 3.13 .60 17 3.50 .68 63 Total 3.80 .75 30 3.67 .67 32 3.68 .46 31 3.21 .56 34 3.58 .65 127 Confidence Seductive graphic 3.33 .52 15 3.42 .55 16 2.93 .50 16 2.86 .72 17 3.13 .62 64 No seductive graphic 3.28 .63 15 2.99 .62 16 2.94 .71 15 2.78 .63 17 2.99 .66 63 Total 3.30 .57 30 3.20 .62 32 2.94 .60 31 2.82 .67 34 3.06 .64 127 Satisfaction Seductive graphic 2.79 .88 15 2.43 1.22 16 2.10 .58 16 2.18 .55 17 2.36 .84 64 No seductive graphic 2.18 .70 15 2.40 .54 16 2.30 .50 15 2.12 .50 17 2.25 .56 63 Total 2.48 .84 30 2.41 .87 32 2.20 .55 31 2.15 .52 34 2.31 .71 127 Note. The possible score range for attitude is 1-5.

A factorial MANOVA indicated that there was no statistically significant difference of students’ attitude score between students who studied the instructional material containing

75 seductive graphics and students who studied the instructional material that does not contain seductive graphics, Wilks’ Lambda=.974, F(3,117)=1.057, p = .37. However, a factorial MANOVA revealed that there was a statistically significant difference of students’ attitude score among four seductive message conditions, Wilks’ Lambda=.782, F(9,285)=3.37, p<.01, η 2 =.08. Follow-up ANOVA indicated that significant differences occurred in relevance, F(3,119)=5.98, p< .01, η 2 =.13, and confidence,

F(3,119)=4.26, p< .01, η 2 =.10. To asses pairwise differences among the four levels for the main effect for relevance, the Tukey HSD follow-up procedure (alpha=.05) was performed. The result indicated that relevance scores for the companion agent condition (M = 3.80, SD = .75), the instructor agent condition (M = 3.67, SD = .67), and the text only condition (M = 3.68, SD = .46) were significantly higher than the no message condition (M = 3.215, SD = .56) respectively. For confidence, pairwise differences among the four levels for the main effect was assessed using the Tukey HSD follow-up procedure (alpha=.05). The result indicated that confidence score for the companion agent condition (M = 3.30, SD = .57) was significantly higher than the no message condition (M = 2.82, SD = .67). The result is presented in Table 4.14.

Table 4.14 Multivate and Univariate Analysis of Variance F rations for attitude

ANOVA a

Relevance Confidence Satisfaction MANOVA a Variable F b F c F F Seductive graphic (G) 1.06 2.24 1.58 1.03 Seductive message source (M) 3.37** 5.98** 4.26** 1.70 G x M 1.07 .50 .87 1.82 Note. a. MANOVA = multivariate analysis of variance; ANOVA = univariate analysis of variance. b. Multivariate F ratios were generated from Wilks’ Lambda statistic; Multivariate df for seductive graphic = 3,117. Multivariate df for seductive message source = 9,285. c. Univariate df for seductive graphic = 1,119. Univariate df for seductive message source = 3,119. A Bonferroni adjusted alpha level = .01 was used. ** p < .01

76

The effect of the role of agent on attitude:

Operational hypothesis 3.3: The mean score for attitude for the Ca condition will be significantly higher than the Ia condition, when a pedagogical agent is presented.

A one way MANOVA was conducted to test hypothesis 3.3. Among students who were presented with a pedagogical agent and seductive graphics, the attitude score was compared between students who were given instructional material containing a companion role of agent and students who were given instructional material containing an instructor role of agent. The results of the analysis revealed no significant difference between the two conditions, Wilks’ Lambda=.957, F(3,27)=.40, p= .75, η 2 =.04. The result is summarized in Table 4.15.

Table 4.15 Multivate and Univariate Analysis of Variance F rations for attitude (Comparison between the companion role of agent and the instructor role of agent)

ANOVA a

Relevance Confidence Satisfaction MANOVA a Variable F b F c F F The role of agent .401 .21 .26 .99 (companion vs. instructor) Note. a. MANOVA = multivariate analysis of variance; ANOVA = univariate analysis of variance. b. Multivariate F ratios were generated from Wilks’ Lambda statistic; Multivariate df = 3,27. c. Univariate df = 1,29.

77 Summary of the Hypotheses Tests

A significant difference was found between students who studied the instructional material containing seductive graphics and students who studied the instructional material that doesn’t contain seductive graphics in terms of learning interest, especially attention. The attention score for the seductive graphic condition was significantly higher than the no seductive graphic condition. Significant differences were also found among four seductive message conditions in terms of learning interest, especially arousal and attention. Both arousal and attention scores for the companion agent condition and instructor agent condition were significantly higher than text only condition and no message condition. There was also significant difference among four seductive message conditions in terms of attitude toward instruction material, especially relevance and confidence. For relevance, no message condition scored significantly lower than other three conditions. The confidence score for the companion agent condition was significantly higher than no message condition. A summary of the results is presented in Table 4.16.

Table 4.16 Summary of hypotheses tests

Learning interest Achievement Attitude Arousal Involvement Attention Recall Comprehension Relevance Confidence Satisfaction H1.1 Sga> Ngb (overall effect)

c e Ca > Tx Ca > Tx f Ca > No Ca > No H1.2 NSD Iad > Tx Ia > Tx Ia > No Ia > No H1.3 NSD NSD NSD H2.1 NSDg NSD H2.2 NSD NSD H2.3 NSD NSD

78 Table 4.17 Continued.

H3.1 NSD NSD NSD Ca > No H3.2 Ia > No Ca > No NSD Tx > No H3.3 NSD NSD NSD Note. a. Sg: seductive graphic condition. b. Ng: No graphic condition. c. Ca: Companion role of agent condition. d. Ia: Instructor role of agent condition, e. Tx: Text message only condition. f. No: No message condition. g. NSD: No significant difference.

79

CHAPTER V DISCUSSION

Introduction

This study investigated the effects of the presence of seductive graphics, the types of seductive messages, and the role of pedagogical agent on learning interest, achievement, and attitude toward instructional material in pedagogical agent supported learning. Learning interest was measured according to the distinction between individual interest and situational interest. Based on the research approach of learning interest suggested by Krapp (1992), two components of individual interest were measured: (a) feeling-related interest and (b) value-related interest. Also, three components of situational interest were measured: (a) arousal, (b) involvement, and (c) attention. Achievement was measured from two tests: (a) recall test, and (b) comprehension test. Three components of students’ attitude toward instructional material were also measured: (a) relevance, (b) confidence, and (c) satisfaction. These components were implemented to find out whether learning interest really mediated the relationship between the independent variables (the use of seductive graphics, the types of seductive messages) and dependent variables (five components of learning interest, two aspects of achievement, and three components of attitude). Further, the relationship between two types of roles of agent and each dependent variable were investigated. The overall findings are discussed in terms of the three main research concerns: (a) the effects of seductive graphics on learning interest, achievement, and attitude, (b) the effects of the type of seductive messages on learning interest, achievement, and attitude, and (c) the effects of

80 two different roles of agent on learning interest, achievement, and attitude. Implications for and suggestions for future research based on these findings are also discussed.

Seductive graphics, learning interest, achievement, and attitude

One main research concern for this study was investigating the effects of seductive graphics on learning interest, achievement, and attitude.

Effects of seductive graphics on learning interest

There was overall significant difference between students who were given seductive graphics and students who were not given seductive graphics in terms of learning interest. This result support the hypothesis 1.1 of this study that the mean score for learning interest for the condition with seductive graphics will be significantly higher than that for the condition without seductive graphics. Although univariate analysis results revealed none of the sub dependent variables were statistically significant, overall effect of MANOVA and discriminant analysis confirmed that the use of seductive graphics in instructional material significantly promoted students’ attention in terms of learning interest. However, the other two components, arousal and involvement, were not significantly different between the two conditions. First of all, it should be noticed that the concept of situational interest consists of two sub constructs, emotional interest and cognitive interest. Situational interest in this study only considered emotional interest. Seductive details /augmentations are mainly implemented to promote students’ emotional interest by means of seductive graphics, text, and messages. On the other hand, cognitive interest is promoted using expository graphics and messages. This is a very important distinction, because seductive graphics and seductive messages were two key treatments in this study. Therefore, situational interest in this study simply means emotional interest promoted by seductive graphics and seductive messages. This study extends the scope of previous studies by testing the effect of seductive graphics designed to promote emotional interest, especially in terms of three aspects: attention, arousal, and involvement. Previous studies on seductive graphics (or illustrations) mainly

81 investigated the effects of seductive graphics (or illustrations) on situational interest by distinguishing emotional interest and cognitive interest (Harp & Mayer, 1997). However, there has been no further investigation of the effects of seductive graphics (or illustrations) on the sub components of emotional interest. The result of this study shows that seductive graphics positively affect students’ level of attention, but fails to affect arousal and involvement. It not only confirms the results of previous research on “seductive detail effect”, especially when seductive detail is of the graphic type, but also further explains why “seductive detail effect” occurs. It can be inferred from the results that students simply paid attention to the instructional material when there were seductive graphics, but were not involved in deeper level of involvement.

Effects of seductive graphics on achievement

The effects on achievement were measured by two aspects: (1) recall test and (2) comprehension test. There were no significant differences between the seductive graphic condition and non-seductive graphic condition in both the recall test and comprehension test. This result fails to confirm the hypothesis 2.1 of this study that the mean score for the achievement test between the Sg condition and the Ng condition will be significantly different. This hypothesis 2.1 was intended to examine whether seductive detail effects still exist when seductive graphics are implemented. It was expected that if seductive detail effect has occurred, students without seductive graphics will show a higher number of recalled keywords and a higher score on the comprehension test. Or if emotional interest promoted by seductive details is activated, students with seductive graphics will show a higher number of recalled keywords and a higher score on the comprehension test. No significant difference in terms of both the recall test and comprehension test indicates that there was no seductive detail effect. The results of this study are consistent with the results of previous research in regard to the effect of seductive graphics added to expository text on recalled main ideas and problem solving. Previous research showed that seductive augmentation effect hurts both recall of main ideas and problem solving transfer of critical information when seductive illustrations are added to expository text (Harp & Mayer, 1997; 1998). Although the scores for the recalled test and comprehension test were almost same between two conditions in this study, it was found that seductive graphics does not facilitate students’ achievement.

