Applying the Technology Acceptance Model (TAM) to educational hypermedia: a field study.
This article applies the Technology Acceptance Model (TAM) to an online course companion site of a textbook to be used by participants in this study. This article reviews literature on TAM and evaluates a set of hypotheses based on the theoretical relationships established in the TAM model. Results suggest that TAM is overall an effective tool for predicting user acceptance of such web-based course support systems and for evaluating competing hypermedia-based educational products. Discussion and suggestions for future research are offered.
The use of the Internet is increasingly prevalent in the educational environment. We observed that many textbook authors and publishers put up companion websites to stay competitive in the textbook market. Traditionally, only instructors receive something such as an instructor's manual, presentation slides, and test files from the publisher. With the companion sites, students can also obtain a lot of information such as chapter outlines, lecture files, simulated case studies, practice tests, web-based exercises, and hyperlinks to websites related to course coverage, including examples, cases, and pertinent government regulations and standards. Such course companions are not born equal. Factors determining acceptance and use of such sites need to be uncovered so that student input can be used in the selection process of textbooks based on a simple, useful, yet cost-effective mechanism. However, the lack of theoretical or conceptual frameworks in many past studies dealing with the effectiveness of web-based learning systems resulted in inconsistent results and left the question of what constitute the determining factors of effective delivery of educational hypermedia unanswered (Psaromiligkos & Retalis, 2003). The technology acceptance model's (TAM) parsimonious nature makes it a good candidate for such a purpose. This study examines the validity of TAM's theoretical relationships when applied to a course companion site. This article first presents the theoretical foundation of the TAM model, followed by a brief review of recent literature on TAM. Then it proposes and tests six hypotheses based on the TAM model through a field study in an educational institution. Finally, discussion of the findings, limitations, and future research opportunities are provided.
THE TECHNOLOGY ACCEPTANCE MODEL (TAM)
In information systems research, the user's attitude toward using and the actual usage of a technology are addressed in the TAM (Davis, 1989; Davis, Bagozzi, & Warshaw, 1989). TAM is rooted in the Theory of Reasoned Actions (TRA) (Fishbein & Ajzen, 1975; Ajzen & Fishbein, 1980) in psychology research. It proposes that perceived ease of use and perceived usefulness of technology are predictors of user attitude toward using the technology, subsequent behavioral intentions, and actual usage. Perceived ease of use was also considered to influence perceived usefulness of technology. TAM has been applied in numerous studies testing user acceptance of information technology, for example, word processors (Davis et al., 1989), spreadsheet applications (Mathieson, 1991), e-mail (Szajna, 1996), web browser (Morris & Dillon, 1997), telemedicine (Hu, Chau, Sheng, & Tam, 1999), and websites (Koufaris, 2002). In this study, a hypermedia-based course support site was considered a system that makes use of internet and web technology in accomplishing its mission of delivering information to and interacting with the user through a computer interface. Thus such a system fits into the paradigm of research on technology acceptance. The following is a diagram of TAM based on Davis et al. (1989).
[FIGURE 1 OMITTED]
TAM can serve the purpose of predicting user acceptance of a technology before the users get heavily involved in the technology, and thus is a cost-effective tool in screening potential candidate systems or programs.
In TAM, perceived usefulness refers to the degree to which the user believes that using the technology will improve his or her work performance, while perceived ease of use refers to how effortless he or she perceives using the technology will be. Both are considered distinct factors influencing the user's attitude toward using that technology, though perceived ease of use is also hypothesized to influence perceived usefulness. Subsequently, intention to use the technology is considered a function of perceived usefulness and attitude toward using the technology. Finally, such behavioral intention determines the actual usage of the technology (Davis et al., 1989).
This section presents the list of hypotheses based on the diagram of the TAM model. Though these hypotheses are stated below in abbreviated forms for simplicity and clarity, all constructs in these hypotheses are to be considered in the context of a web-based course support system.
H1: Perceived ease of use is positively related to perceived usefulness.
H2: Perceived ease of use is positively related to attitude toward using.
H3: Perceived usefulness is positively related to attitude toward using.
H4: Perceived usefulness is positively related to intention to use.
H5: Attitude toward using is positively related to intention to use.
H6: Intention to use is positively related to actual use.
A field study was conducted to evaluate the application of TAM to one type of educational hypermedia, a course companion site. It was related to a textbook that was to be used by junior and senior students in the information systems major of a northeastern college in the US. The site contained a section of study guides, through which a student could review the key concepts of each chapter and take practice tests generated randomly from a test bank, and then have these tests graded. They could view PowerPoint slides of lecture notes corresponding to their textbook. They could also preview or work on some of the web-based exercises or case studies related to each chapter. Additionally, the site offered links to firms or subjects of interest related to the covered topics in the textbook.
Accordingly, participants were drawn from among students who were either using the textbook or would potentially use the textbook in a subsequent semester in their curriculum. They were told to provide an evaluation of a course companion site for the purpose of a feedback on the choice of the current textbook and textbook selection in the future. Each participant was asked to explore the site for about 10 minutes and fill out a single-page questionnaire indicating his/her agreement or disagreement with each statement on a 7-point semantic differential scale. Scale items appearing on the survey were adapted from scales measuring variables in Davis et al. (1989). The instrument items used in this study are shown in Appendix A.
Sample demographic information with respect to age, gender, and prior experience with the Internet was also taken for potential control purposes in data analysis. Descriptive statistics collected from the survey showed that the majority of the participants were computer-savvy, and most spent over 10 hours a week on the Internet. Slightly over half of the participants were female, and the average age of the participants was about 21, reflecting the population from which we drew our sample. Since the research focus was educational hypermedia, the use of student sample here was deemed appropriate.
Cronbach's alpha of each multi-item scale was obtained and summarized in the following table. They all exceeded the recommended level of 0.70 (Nunnally, 1978). Thus all scales were reliable and had high internal consistency.
To examine construct validity of measures adopted in this study, a factor analysis was performed. The validity of the scales were tested by a principle component analysis with direct oblimin rotation, an appropriate method when there was reason to expect correlated factors (Pedhazur & Schmelkin, 1991). The results indicate that items within each scale loaded highest on the same factor, and items belonging to different scales loaded on their own separate factors, showing evidence of discriminant validity. Table 2 shows that perceived usefulness loaded on the first factor, perceived ease of use on the second factor, intention to return on the third factor, and attitude on the fourth factor, with all factor loadings exceeding .7. Thus evidence supports construct validity for the scales adopted in this study.
No significant correlation was found between participants' age, gender, or time spent on the Web and the two dependent variables--attitude and intention to return. These demographic variables were dropped from further analysis.
Separate linear regression analyses were conducted based on 56 completed surveys collected from the study. In testing Hypothesis H1, a regression analysis was performed, with perceived ease of use as an independent variable and perceived usefulness as the dependent variable. Results indicate