82

Effects of seductive graphics on attitude

In regard to the effects of seductive graphics on students’ attitude toward instructional material, there was no statistically significant difference of attitude between students who studied the instructional material containing seductive graphics and students who studied the instructional material that did not contain seductive graphics. This result fails to confirm the hypothesis 3.1 of this study that the mean score for attitude for the Sg condition will be significantly higher than that for the Ng condition. It was found from the hypothesis 1.1 that students who were given seductive graphics showed higher attention level than students who were not given seductive graphics. However, this attention did not continue to build students’ positive attitude toward the instructional material in this study. In other word, using graphics led students to pay attention to the graphics, but it did not facilitate the students’ perception of relevance of the instructional material. In addition, it is notable that the relevance score was quite high for both conditions. The mean score for the seductive graphic condition was 3.66 and non-seductive graphic was 3.50, which are placed between moderately true and mostly true in five choice Likert scale. One possible reason why students in both conditions scored high on relevance can be explained by students’ individual interest. One of the components of individual interest was the value-related interest. The means scores for the seductive graphic condition and the non- seductive graphic condition were 3.88 and 4.6 respectively. Therefore, it can be concluded that students’ level of relevance to the instructional material in this study was determined simply by their level of value-related interest in terms of individual interest. According to Keller’s ARCS motivational model, one of sub components of the relevance factor is motive matching (Keller, 2002). Motive matching is about how and when instruction can be linked to students’ learning style and personal interest. Therefore students feel that they are linked to the instructional content when they have a certain level of personal interest. In this study, students’ personal interest level measured as an individual interest was approximately 7 out of 10 because feeling-related interest score was 3 out of 5 and value related- interest score was 4 out of 5. Thus, it can be inferred that the topic of the instructional material in this study was quite an interesting topic to students and students’ level of relevance to the instructional material was determined by the high level of individual interest.

83

Seductive messages, learning interest, achievement, and attitude

The second main research concern for this study was investigating the effects of different sources of seductive messages on learning interest, achievement, and attitude.

Effects of seductive messages on learning interest

The students who were given seductive messages delivered by a pedagogical agent earned significantly higher scores in arousal and attention in terms of learning interest than the students who were given either the text type of seductive message or no seductive message. Students in the no seductive message condition were not given any type of seductive messages, thus their level of learning interest was lower than the other conditions. An interesting finding is that there was significant difference of the arousal score and attention score between agent delivered seductive message and text based seductive message. The result of contrast analysis confirms that seductive messages are more effective to promote students’ learning interest when they were delivered by a pedagogical agent rather than in text. There are at least two possible explanations for why the text type of seductive message was not as effective as the one delivered by an agent. First, the presence of an agent could affect students’ learning interest. Previous research on the effects of pedagogical agent showed that users’ interactions with the computer are much smoother when likable animated pedagogical agents are present. They also help students develop an emotional connection with the agent and facilitate their enjoyment of the learning situation (Dehn & Mulken, 2000). Therefore, although identical seductive messages were presented to both agent conditions and text only condition, students with agents could have been more interested in the learning of instructional material. A second reason is that the source of the messages was different in this study. In other word, seductive messages delivered by a pedagogical agent were audio messages via agent voice that is very different from text type of message. Therefore it is also possible that students perceived a sound type of seductive message more positively then the text type of message. The

84 second reason was not confirmed, because the sound only condition was not implemented in this study.

Effects of seductive messages on achievement

It was found that there was no significant difference of recall test score and comprehension score in terms of the achievement among four different sources of seductive message conditions. The reason why these scores were not different can be explained by the “seductive details effect”. As for the recall test, previous research on the seductive details shows that adding interesting text that is irrelevant to the theme of a descriptive or narrative passage either reduces or doesn’t facilitate students’ remembering of the main idea in the passage (Garner et al., 1992; Garner et al., 1989; Hidi & Baird, 1988; Mohr, Glover, & Ronning, 1984; Shirey, 1992; Shirey & Reynolds, 1988; Wade, 1992; Wade & Adams, 1990). In this study, the mean score for text only and no message condition was 4.81 and 5.06 respectively. Although the result was not significantly different, the difference between mean scores shows that a seductive details effect occurred. Regarding comprehension test, Harp and Mayer (1997; 1998) found that students scored highest when only basic text was presented. Consistent with their study, the result of the comprehension study in this experiment also found that students who were in the no seductive message condition scored highest among the four message conditions although the difference was not statistically significant. However, it should be highlighted from the contrast analysis that students who were in both agent conditions scored higher than students in text only condition when the total number of keywords is considered. It means that students in both agent conditions actually recalled more keywords than the students in the text only condition. It can be inferred from this result that if the messages delivered by agents were not seductive in their content, the result might be different. It is because students actually recalled more information from an agent than they did from text although the messages were identical in all conditions. It emphasizes the importance of message design especially when pedagogical agents are embedded into computer based learning.

85 Effects of seductive messages on attitude

There were significant differences in relevance scores and confidence scores among the four message conditions. When seductive messages were presented, students perceived a higher relevance toward instructional material than did students in the no seductive message condition. However, students perceived higher confidence toward instructional material when only a companion role of agent was presented than the control condition where no seductive message was delivered. With respect to relevance, the reason students perceived higher score when seductive messages were presented then no seductive message was presented can be explained by the content of the seductive messages. Although the content of seductive messages was related to the main theme, it contained unimportant information to understand the main topic. However, students were exposed to a couple of examples that illustrated some stories and cases about intellectual property. According to Keller (1983, 2002), relevance is referred as value where personal worth and future value are immersed. Therefore, it is possible that students felt connected to the content of the seductive messages when seductive messages stimulated students’ perception of value by providing comment, anecdotes, or examples about intellectual property. Confidence scores were significantly different only between the companion role of agent and the control condition where no seductive messages were given. However, no significant difference was found between instructor role of agent and control condition. Confidence is defined as a positive expectation and support for success by enhancing students’ belief in their competence (Keller, 1983, 2002). The fact that students recognized a companion agent and an instructor agent differently in this study implies that students felt more competence with a companion agent, because delivered messages and voice of the companion agent were consistent with students’ perception of an agent as a companion. However, it also should be noted that the difference on confidence scores between the companion agent and the instructor agent was minimal.

86 Pedagogical agents, learning interest, achievement, and attitude

In regard to the difference between the companion role of agent and the instructor role of agent, there were no significant differences in terms of learning interest, achievement, and attitude. This result fails to support the hypothesis 1.3, 2.3, and 3.3 of this study, that the companion role of agent would be more effective in increasing learning interest, achievement, and further attitude toward the instructional material. This result is not consistent with those of earlier studies on the role of pedagogical agents which have found that agent perception and learning outcomes are different depending on what role of agent is implemented (Baylor, 2003). One possible reason why no significant difference was found between the roles of the pedagogical agent is that students did not recognize the agent persona of each agent. An agent has its character or personality based on students’ perceived persona combining image, animation, voice, and emotional expressions. In this study, although image, voice, and emotional expression were different between two agents, the messages delivered were identical each other. On the other hand, the study conducted by Baylor (2003) used different types of messages for each different role of agent. For example, in terms of messages, the expert agent provided information whereas the motivational agent provided encouragement and support, and the mentor agent provided both information and encouragement/support. Therefore, differentiating the delivered messages plays an important role for students to acknowledge the role of agent. However, identical seductive messages were employed to control the effect of seductive messages in this study. It led research to find no differences in regard to the learning interest, achievement, and attitude between the companion role of agent and the instructor role of agent.

Limitation of the Study

There are several limitations to be considered in regard to this study.

87 First, as described in the definition of learning interest, it is theoretically possible to distinguish individual interest from situational interest. However, one should be cautious because it is not an easy to differentiate situational interest and individual interest practically. Therefore, this research considered individual interest and situational interest together when the learning interest dependent variables were analyzed. Yet, the results of this study need to be interpreted cautiously because there are still possibilities that individual interest could affect the level of situational interest. Second, situational interest was considered from only one dimension of it. By definition, situational interest is affected by two sub components, cognitive interest component and emotional interest component. However, this study only implemented seductive augmentation to promote emotional interest. Therefore it is unknown whether there were any effects of cognitive interest on situational interest. Third, the effects of pedagogical agents were not fully investigated because there was no sound-only condition that the text only condition needed to be compared with. Also, the gender of agents used in this study was male. Therefore, follow up study is needed to investigate the effects of agents in various ways. Fourth, recall test was conducted by asking students “Please type in every keyword you can remember from this material.” Previous studies using recall test as a variable asked students to type in everything they can remember (Harper & Mayer, 1997; 1998). Therefore, the “keyword” in the recall test question could hinder students from typing in everything they remember, which caused no significant differences among conditions. Fifth, the instructional material used in this study was lecture based. There were limited interactions available between a student and the instructional material. Therefore, it is necessary to conduct a follow up study with an interactive instructional module that students can actively interact with. Sixth, although two different roles of agent were validated from the previous research (Baylor, 2003), these roles were not validated for this study. Therefore, it is possible that the role of agent might not have been perceived as being planned. In other word, students might have perceived two different agents as not being different. Lastly, it should be noted that the time taken to complete this study was 30 to 40 minutes depending on the condition. Comparing the previous studies that mainly focused on text based

88 instructional material, this study could be considered to have taken reasonable time. Most of the previous studies took less than 10 minutes. The study conducted by Schraw (1998) took the longest time, 12.5 minutes. Although the time taken in this study was relatively longer than other previous seductive augmentation studies, it would be important to have longer treatment times to ensure the treatment effect.

Implications

The findings of this study have both theoretical and practical implications. On the theoretical side, the results of this study illuminate the concept of learning interest as it concerns seductive augmentation. Especially, unlike the previous studies, the result of this study shows how seductive augmentation could affect three sub components of situational interest: attention, arousal, and involvement. In addition, this study generalized the use of seductive augmentation in a multimedia learning context where a pedagogical agent is presented. Considering that most of previous research has been conducted with paper based instructional materials, this study investigated whether the “seductive detail/augmentation effect” is still evident in the multimedia learning environment where a pedagogical agent is present. Finally, one of the findings was that there is different effectiveness of seductive augmentation when sources of messages vary. In particular, same seductive messages can have different effects when they are delivered by a pedagogical agent. On the practical side, the important finding is the strategies employed to promote learner’s interest in learning. It is obvious that adding humorous graphics or interesting graphics does not affect students’ involvement with the instructional material. Instead, there should be a balance between the use of seductive augmentation and the contents of instructional material so that students’ can promote cognitive interest as well as emotional interest. Only enhancing emotional interest does not help multimedia designers get students engaged with the learning. A second implication would be an importance of message design for pedagogical agents. It was found that students tend to believe and remember more when the message is delivered by agents. Therefore pedagogical agents need to be designed cautiously. Also the role of agent should be defined clearly before creating them.

89

Further Research

Future research needs to move in four directions. First, as this study mainly focused on the effects of emotional interest, it is still necessary to consider the other component of situational interest that is cognitive interest. It has been suggested that cognitive interest adjuncts such as explanative summary, illustrations or messages influence students’ cognition by promoting structural understanding of the expository passage. And the attainment of structural understanding promoted a sense of positive affect about the passage. Therefore situational interest is affected by cognitive interest as well as emotional interest. This line of study will provide researchers with a big picture of how situational interest is promoted from two aspects, emotional interest and cognitive interest. Second, one of the assumptions in this study was that students will perceive the role of agents differently. Although no difference was found, many studies observed agent characteristics such as gender and age affect differently on the perception of students. Especially, the effects of messages were found to be very important in this study when they are delivered by agents. Therefore, future study is needed to examine how the cognitive interest messages or emotional interest messages affect students’ perceived situational interest when these messages are delivered by agents. Third, the contents of this study were mainly about intellectual property that was a type of expository information. Future study should extend the scope of material content to other types of text information such as math or science content where students actually need to apply what they have learned from the instructional material. Also, the narrative type of text needs to be tested. Previous research shows that the effects of seductive details/augmentation are different in opposite way depending on whether the context is expository information or narrative biography text. When a narrative biography text was used as a main text, none of previous studies showed the seductive-augmentation effect. In fact, the significant comparison showed that seductive augmentations helped learning of main ideas (Garner & Gillingham, 1991; Schraw, 1998).

90 Fourth, individual difference needs to be considered in future research. If students’ pre interest level is different at the beginning of instruction, the use of seductive details might be an effective strategy to improve students’ situational interest so that students will pay attention to the instructional material. But if students are already highly interested in the instructional material, there will be no need to improve students’ situational interest.

Conclusion

This study investigated the effects of the presence of seductive graphics, the types of seductive messages, and the role of pedagogical agents on learning interest, achievement, and attitude toward instructional material in pedagogical agent supported learning. The results indicated that the use of seductive graphics and seductive messages was effective to improve learner’s attention to the learning material in terms of learning interest, and also learner’s attitude especially, relevance scores. However, no significant difference was found for recall test and comprehension test. Two different roles of agent didn’t make any difference in all three dependent variables. The main reason this study failed to find the effects of seductive augmentation on achievement can be explained by he lack of cognitive interest stimulus. In order to fully promote situational interest, it is necessary to consider cognitive interest as well as emotional interest. Therefore future research is recommended to examine the effects of cognitive interest stimulus and emotional interest stimulus when they are presented separately and also when they are presented together. Although this study showed the “seductive details effect” as in previous studies, the findings of this study are expected to contribute on the issue of using seductive augmentation in agent supported learning. First, the difference in affects of seductive augmentation on students’ interest, specifically on arousal, involvement, and attention were empirically tested in this study. The data shows that students paid attention to the seductive augmentation, but did not get deeply involved in it. Second, this study showed the importance of message design for pedagogical agents. It was found that students tend to believe and remember more content when the message is delivered by agents. Finally, while most previous research on seductive augmentation was conducted with paper-based material, this study extended the type of material to a multimedia

91 learning environment with the presence of a pedagogical agent. Improving learning interest for successful learning is not a simple process. Many psychological constructs are involved from an individual aspect and also a situational aspect. As an instructional designer, therefore, it is necessary to approach the issue of learning interest from both of these aspects.

92

APPENDIX A INSTRUCTIONAL MATERIAL TEXT

93 Introduction to Intellectual Property

The purpose of this material is to give a quick and basic overview of Intellectual Property, focusing on the procurement of a United States. After reading this material, (1) you will be able to tell what intellectual property is, (2) what a patent is, (3) what a trademark is, and (4) what a copyright is.

What is an intellectual property (IP)? Intellectual property allows people to own their creativity and innovation in the same way that they can own physical property. The owner of intellectual property can control and be rewarded for its use, and this encourages further innovation and creativity to the benefit of us all. In some cases intellectual property gives rise to protection for ideas but in other areas there will have to be more elaboration of an idea before protection can arise. It will often not be possible to protect intellectual property and gain intellectual property rights unless they have been applied for and granted, but some intellectual property protection such as copyright arises automatically, without any registration, as soon as there is a record in some form of what has been created.

There are three main types of intellectual property as follows:

1. Patents for inventions - new and improved products and processes that are capable of industrial application 2. Trademarks for brand identity - of goods and services allowing distinctions to be made between different traders 3. Copyrights for material - literary and artistic material, music, films, sound recordings and broadcasts, including software and multimedia

However, Intellectual property is much broader than this extending to trade secrets, plant varieties, geographical indications, performers rights and so on. To understand exactly what can be protected by intellectual property, you will need to check the three main areas of copyright, patents, and trade marks as well as other intellectual property. Often, more than one type of intellectual property may apply to the same creation.

What is a patent? A patent as one of three main types of intellectual property gives an inventor the right for a limited period to stop others from making, using or selling an invention without the permission of the inventor. It is a deal between an inventor and the state in which the inventor is allowed a short term monopoly in return for allowing the invention to be made public. There are some special features of a patent in terms of its scopes. In addition, specific conditions must be fulfilled to get a patent. Patents as one of three main types of intellectual property are about functional and technical aspects of products and processes. Most patents are for incremental improvements in known technology - evolution rather than revolution. The technology does not have to be complex. Patent rights are territorial; a US patent does not give rights outside of the US. Patent rights last for up to 20 years in the US. A patent can be of value to an inventor - as well as protecting his business, patents can be bought, sold, mortgaged, or licensed to others. They also benefit people other than the

94 inventor since large amounts of information can be learnt from other people’s patents - they can stop you from reinventing things or you can monitor what your competitors are doing. Patents also spur you or others on to develop your idea further, and once the term of the patent expires it can be freely performed by anyone who benefits the public and the economy. In order to get a patent as one of three main types of intellectual property specific conditions must be met. That is, the invention must……..

• Be new. The invention must not form part of the "state of the art". The state of the art is everything that has been made available to the public before the date of applying for the patent. This includes published documents and articles, but can also include use, display, spoken description, or any other way in which information is made available to the public. • Involve an inventive step. As well as being new, the invention must not be obvious from the state of the art. Obviousness is from the viewpoint of a person skilled in the area of technology that the invention is in. • Be industrially applicable. This condition requires that the invention can be made or used in any kind of industry.

What is a trade mark? A trade mark as one of three main types of intellectual property is any sign which can distinguish the goods and services of one trader from those of another. A sign includes words, logos, colors, slogans, three-dimensional shapes and sometimes sounds and gestures. Examples of trademark include ‘aspirin’, ‘cellophane’, and ‘escalator’.

A trade mark is therefore a "badge" of trade origin. It is used as a marketing tool so that customers can recognize the product of a particular trader. You don't necessarily have to register your trademark to be protected. However, federal registration has several advantages. If you register you trademark, one of three main types of intellectual property federal registration offers several advantages. They include notice to the public of the registrant's claim of ownership of the mark, a legal presumption of ownership nationwide, and the exclusive right to use the mark on or in connection with the goods or services set forth in the registration.

What is a copyright? A copyright as one of three main types of intellectual property gives the creators of a wide range of material, such as literature, art, music, sound recordings, films and broadcasts, economic rights enabling them to control use of their material in a number of ways, such as by making copies, issuing copies to the public, performing in public, broadcasting and use on-line. It also gives moral rights to be identified as the creator of certain kinds of material, and to object to distortion or mutilation of it. (Material protected by copyright is termed a "work".) However, copyright does not protect ideas, or such things as names or titles. Like other types of intellectual property, copyright has its own purpose. The purpose of copyright is to allow creators to gain economic rewards for their efforts and so encourage future creativity and the development of new material which benefits us all. Copyright material is usually the result of creative skill and/or significant labor and/or investment, and without protection, it would often be very easy for others to exploit material without paying the creator. Most uses of copyright material therefore require permission from the copyright owner.

95 However there are exceptions to copyright, so that some minor uses may not infringe copyright. Like other types of intellectual property copyright protection could automatic as soon as there is a record in any form of the material that has been created, and there is no official registration or form or fee. But creators can take certain steps to help prove that material is theirs.

96

APPENDIX B PRE-INTEREST INSTRUMENT

97

Instruction: Each of the statements in the following describes feelings or mood you expect while studying the instructional material on “Intellectual Property.” Please choose the description that best describes you.

true true Very true Very Moderately Moderately Mostly true Slightly true true Slightly Not at all true

1 2 3 4 5

I feel bored when I think of an intellectual property

I feel stimulated when I think of an intellectual property

I feel interested when I think of an intellectual property

I feel indifferent when I think of an intellectual property

I feel engaged when I think of an intellectual property

I feel studying intellectual property is meaningful.

I feel studying intellectual property is unimportant.

I feel studying intellectual property is useful.

I feel studying intellectual property is worthless.

98

APPENDIX C

POST INTEREST INSTRUMENT

99

Instruction: Each of the statements in the following describes your feeling while you were studying the instructional material on “Intellectual Property.” Please choose the description that best describes you.

true true Very true Very Moderately Mostly true Mostly Slightly true Slightly Not at all true

1 2 3 4 5

I felt active at the moment while I was studying.

I felt energetic at the moment while I was studying.

I felt vigorous at the moment while I was studying.

I felt lively at the moment while I was studying.

I felt full-of-peep at the moment while I was studying.

I was completely caught up in what I was studying.

I was concentrated while I was studying.

I found something interesting at the beginning of this learning material that got my attention.

I felt the quality of the writing of learning material helped to hold my attention.

I felt the contents in this learning material were eye-catching.

I felt the learning material was so abstract that it was hard to keep my attention on it.

I felt the pages look dry and unappealing.

I felt the way the information is arranged on the screens helped keep my attention.

100 I felt the learning material has things that stimulated my curiosity.

I learned some things that were surprising or unexpected in this learning material.

I found the variety of reading passages, exercises, illustrations, etc., helped keep my attention on the learning material.

I found the style of writing of the learning material is boring.

I found much irritating information on each screen of the learning material.

The amount of repetition in the learning material caused me to get bored sometimes.

101

APPENDIX D

RECALL TEST

102

Instruction: Please type in every keyword you can remember from the passage you read on Intellectual Property.

103

APPENDIX E ACHIEVEMENT TEST ITEMS

104

Instruction: In the following section, seven statements are listed. Choose a "True" if you think the statement is correct. Choose a "False" if you think the statement is incorrect.

Copyright arises automatically without any registration. True False Patents can be bought, sold, mortgaged, or licensed to others. True False A US patent does not give rights outside of the US. True False A trademark is any sign that includes words, logos, colors, slogans, three- True False dimensional shapes, but not sounds and gestures. People don’t necessarily have to register their trademark to be protected. True False Copyright does not protect things as names or titles. True False

Instruction: Please answer the following questions.

• Please choose an incorrect statement about copyright in the following statements. o A copyright gives moral rights to identify the creator of certain kinds of material and to object to distortion or mutilation of it. o A copyright allows creators to gain economic rewards for their efforts and so encourage future creativity. o A copyright gives the creators of a wide range of material economic rights enabling them to control use of their material in a number of ways. o All uses of copyright material require permission from the copyright owner with no exception.

• Which one of the followings is not considered as a condition of patent? o In order to get a patent, the invention must be new. o In order to get a patent, the invention must involve an inventive step. o In order to get a patent, the invention must use complex technology. o In order to get a patent, the invention must be industrially applicable.

• Which one of the followings is not an advantage of registering trademark? o Notice to the public of the registrant's claim of ownership of the mark. o A legal presumption of ownership nationwide. o The exclusive right to use the mark on or in connection with the goods or services set forth in the registration. o Economic rewards for creator’s efforts.

• Please describe in writing how the concepts of intellectual property, copyright, trademark, and patent are related each other.

105

APPENDIX F

ATTITUDE INSTRUMENT

106 Instruction: There are 24 statements in this questionnaire. Please think about each statement in relation to the learning materials you have just studied and indicate how true it is. Think about each statement by itself and indicate how true it is. Do not be influenced by your answers to other statements.

true Very true true Very Moderately Moderately Mostly true Slightly true Not at all trueNot at all

1 2 3 4 5

When I first looked at this learning material, I had the impression that it would be easy for me.

This learning material was more difficult to understand than I would like for it to be.

After reading the introductory information, I felt confident that I knew what I was supposed to learn from this learning material.

Completing the tests on this learning material gave me a satisfying feeling of accomplishment.

I found some stories, pictures, or examples that showed me how this learning material could be important to some people.

It is clear to me how the content of this learning material is related to things I already know.

Completing this learning material successfully was important to me.

Many of the screens in this learning material had so much information that it was hard to pick out and remember the important points.

As I worked on this learning material, I was confident that I could learn the content.

I enjoyed this learning material so much that I would like to know more about this topic.

107 The content of this learning material was relevant to my interests.

There are explanations or examples of how people use the knowledge in this learning material.

The exercises in this learning material were too difficult.

I really enjoyed studying this learning material.

The content and style of writing in this learning material convey the impression that it was worth knowing.

After working on this learning material for awhile, I was confident that I would be able to pass a test on it.

This learning material was not relevant to my needs because I already knew most of it.

The wording of feedback after the exercises, or of other comments in this learning material, helped me feel rewarded for my effort.

I could relate the content of this learning material to things I have seen, done, or thought about in my own life.

It felt good to successfully complete this learning material.

The content of this learning material will be useful to me.

I could not really understand quite a bit of the content in this learning material.

The good organization of the content helped me be confident that I would learn this learning material.

It was a pleasure to work on such a well-designed learning material.

108

APPENDIX G ACHIEVEMENT TEST ITEMS ANALYSIS

109 Item Item Item type Item class Information used to develop test items Note number difficulty a 1 True / Copyright Copyright protection could automatic as soon as there is a record .21 False in any form of the material that has been created, and there is no official registration or form or fee. 2 True / Patent A patent as one of three main types of intellectual property gives .94 Dropped False an inventor the right for a limited period to stop others from making, using or selling an invention without the permission of the inventor. 3 True / Patent Patents can be bought, sold, mortgaged, or licensed to others .63 False 4 True / Patent Patent rights are territorial; a US patent does not give rights .53 False outside of the US. 5 True / Trademark A sign includes words, logos, colors, slogans, three-dimensional .37 False shapes and sometimes sounds and gestures. Examples of trademark include ‘aspirin’, ‘cellophane’, and ‘escalator’ 6 True / Trademark You don't necessarily have to register your trademark to be .47 False protected. 7 True / Copyright A copyright does not protect ideas, or such things as names or .53 False titles. 8 Multiple Copyright Most uses of copyright material therefore require permission .32 choice from the copyright owner. However there are exceptions to copyright, so that some minor uses may not infringe copyright. 9 Multiple Patent • Be new. The invention must not form part of the "state of .68 choice the art". • Involve an inventive step. As well as being new, the invention must not be obvious from the state of the art. • Be industrially applicable. This condition requires that • the invention can be made or used in any kind of industry.

110 10 Multiple Trademark If you register you trademark, one of three main types of .37 choice intellectual property federal registration offers several advantages. They include notice to the public of the registrant's claim of ownership of the mark, a legal presumption of ownership nationwide, and the exclusive right to use the mark on or in connection with the goods or services set forth in the registration. 11 open- Intellectual There are three main types of intellectual property as follows: ended property - Patents for inventions - new and improved products and processes that are capable of industrial application - Trade marks for brand identity - of goods and services allowing distinctions to be made between different traders - Copyright for material - literary and artistic material, music, films, sound recordings and broadcasts, including software and multimedia

Note. a. The percentage of students who answer an item correctly; average test difficulty was .46.

111

APPENDIX H AGENT SCRIPTS

112

Condition Webpage ID Script Condition 1 Con1main Hi! My name is Mike (or Dr. Hendriks). Thanks so much for attending this course. I’m your guide to teach you about 70 words intellectual property. We’re going to study this together.

First of all, I’d like to know about you little more. There are a couple of questions you need to answer for me in next screen. Please be serious about answering them so that I can prepare my instruction properly for you.

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button. Con1mainp2 (Background information) Con1mainp3 Thank you! Now I’m ready to talk about “intellectual property”. 12 words Are you?

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button. Con1mainp4 I’m guessing that you probably don’t know much about what intellectual property is. That’s why I’m here with you. I’m 38 words going to explain what intellectual property is and how it impacts you in school and in your life.

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button. Con1mainp5 The Annual Consumer Electronic Show in Las Vegas is an internationally known show. Just last January over 120,000 people from over 110 countries attended the show to see the latest product innovations from the world's leading companies! 152 words That’s amazing…Don’t you think? It must be important. Well, without intellectual property protection, many of these companies would not invest billions every year to launch new waves of products. Why is that? Without the protection of the patents covering the iPod, Apple probably wouldn’t have launched the product in the first place. Why launch it if it’s going to be stolen right away? And even if they did launch it, they certainly would not have developed the iPod mini after the first version had been ripped off by counterfeiters. You see my point ? If Apple hadn’t had the patent on the iPod anyone could have taken the idea and made money off of it.

- If you want to listen to me again, click on the button.

- If you want to proceed to the next screen, click on the < Next

113 > button.

Con1mainp6 I have one more story about intellectual property. There’s this new medical student, Todd, who invents a process for inserting 110 words modified genes into cancer cells. He applies for and receives a patent. A patent is basically a monopoly awarded to Todd by the federal government. It allows him to require anyone who wants to use the process to pay him a negotiated license fee. The problem is, if no one wants to use the process, Todd won’t make any money off it. If that happens, he’ll probably just use it in his own gene therapy clinic. It’ll be useful – but WON’T make him the money. Well.. That’s life, ya know.

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button. Con1mainp7 Here’s something interesting… Do you drive your own car? You’ve probably never wondered who invented “power steering 169 words valves”? Well, let me tell you. Bishop power steering valves are used in about one fifth, or 20%, of the world's cars. A.E. Bishop & Associates, the company that makes the valves, receives a royalty payment of up to $1.00 for each unit someone manufactures and generates its financial return through effective use of intellectual property. The story goes that Arthur Bishop began developing and patenting automotive power steering systems after World War II. His company now has over 300 patents worldwide AND it makes about $5 million a year in royalties. This means that his company doesn’t manufacture any products itself but licenses car and component manufactures and specialist manufactures of machinery around the world to use its designs. Obviously they make sure their patents are vigorously defended. And I can tell you that they prosecute many infringers. Infringers are people who try to use the valve without paying the royalty.

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button. Con1mainp8 Sometimes we’renot sure who’s as inventor and who is not, especially in a when people work together and create something 134 words together. Here’s a perfect example of that kind of situation. Tom Haberfield and Bonnie Rand jointly conceived of and designed a miniature EEG machine (a machine that measures brain waves). This machine allows its wearer to monitor his or her own brain waves by using a wrist watch-like device. To make sure the concept was viable, Tom and Bonnie got their engineer friend, Clark Bromsky, to build a test model according to their specifications. In this case, Bonnie and Tom would be called

114 “co-inventors” of the invention because they were the sole creative contributors to the invention’s structure. Clark wouldn’t be an inventor at all. He just did what they told him to do

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button. Con1mainp9 Here’s another example. The Cochlear bionic ear implant is a device designed to help the hearing impaired and the profoundly 189 words deaf can’t benefit from traditional hearing aids. This device is used by over 10,000 people world wide. It consists of two parts. The first part is a head set made up of a 22-channel stimulator in a titanium capsule with platinum electrodes. This stimulator is implanted inside the skull behind the ear, close to residual nerve fibers. The other part is a pocket speech processor.

The cochlear bionic ear implant was developed over 10 years by a team from the University of Melbourne lead by a man named Dr. Graeme Clarke. This group and the University of Melbourne, called the Cochlear Party Limited, has over 180 patents in 7 countries that protect key features of the implant. The University of Melbourne and the Commonwealth also hold patents covering early research that has been licensed to Cochlear. Manufacturing of the Cochlear implant is done entirely in Australia with exports making up 95% of their sales. The Cochlear Party Limited has earned over $8 million in royalties from their intellectual Property.

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button. Con1mainp10 This story happened in 1995. Qualitex Corporation used a green-gold pad that they designed to use in dry cleaning presses. 108 words The green-gold color wasn’t associated with dry cleaning pads as such, had no functional purpose. It operated only to identify the pads as originating with Qualitex. Once these facts were established, the U. S. Supreme Court saw no reason why the color couldn’t qualify as a trademark as long as it could be shown that consumers relied on the color to identify the source of the pads. So, in 1995, the U.S. supreme court ruled that a single color – green –could function as a trademark for ironing pad.

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button.

115 Con1mainp11 Do you know how the trademark “Mercedes” was born? In 1897, Austrian businessman Emil Jellinek, traveled from his 110 words home in Nice, France to buy a car from the Daimler factory in Cannstatt, Germany. When he returned to his home on the French Riviera, his sporting Daimler Phoenix car caused such a sensation that he decided to enter it into a local touring competition. He entered it under the name of "Mercedes" after his favorite 9 year old daughter. Jellinek realized the business potential for the new car, so he placed an order for 36 more and also secured the franchise for to sell them in several countries. Gottlieb Daimler also agreed to have them sold under the name of "Mercedes."

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button. Con1mainp12 People connected to the Internet are already exchanging digital files containing pictures, music, video clips, you name it. 110 words Record companies are afraid that people will simply download their music, instead of buying CDs. That could put the CD manufacturers out of business. According to the US Court of Appeals in 1999, it’s OK to record music on your personal computer. And, you can also rip the music, convert to MP3 and burn it on a CD for YOU, but if you give a copy to anyone else…. well, it’s going to cost you. In fact, if you’re caught for copyright infringement of music, the average fine PER SONG is $750!

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button. Con1mainp13 One of the issues about copyright in schools is plagiarism. As more internet resources have been greatly increased, the 114 words temptation to plagiarize has greatly increased. You can steal articles from a huge variety of sources. Or you can even find perfect web materials such as word or PowerPoint for your class assignment. Well you’d better be careful before you’re sorry….. According to the copy right law, a copyright owner has the right to sue you and collect between $2500 and $25,000 from you for your FIRST copyright violation! Now, THAT’S A LOT OF MONEY! Even if you made some cosmetic changes to the PowerPoint slides from the internet, you still have violated copyright law.

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button. Con1mainp14 Do you know what the law is in Florida about copying software, even if you’re only going to use it yourself? Florida Law says 126 words that if you copy software, you face jail and fines. So, what’s the penalty for copying software? In Florida, you could get a

116 Misdemeanor of the First Degree, go to prison for a year and be fined $1000 or more. OR You could get a Felony of the Second Degree where you can go to jail for up to 15 years and be fined at least $10,000. OR you could get a Felony of the Third Degree, which is up to five years in prison and a fine of at least $5000. Doesn’t make total sense, but that’s what the statute says.

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button. Con1mainp15 The purpose of this material was to give you a quick and basic overview of Intellectual Property and how it works here in the 52 words United States. Now it’s time for me to get some information about what you thought of this material. Let’s see how much you actually learned from me !

- If you want to listen to me again, click on the button. - If you want to proceed to the next screen, click on the < Next > button.

117

APPENDIX I INSTRUCTIONAL MATERIAL SCREEN (COMPANION AGENT + SEDUCTIVE GRAPHICS)

118

Figure I.1 Sample screen showing a main page

Figure I.2 Sample screen showing a companion role of agent

Figure I.3 Sample screen showing a purpose of instructional material

119

Figure I.4 Sample screen showing a concept of intellectual property

Figure I.5 Sample screen showing a three types of intellectual property

Figure I.6 Sample screen showing a concept of a patent (1)

120

Figure I.7 Sample screen showing a concept of a patent (2)

Figure I.8 Sample screen showing a conditions to be a patent

Figure I.9 Sample screen showing a concept of a trademark (1)

121

Figure I.10 Sample screen showing a concept of a trademark (2)

Figure I.11 Sample screen showing a concept of a copyright (1)

Figure I.12 Sample screen showing a concept of a copyright (2)

122

Figure I.13 Sample screen showing a concept of a copyright (3)

Figure I.14 Sample screen showing a closing of material

123

APPENDIX J INSTRUCTIONAL MATERIAL SCREEN (INSTRUCTOR AGENT + SEDUCTIVE GRAPHICS)

124

Figure J.1 Sample screen showing a main page

Figure J.2 Sample screen showing a companion role of agent

Figure J.3 Sample screen showing a purpose of instructional material

125

Figure J.4 Sample screen showing a concept of intellectual property

Figure J.5 Sample screen showing a three types of intellectual property

Figure J.6 Sample screen showing a concept of a patent (1)

126

Figure J.7 Sample screen showing a concept of a patent (2)

Figure J.8 Sample screen showing a conditions to be a patent

Figure J.9 Sample screen showing a concept of a trademark (1)

127

Figure J.10 Sample screen showing a concept of a trademark (2)

Figure J.11 Sample screen showing a concept of a copyright (1)

Figure J.12 Sample screen showing a concept of a copyright (2)

128

Figure J.13 Sample screen showing a concept of a copyright (3)

Figure J.14 Sample screen showing a closing of material

129

APPENDIX K INSTRUCTIONAL MATERIAL SCREEN (TEXT MESSAGE + SEDUCTIVE GRAPHICS)

130

Figure K.1 Sample screen showing a main page

Figure K.2 Sample screen showing a text only seductive message

Figure K.3 Sample screen showing a purpose of instructional material

131

Figure K.4 Sample screen showing a concept of intellectual property

Figure K.5 Sample screen showing a three types of intellectual property

132

Figure K.6 Sample screen showing a concept of a patent (1)

Figure K.7 Sample screen showing a concept of a patent (2)

133

Figure K.8 Sample screen showing a conditions to be a patent

Figure K.9 Sample screen showing a concept of a trademark (1)

134

Figure K.10 Sample screen showing a concept of a trademark (2)

Figure K.11 Sample screen showing a concept of a copyright (1)

Figure K.12 Sample screen showing a concept of a copyright (2)

135

Figure K.13 Sample screen showing a concept of a copyright (3)

Figure K.14 Sample screen showing a closing of material

136

APPENDIX L INSTRUCTIONAL MATERIAL SCREEN (BASE + SEDUCTIVE GRAPHICS)

137

Figure L.1 Sample screen showing a main page

Figure L.2 Sample screen showing a text only seductive message

Figure L.3 Sample screen showing a purpose of instructional material

138

Figure L.4 Sample screen showing a concept of intellectual property

Figure L.5 Sample screen showing a three types of intellectual property

Figure L.6 Sample screen showing a concept of a patent (1)

139

Figure L.7 Sample screen showing a concept of a patent (2)

Figure L.8 Sample screen showing a conditions to be a patent

Figure L.9 Sample screen showing a concept of a trademark (1)

140

Figure L.10 Sample screen showing a concept of a trademark (2)

Figure L.11 Sample screen showing a concept of a copyright (1)

Figure L.12 Sample screen showing a concept of a copyright (2)

141

Figure L.13 Sample screen showing a concept of a copyright (3)

Figure L.14 Sample screen showing a closing of material

142

APPENDIX M SHAPIRO-WILK NORMALITY TESTS FOR LEARNING INTEREST AND ATTITUDE

143 Shapiro-Wilk normality tests for learning interest

Five dependent variables of Condition Shapiro-Wilk df Sig learning interest statistic Feeling-related interest Sg + Ca .912 15 .145 Sg + Ia .853* 16 .015 Sg + Tx .962 16 .693 Sg + No .924 17 .174 Ng + Ca .947 15 .475 Ng + Ia .920 16 .166 Ng + Tx .951 15 .537 Ng + No .905 17 .082 Value-related interest Sg + Ca .943 15 .421 Sg + Ia .942 16 .379 Sg + Tx .886* 16 .048 Sg + No .965 17 .718 Ng + Ca .884 15 .054 Ng + Ia .922 16 .180 Ng + Tx .906 15 .116 Ng + No .951 17 .469 Arousal Sg + Ca .889 15 .064 Sg + Ia .937 16 .318 Sg + Tx .862* 16 .021 Sg + No .930 17 .214 Ng + Ca .918 15 .179 Ng + Ia .940 16 .349 Ng + Tx .896 15 .081 Ng + No .894 17 .053 Involvement Sg + Ca .944 15 .436 Sg + Ia .729* 16 .000 Sg + Tx .910 16 .115 Sg + No .920 17 .150 Ng + Ca .951 15 .537 Ng + Ia .900 16 .079 Ng + Tx .889 15 .064 Ng + No .847* 17 .010 Attention Sg + Ca .869* 15 .033 Sg + Ia .890 16 .056 Sg + Tx .971 16 .855 Sg + No .987 17 .995 Ng + Ca .931 15 .279 Ng + Ia .951 16 .511 Ng + Tx .942 15 .411 Ng + No .931 17 .226 Note. Sg: seductive graphic, Ng: No graphic, Ca: Companion role of agent, Ia: Instructor role of agent, Tx: Text message only, No: No message

144 Shapiro-Wilk normality tests for attitude

Three dependent variables of Condition Shapiro-Wilk df Sig attitude statistic Relevance Sg + Ca .970 15 .854 Sg + Ia .935 16 .292 Sg + Tx .907 16 .103 Sg + No .910 17 .099 Ng + Ca .919* 15 .024 Ng + Ia .939 16 .336 Ng + Tx .835* 15 .011 Ng + No .978 17 .935 Confidence Sg + Ca .972 15 .884 Sg + Ia .916 16 .145 Sg + Tx .960 16 .653 Sg + No .931 17 .226 Ng + Ca .941 15 .400 Ng + Ia .942 16 .373 Ng + Tx .916 15 .167 Ng + No .950 17 .455 Satisfaction Sg + Ca .899 15 .092 Sg + Ia .858* 16 .018 Sg + Tx .965 16 .758 Sg + No .942 17 .339 Ng + Ca .924 15 .222 Ng + Ia .958 16 .622 Ng + Tx .963 15 .749 Ng + No .884* 17 .037 Note. Sg: seductive graphic, Ng: No graphic, Ca: Companion role of agent, Ia: Instructor role of agent, Tx: Text message only, No: No message

145

APPENDIX N BOX TESTS FOR LEARNING INTEREST AND ATTITUDE

146

Box’s test of equality of covariance matrices for learning interest

Box's 121.360 M F .993 df1 105 df2 17610.748 Sig. .502

Box’s test of equality of covariance matrices for attitude

Box's 68.347 M F 1.491 df1 42 df2 23013.778 Sig. .021

147

APPENDIX O LEVENE’S TESTS FOR LEARNING INTEREST AND ATTITUDE

148

Levene’s test of equality of error variances for learning interest

F df1 df2 Sig. Feeling related 1.249 7 119 .282 interest Value related 1.437 7 119 .197 interest Arousal .305 7 119 .951 Involvement .967 7 119 .458 Attention 1.259 7 119 .277

Levene’s test of equality of error variances for attitude

F df1 df2 Sig. Relevance 1.928 7 119 .071 Confidence .577 7 119 .773 Satisfaction 2.403 7 119 .025

149

APPENDIX P SHAPIRO-WILK NORMALITY TESTS FOR RECALL TEST AND COMPREHENSION TEST

150

Shapiro-Wilk normality tests for recall test

Dependent variable Condition Shapiro-Wilk df Sig statistic Recall test Sg + Ca .922 15 .204 Sg + Ia .838* 16 .009 Sg + Tx .774* 16 .001 Sg + No .942 17 .349 Ng + Ca .917 15 .173 Ng + Ia .857* 16 .017 Ng + Tx .785* 15 .002 Ng + No .934 17 .250 Note. Sg: seductive graphic, Ng: No graphic, Ca: Companion role of agent, Ia: Instructor role of agent, Tx: Text message only, No: No message

Shapiro-Wilk normality tests for comprehension test

Dependent variable Condition Shapiro-Wilk df Sig statistic Comprehension test Sg + Ca .943 15 .427 Sg + Ia .962 16 .693 Sg + Tx .928 16 .230 Sg + No .833* 17 .006 Ng + Ca .889 15 .064 Ng + Ia .914 16 .137 Ng + Tx .925 15 .230 Ng + No .936 17 .279 Note. Sg: seductive graphic, Ng: No graphic, Ca: Companion role of agent, Ia: Instructor role of agent, Tx: Text message only, No: No message

151

APPENDIX Q LEVENE’S TESTS FOR RECALL TEST AND COMPREHENSION TEST

152

Levene’s test of equality of error variances for recall test

F df1 df2 Sig. 1.211 7 119 .302

Levene’s test of equality of error variances for comprehension test

F df1 df2 Sig. 1.641 7 119 .130

153

APPENDIX R HUMAN SUBJECT COMMITTEE APPROVAL

154 155

156

REFERENCES

Alexander, P. A., & Jetton, T. L. (1996). The role of importance and interest in the processing of text. Educational Psychology Review, 8, 89-121.

Alexander, P. A., Jetton, T. L., & Kulikowich, J. M. (1995). The relationship of knowledge, interest, and recall: Assesing a model of domain learning. Journal of Educational Psychology, 87, 559-575.

Alexander, P. A., & Kulikowich, J. M. (1991). Domain-specific and strategic knowledge as predictors of expository text comprehension. Journal of Reading Behavior, 23, 165-190.

Alexander, P. A., Kulikowich, J. M., & Jetton, T. L. (1994). The Role of Subject-Matter Knowledge and Interest in the Processing of Linear and Nonlinear Texts. Review of Educational Research, 64(2), 201-252.

Alexander, P. A., Schallert, D. L., & Hare, V. C. (1991). Coming to terms: How researchers in learning and literacy talk about knowledge. Review of Educational Research, 61, 315-343.

Anglin, G. J., Towers, R. L., & Levis, W. H. (1996). Visual Message Design and Learning: The role of Static and Dynamic Illustrations. In D. H. Jonassen (Ed.), Handbook of Research for Educational Communications and Technology (pp. 755-794). New York, NY: Simon & Schuster Macmillan.

Baylor, A. L. (2000). Beyond butlers: Intelligent agents as mentors. Journal of Educational Computing Research, 22(4), 373-382.

Baylor, A. L. (2002a). Agent-based learning environments for investigating teaching and learning. Journal of Educational Computing Research, 26(3), 249-270.

Baylor, A. L. (2002b). Expanding pre-service teachers' metacognitive awareness of instructional planning through pedagogical agents. Educational Technology Research & Development, 50(2), 5-22.

Baylor, A. L. (2003). The Impact of Three Pedagogical Agent Roles. Paper presented at the Autonomous Agents & Multi-Agent Systems (AAMAS), Melbourne, Australia.

157 Baylor, A. L., & Ryu, J. (2003). Does the presence of image and animation enhance pedagogical agent persona? Journal of Educational Computing Research, 28(4), 373-395.

Berlyne, D. E. (1965). Curiosity and education. In J. D. Krumboltz (Ed.), Learning and the educational process. Chicago, IL: Rand McNally.

Brody, P. J. (1984). In search of instructional utility: a function based approach to pictorial research. Instructional Science, 13, 47-61.

Concannon, S. J. (1975). Illustrations in books for children: review of research. The Reading Teacher, 29, 254-256.

Craig, S. D., & Gholson, B. (2002). Does an agent matter?: The Effects of Animated Pedagogical Agents on Multimedia Environments. Paper presented at the ED- Media, Norfolk, VA.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2 ed.). Hillsdale, NJ: Lawrence Erlbaum Associates.

Dehn, D., M., & Mulken, S. (2000). The impact of animated interface agents: A review of empirical research. International Journal of Human-Computer Studies, 52, 1-22.

Dewey, J. (1913). Interest and effort in education. Boston, MA.

Eccles, J., S., & Wigfield, A. (2002). Motivational Beliefs. Values, and Goals. Annual Review of Pstchology, 53, 109-132.

Forgas, J. P. (1998). On feeling good and getting your way: Mood effects on negotiator cognition and bargaining strategies. Journal of Personality and Social Psychology, 74, 565-577.

Fredrickson, B. L. (1998). What Good are Positive Emotions?. Review of General Psychology, 2(3), 300-319.

Gadanho, S. C., & Hallam, J. (2001). Robot learning driven by emotions. Adaptive Behavior, 9(1).

Gagne, R. M., Briggs, L. J., & Wager, W. W. (1988). Principles of instructional design (3rd ed.). New York, NY: Holt, Rinehart and Winston, Incorporated.

Garner, R., & Gillingham, M. G. (1991). Topic knowledge, cognitive interest, and text recall: A microanalysis. Journal of Experimental Education, 59, 310-319.

158 Garner, Brown, R., Sanders, S., & Menke, D. J. (1992). Seductive details and learning from text. In A. Renninger, S. Hidi & A. Krapp (Eds.), The role of interest in learning and development (pp. 239-254). Hillsdale, NJ: Erlbaum.

Garner, R. (1992). Learning from school texts. Educational Psychologist, 27, 53-63.

Garner, R., Alexander, P. A., Gillingham, M. G., Kulikowich, J. M., & Brown, R. (1991). Interest and Learning From Text. American Educational Research Journal, 28(3), 643-659.

Garner, R., Gillingham, M. G., & White, C. S. (1989). Effects of "seductive details" on macroprocessing and microprocessing in adults and children. Cognition and Instruction, 6, 41-57.

Harp, S. F., & Mayer, R. E. (1997). The Role of Interest in Learning From Scientific Text and Illustrations: On the Distinction Between Emotional Interest and Cognitive Interest. Journal of Educational Psychology, 89(1), 92-103.

Harp, S. F., & Mayer, R. E. (1998). How seductive details to their damage: A theory of cognitive interest in science learning. Journal of Educational Psychology, 90, 414-434.

Hidi, S. (1990). Interest and its contribution as a mental resource for learning. Review of Educational Research, 60, 549-572.

Hidi, S. (2001). Interest and reading: Theoretical and practical considerations. Educational Psychology Review, 13(3), 191-208.

Hidi, S., & Anderson, V. (1992). Situational interest and its impact on reading and expository writing. In A. Krapp, S. Hidi & A. Renninger. (Eds.), The role of interest in learning and development (pp. 215-238). Hillsdale, NJ: Lawrence Erlbaum Association.

Hidi, S., & Baird, W. (1986). Interestingness - a neglected variable in discourse processing. Cognitive Science, 10, 179-194.

Holliday, W. G. (1973). Critical analysis of pictorial research related to science education. Science Education, 57(2), 201-214.

Isen, A. M., & Baron, R. A. (1991). Positive affect as a factor in organizational behavior. Research in Organizational Behavior, 13, 1-53.

Isen, A. M., & Daubman, K. A. (1984). The influence of affect on categorization. Journal of Personality and Social Psychology, 47, 1206-1217.

159 Isen, A. M., Daubman, K. A., & Nowicki, G. P. (1987). Positive affect facilitates creative problem solving. Journal of Personality and Social Psychology, 52, 1122-1131.

Isen, A. M., Johnson, M. S., Mertz, E., & Robinson, G. F. (1985). The influence of positive affect on the unusualness of word associations. Journal of Personality and Social Psychology, 48, 1413-1426.

Isen, A. M., & Means, B. (1983). The influence of positive affect on decision-making strategy. Social Cognition, 2, 18-31.

Izard, C. E. (1972). Anxiety: a variable combination of interaction fundamental emotions. In C. D. Spielberger (Ed.), Anxiety: Current trends in theory and research. New York, NY: Academic press.

Izard, C. E., & Ackerman, B. P. (2000). Motivational, Organizational, and Regulatory Functions of Discrete Emotions. In M. Lewis & J. M. Haviland-Jones (Eds.), Handbook of Emotion (Vol. II, pp. 253-264). New York, NY: Guilford Publications, Inc.

Jetton, T. L., & Alexander, P. A. (2001). Interest Assessment and the Content Area Literacy Environment: Challenges for Research and Practice. Educational Psychology Review, 13(3), 303-318.

Johnson, W. L., Rickel, J. W., & Lester, J. C. (2000). Animated pedagogical agents: face- to-face interaction in interactive learning environments. International Journal of Artificial Intelligence in Education, 11, 47-78.

Keller, J. M. (1983). Motivational design of instruction. In C. M. Reigeluth (Ed.), Instructional design theories and models: An overview of their current status. Hillsdale, NJ: Erlbaum.

Keller, J. M. (1987). Strategies for stimulating the motivation to learn. Performance & Instruction, 26(8), 1-7.

Keller, J. M., & Subhiyah, R. (1993). Manual for the Course Interest Survey (CIS).Unpublished manuscript, Tallahassee, FL.

Keller, J. M. (1993). Manual for the Instructional Materials Motivational Survey (IMMS). Unpublished manuscript, Tallahassee, FL.

Keller, J. M. (2002). Motivation and performance. In R. A. Reiser (Ed.), Trends and issues in instructional design and technology. Upper saddle river, NJ: Pearson.

Kintsch, W. (1980). Learning from text, levels of comprehension. Poetics, 9, 87-98.

160 Krapp, A. (1999). Interest, motivation and learning: An educational-psychological perspective. European Journal of Psychology of Education, 15(1), 23-40.

Krapp, A. (2002). Structural and dynamic aspexts of interest development: theoretical considerations from an ontpgenetic perspective. Learning and Instruction, 12, 383-409.

Krapp, A., Hidi, S., & Renninger, A. (1992). Interest, Learning. and Development. In A. Krapp, S. Hidi & A. Renninger (Eds.), The role of interest in learning and development (pp. 3-25). Hillsdale, NJ: Lawrence Erlbaum Association.

Levi, W. H., & Lentz, R. (1982). Effects of text illustrations: a review of research. Educational Communication and Technology Journal, 30(4), 195-232.

Levin, J. R., & Mayer, R. E. (1993). Understanding illustrations in text. In B. Britton, A. Woodward & M. Binldey (Eds.), Learning from textbooks (pp. 95-113). Hillsdale, NJ: Lawrence Erlbaum Association.

Mayer, R. E. (1989). Systematic thinking fostered by illustrations in scientific text. Journal of Educational Psychology, 81, 240-246.

Mayer, R. E., Heiser, J., & Lonn, S. (2001). Cognitive Constraints on Multimedia Learning: When Presenting More Material Results in Less Understanding. Journal of Educational Psychology, 93(1), 187-198.

Mayer, R. E., & Moreno, R. (1998). A split-attention effect in multimedia learning: Evidence for dual information processing systems in working memory. Journal of Educational Psychology, 90, 312-320.

Mayer, R. E., Steinhoff, K., Bower, G., & Mars, R. (1995). A generative theory of textbook design: Using illustrations to foster meaningful learning of science text. Educational Technology Research and Development, 43, 31-43.

Moreno, R., & Mayer, R. E. (1999). Cognitive principles of multimedia learning: The role of modality and contiguity. Journal of Educational Psychology, 91, 358-368.

Moreno, R., & Mayer, R. E. (2000). A coherence effect in multimedia learning: The case for minimizing irrelevant sounds in the design of multiledia instructional messages. Journal of Educational Psychology, 92, 117-125.

Moreno, R., & Mayer, R. E. (2002). Verbal redundancy in multimedia learning: When reading helps listening. Journal of Educational Psychology, 94, 156-163.

161 Moreno, R., Mayer, R. E., Spires, H. A., & Lester, J. C. (2001). The case for social agency in computer-based teaching: do students learn more deeply when they interact with animated pedagogical agents? Cognition and Instruction, 19(2), 177- 213.

Picard, R. W. (1997). Affective computing. Cambridge, MA: The MIT Press.

Pressley, M., El-Dinary, P. B., Marks, M. B., Brown, R., & Stein, S. (1992). Good strategy instruction is motivating and interesting. In A. Renninger, S. Hidi & A. Krapp (Eds.), The Role of Interest in Learning and Development (pp. 333-357). Hillsdale, NJ: Erlbaum.

Renninger, K. A., Hidi, S., & Krapp, A. (1992). The role of interest in learning and development. Hillsdale, NJ: Erlbaum.

Reynolds, R. E. (1992). Selective attention and prose learning: Theoretical and empirical research. Educational Psychology Review, 4, 1-48.

Samuels, S. J. (1970). Effects of pictures on learning to read, comprehension and attitudes. Review of Educational Research, 40, 397-407.

Sansone, C., Weir, C., Harpster, L., & Morgan, C. (1992). Once a boring task always a boring task? Interest as a self-regulatory mechanism. Journal of Personality and Social Psychology, 63, 379-390.

Schallert, D. L. (1980). The role of illustrations in reading comprehension. In B. Spiro & W. F. Brewer (Eds.), Theoretical issues in reading comprehension (pp. 503-523). Hillsdale,NJ: Erlbaum.

Schiefele, U. (1991). Interest, learning, and motivation. Educational Psychologist, 26, 299-323.

Schiefele, U. (1992). Topic interest and levels of text comprehension. In A. Krapp, S. Hidi & A. Renninger (Eds.), The role of interest in learning and development (pp. 152-182). Hillsdale, NJ: Lawrence Erlbaum Association.

Schiefele, U. (1996). Topic interest, Text representation, and quality of experience. Contemporary Educational Psychology, 21, 19-42.

Schiefele, U. (1998). Individual Interest and Learning: What we know and What we do not know. In L. Hoffman, A. Krapp, K. A. Renninger & J. Baumert (Eds.), Interest and learning (pp. 91-104). Kiel, Germany: Institut fur die padogogik der Naturwissenshaften.

Schiefele, U. (1999). Interest and learning from text. Science Study and Reading, 3, 257- 280.

162 Schiefele, U., & Krapp, A. (1996). Topic Interest and Free recall of Expository Text. Learning and Individual Differences, 8(2), 141-160.

Schraw, G. (1998). Processing and recall differences among seductive details. Journal of Educational Psychology, 90, 3-12.

Schraw, G., Flowerday, T., & Lehman, S. (2001). Increasing situational interest in the classroom. Educational Psychology Review, 13(3), 211-224.

Schraw, G., & Lehman, S. (2001). Situational Interest: A Review of the Literature and Directions for Future Research. Educational Psychology Review, 13(1), 23-52.

Short, J., Williams, E., & Christie, B. (1976). The social psychology of telecommunications. London: John Willey & Sons.

Silvia, P. J. (2001). Interest and Interests: The Psychology of Constructive Capriciousness. Review of General Psychology, 5(3), 270-290.

Song, S. H. (1998). The effects of Motivationally Adaptive Computer-Assisted Instruction Developed through ARCS Model. Unpublished Unpublished Doctoral Dessertation, Florida State University, Tallahassee.

Spaulding, S. (1955). Research on pictorial illustration. AV communication Review, 3(35- 45).

Stipek, D. (2002). Motivation To Learn: Integrating Theory and Practice (4 ed.). Boston, MA: Allyn & Bacon.

Sweet, A. P., Guthrie, J. T., & Ng, M. M. (1997). Tearcher perceptions and student reading motivation. Journal of Educational Psychology, 90, 210-223.

Thalheimer, W. (2004). Bells, whistles, neon, and purple prose: When interesting words, sounds, and visuals hurt learning and performance - a review of the seductive- augmentation research. Retrieved August 16,, 2004, from http://www.worklearning.com/seductive_augmentations.htm

Thayer, R. E. (1985). Activation (arousal): The shift from a single to multidimensional perspective. Biological Bases of Personality and Behavior, 1, 115-127.

Thayer, R. E. (1986). Activation-Deactivation adjective check list: Current overview and structural analysis. Psychological Reports, 58, 606-614.

Tobias, S. (1994). Interest, Prior Knowledge, and Learning. Review of Educational Research, 64(1), 37-54.

163 Wade, S. E., & Adams, R. B. (1990). Effects of Importance and Interest on Recall of Biographical Text. Journal of Reading Behavior, 12(4), 331-353.

Wade, S. E., Schraw, G., Buxton, W. M., & Hayes, M. T. (1993). Seduction of the strategic reader: Effects of interest on strategies and recall. Reading Research Quartly, 28, 3-24.

Warren, D. R., Shen, E., Park, S., & Balyor, A. L. (2004). A phenomenological Study of the perceptions of Adult Education Students toward Pedagogical Agents with Evasive Emotion: What do Real Talking Heads want from the Little Talking Heads.Unpublished manuscript, Tallahassee, FL.

164

BIOGRAPHICAL SKETCH

Sanghoon Park was born March 16, 1974 in Seoul, Korea

EDUCATION______Florida State University Tallahassee, FL Ph.D. in Instructional Systems 2005 Minors in Measurement and Statistics Dissertation Chair: Dr. John M. Keller Dissertation Title: The Effects of Seductive Augmentation and Agent Role on Learning Interest, Achievement, and Attitude.

Hanyang University Seoul, Korea M.S. in Educational Technology 2001 Thesis Advisor: Dr. Sungho Kwon Thesis Title: The effects of tutor’s motivational strategy on learning motivation in Web-based Instruction: Focused on personal relevance messages.

Kookmin University Seoul, Korea B.A. in Educational Foundations 1999 Minors in English & English Literature

TEACHING EXPERIENCE______Instructor Fall 2003 EME2040 Introduction to Educational Technology - Fall 2004 Taught required technology course for pre-service teachers & Fall 2002 earning teaching credentials.

Teaching Assistant Spring 2004 EDF 5464 Qualitative Methods for Program Evaluation Assisted faculty with preparing course materials Managed course blackboard for on-line discussion Mentored students’ class works

Teaching Assistant Spring 2003 EME 6635 Advanced Seminar on Motivation Issues Assisted faculty with preparing supplemental materials Managed course blackboard for on-line discussion and assignment submission; Coordinated course schedule

165 Teaching Assistant Spring 2003 EME 5601 Introduction to Instructional Systems Assisted faculty with preparing course materials & grading papers Managed course blackboard for on-line discussion

On-line Tutor Fall 2000 Hanyang Cyber Learning Center Designed on-line course for college of information and communication (Course title: ASIC design and development) Developed course materials: Workbook, on-line references

Teaching Assistant Spring 1999 Taught courses in video and photography production - Spring 2000

RESEARCH EXPERIENCE______Motivation & Volitional Strategy Spring 2005 Research Assistant - Present Responsibilities include reviewing related literatures on motivation and volition, building motivational and volitional strategies, designing research study, analyzing collected data, and writing up the findings.

Pedagogical Agents Research Summer 2004 Research Assistant - Present Responsibilities include designing research study, developing pedagogical agents, and analyzing collected data for the experiments to examine the effectiveness of agent-based learning environment that Dr. Baylor is funded by National Science Foundation (NSF).

Cognitive Load Theory Spring 2004 Research Assistant - Present Responsibilities include designing research study, developing instruments, preparing study materials, conducting experiments, analyzing collected data.

Mobile Learning Fall 2003 Researcher - Present Responsibilities include reviewing related literatures on using PDA/Table PC for educational purpose, designing research study, developing instruments, designing study material, analyzing collected data, writing up the findings.

166 PROJECTS EXPERIENCE______Florida Department of Revenue, Tallahassee, FL Summer 2003 Instructional Designer - Spring 2004 Designed and Developed Web training courses for certificate of tax collectors and staff members in the State of Florida

Dept. of Educational Psychology and Learning Systems Fall 2002 Florida State University, Tallahassee, FL - Spring 2003 Database developer Developed a database (Advisee ver. 1.0) for managing graduate students information for Department of Educational Psychology & Learning Systems

AECT (Association for Educational Communications and Fall 2001 Technology), Bloomington, IN - Summer 2002 Web content developer Developed PT3 Web content: criteria and examples to promote teaching practices in math and science fields

Joongang IT Academy, Seoul, Korea Spring 2001 Contents Designer and Programmer Designed and developed a Knowledge Management System for the information technology professions

Korea Research Foundation, Seoul, Korea Fall 2000 Research Assistant - Spring 2001 Designed and developed learner support strategies for collaborative Interaction in Web based learning environment

Ministry of Labor, Seoul, Korea Summer/2000 E-learning designer - Fall/2000 Designed E-learning system on safety education for Korean middle and small size companies

Institute of Educational Technology (IET), Hanyang university, Spring 2000 Seoul, Korea - Fall 2000 Research Assistant Assisted faculties with preparing instructional materials, Coordinated workshops of IET, Wrote research proposals for research fund, and served as a editor of IET journals

SK telecom, Seoul, Korea Summer 1999 E-learning designer - Fall 1999 Designed E-learning system on e-commercial for SK telecom

167 Ministry of Education, Seoul, Korea Fall 1999 Research Assistant - Fall 2000 Worked as a research team member of Brain Korea 21 Project focusing on the design and development of On-line distance learning system

PUBLICATION ______Articles in Refereed Journals

Park, S., & Kwon, S. (2001). The Effects of Motivational Strategy on Learning and Motivation. Research in educational information & broadcasting, 7-2.

Park, S., Kim, M., Lee, Y., Son, C., & Lee, M. (2006, In press). The Effects of Visual Illustrations on Learners' Achievement and Interest in PDA- (Personal Digital Assistant) Based Learning. Journal of Educational Computing Research, 33(2)

Selected papers in Refereed Proceedings

Park, S., & Lim, J. (2005, October). Measuring Flow Experience of Pre-Service Teachers in Performing E-Portfolio. Proceeding of E-Learn 2005 World Conference of Association for the Advancement of Computing in Education (AACE), Vancouver, BC Canada.

Park, S., & Lim, J. (2005, October). Promoting Positive Emotion in Multimedia Learning using Visual Illustrations. Proceeding of E-Learn 2005 World Conference of Association for the Advancement of Computing in Education (AACE), Vancouver, BC Canada

Olina, Z., Reiser, R., Huang, X., Lim, J., & Park. S. (2005, August). Applying Theory to Teaching Comma Rules to US High School Students, 11th Biennial Conference of European Association for Research on Learning and Instruction (EARLI), Nicosia, Cyprus.

Warren, D., Shen, E., Park, S., Baylor, A., & Perez, R. (2005, July). Adult Learner Perceptions of Affective Agents: Experimental Data and Phenomenological Observations, 12th International Conference on Artificial Intelligence in Education (AIED 2005), Armsterdam, Netherlands.

Baylor, A., Warren, D., Park, S., Shen, E., & Perez, R. (2005, July). The Impact of Frustration-mitigating Messages Delivered by an Interface Agent, 12th International Conference on Artificial Intelligence in Education (AIED 2005), Armsterdam, Netherlands

Park, S. (2004, October). Building Bridge Between Learning and Positive Emotion: How to Apply Emotional Factor in Instructional Designing Process?, Proceeding of International Conference of the Association for Educational Communications and Technology (AECT), Chicago , Illinois .

168 Park, S., & Lim, J. (2004, October). The Effect of Graphical Representation of Science Text on the Learner's Learning Interest And Achievement in Multimedia Learning , Proceeding of International Conference of the Association for Educational Communications and Technology (AECT), Chicago, Illinois.

Olina, Z., Huang, X., Park, S., Lim, J., & Reiser, R. (2005, April). Effects of Problem Format and Presentation Sequence on High School Student Cognitive Load and Learning, 2005 Paper presented at the Annual meeting of the American Educational Research Association (AERA), Montreal, Canada.

Park, S., & Ryu, J. (2004, March). The effects of pre-service teacher's self-reflection and achievement on the satisfaction with E-portfolio, Proceeding of Association for the Advancement of Computing in Education (AACE) SITE 2004 World Conference, Atlanta, Georgia.

Park, S., Kim, M. & Ryu, J. (2003, November). The Effect of Character Spacing on Reading in Personal Digital Assistants (PDA), Proceeding of Association for the Advancement of Computing in Education (AACE) E-Learn 2003 World Conference, Phoenix, Arizona.

Ryan, W. & Park, S. (2003, October). An Integration of Situated Cognition and Interaction Theory in Designing of Web-based Learning Environments , Proceeding of International Conference of the Association for Educational Communications and Technology (AECT), Anaheim, California.

PRESENTATIONS______Refereed Presentations

Park, S. & Keller, J. (2005, October). The Effects of Seductive Motivational Message on Study habit, Motivation and Achievement: How to keep them interested. Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Orlando, Florida.

Olina, Z., Huang, X., Kim, H., Lim, J., Liu, Y., Park, S., & Reiser, R. (2005, October). Use of Instructional Explanations versus Self-Explanations in Worked Examples during Learning of Comma Rules. Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Orlando, Florida.

Perez, R., Warren, D., & Park, S. (2005, October). An Online Learning Community Tool for the Collaborative Creation of an Instructional Systems Dictionary. Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Orlando, Florida.

West, R., Waters, S., Polly, D., Subramony, D, & Park, S. (2005, October). Integrating New Students into the Field of Instructional Technology. Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Orlando, Florida.

169 Subramony, D., Park, S., Strobel, J., Tillberg, H., & Polly, D. (2005, October). Maximizing AECT Involvement as Graduate Student Reflections from the 2004 AECT Interns. Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Orlando, Florida.

Park, S. (2005, March). The Effects of personal relevance motivational message on Learning Motivation and Participation in Web-Based Instruction. Paper presented at the 1st Southeastern Conference in Instructional Design and Technology, Mobile, AL

Lim, J., & Park, S. (2005, March). A Strategy Supporting Self-regulated Learning in a Hybrid Learning Environment. Paper presented at the 1st Southeastern Conference in Instructional Design and Technology, Mobile, AL

Kim, M., Lee, M., Lee, Y., Park, S., & Son, C. (2005, January). Overviewing Return On Investment (ROI) in E-Learning. Paper presented at the Southeast Evaluation Association (SEA) 17th Annual conference, Tallahassee, Florida.

Solomon, H. & Park, S. (2004, October). The Move from Training to Learning and Performance at Obscurity Mutual, Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Chicago, Illinois. (Paper nominated for the PacifiCorp Design and Development Award)

Olina, Z., Huang, X., Park, S., Lim, J., & Reiser, R. (2004, October). Effects of Cued Problems and Problem Presentation Sequence during Practice on Cognitive Load, Achievement and Transfer Test Performance, Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Chicago, Illinois.

Son, C., Kim, M., & Park, S. (2004, March). The case analysis of PDA integrated classroom, Paper presented at the Association for the Advancement of Computing in Education (AACE) SITE 2004 World Conference, Atlanta, Georgia.

Park, S. (2004, January). A Formative Product Evaluation of Course Resource Management Tool. Paper presented at the Southeast Evaluation Association (SEA) 16 the Annual conference, Tallahassee, Florida. (Paper nominated for the award for best student paper)

Park, S., & Kim, M. (2004, January). Qualitative Approach to International Instructor Evaluation: A Case from field. Paper presented at the Southeast Evaluation Association (SEA) 16th Annual conference, Tallahassee , Florida.

Park, S. & Lee, Y. (2003, October). How to Select the Best- Fitting EPSS (Electronic Performance Support System) for My Organization?, Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Anaheim, California .

170 Park, S. & Oh, S. (2003, October). Expanding the Motivational Design Process throughout On-line Courses in Higher Education, Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Anaheim, California.

Lee, Y. & Park, S. (2002, November). How far were we from the ideal knowledge management system (KMS)?: A case study for developing the KMS , Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Dallas, Texas.

Kwon, S. & Park, S. (2000, September). New trends of cyber education in Korea. Paper presented at the symposium on virtual learning, Northern Arizona University. Flagstaff, Arizona.

Poster Presentations

Kim, M., Lee, M., Lee, Y., Park, S., Ryu, J., & Son, C. (2005, October). Potential Use of PDAs (Personal Digital Assistant) in Education: A Result of Survey, Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Orlando, Florida.

Park, S., Kim, M., Lee, Y., Lee, M., & Son, C. (2005, October). The Effects of Visual Illustration on Learning Interest and Achievement in PDA (Personal Digital Assistant) based Learning, Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Orlando, Florida

Park, S., & Son, C. (2004, October). Designing and Development of EPSS (Electronic Performance Support System): Case of Course Resource Management Tool (CRMT), Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Chicago, Illinois.

Chow, A., Park, S., & Isomi, T. (2004, October). Opening Doors: The Use of Performance Systems Analysis in Identifying Gaps and Assisting in Organizational Alignment, Paper presented at the International Conference of the Association for Educational Communications and Technology (AECT), Chicago, Illinois.

Park, S. (2004, March). The Advisee 1.0 for Advisor faculty: Tool for Effective Advising Service in Higher Education, Paper presented at the Association for the Advancement of Computing in Education (AACE) SITE 2004 World Conference, Atlanta, Georgia.

SCHOLARLY AWARDS______Gagne Briggs Outstanding Doctoral Student Award (2004-2005), EPLS, College of Education , Florida State University.

Ruby Diamond Future Professor Award (2004-2005), EPLS, College of Education , Florida State University.

171 Award of Cochran Intern by Educational Communication Technology (ETC) Foundation, the International Conference of Association of Educational Communications and Technology (AECT) of 2004 in Chicago, Illinoise.

Co-recepient of AECT's 2002 Service Award : PT3 AECT project awarded by Association for Educational Communications and Technology (AECT)

Award of Graduate Student Professional Meeting by the Council on Research in Education (CORE) for the Spring Semester of 2004/2005 funded by the College of Education , Florida State University .

Finalist for the Liliana Muhlman Masoner Award: Excellent international student award (2004-2005), EPLS, College of Education , Florida State University.

Finalist for the PacifiCorp Design and Development Award, Design and Development Division, the International Conference of Association of Educational Communications and Technology (AECT) of 2004 in Chicago, Illinoise.

Finalist for the Gagne Briggs Outstanding Service award (2004-2005), EPLS, College of Education , Florida State University.

Finalist for the Liliana Muhlman Masoner Award: Excellent international student award (2003-2004), EPLS, College of Education , Florida State University.

Award of the Brain Korea 21 Research Grant (1999-2000) by Ministry of Education in Korea

SERVICE FOR THE PROFESSIONAL COMMUNITY__ Graduate Student Volunteer of E-Learn 2005, Association for the Advancement of Computing in Education (AACE) world conference, Vancouver, British Columbia, Canada.

Delegation Leader for the American Society for Training and Development (ASTD), June /2005 and May/2004. Delegation coordinator of the Korea Banking Institute (KBI).

Proposal Reviewer of AECT (2003 - ), Association for Educational Communications and Technology (AECT) international conference, Orlando, Florida, USA.

Proposal Reviewer of AERA (2004 - ), American Educational Research Association (AERA) 2005 Annual Meeting, Montreal, Canada.

Graduate Student Volunteer of SITE 2004, Association for the Advancement of Computing in Education (AACE) world conference, Atlanta, Georgia, USA.

Graduate Student Volunteer of AECT 2003, Association for Educational Communications and Technology (AECT) international conference, Anaheim, California, USA.

172 Graduate Student Volunteer of WebNet 2001, Association for the Advancement of Computing in Education (AACE) world conference, Orlando, Florida, USA.

President (Summer, 2004 – Summer 2005), Korean Student Association in Instructional Systems Program, Florida State University.

CERTIFICATES______Certificate of Measurement and Statistics by Department of Educational Psychology and Learning Systems, Florida State University, Tallahassee, Florida. Certificate of Web master by Joongang Information Technology Academy , Seoul , Korea .

Certificate of Teacher in English for Secondary Education by Minister of Education, Seoul, Korea.

Certificate of Specialist for life-long education by Minister of Education, Seoul, Korea.

PROFESSIONAL MEMBERSHIPS ______American Educational Research Association (AERA)

Association for Educational Communications and Technology (AECT)

Association for the Advancement of Computing in Education (AACE)

Southeastern Evaluation Association (SEA)

The Korean Society for Educational Technology (KSET)

The Korea Association of Educational Information & Broadcasting (KAEIB)

173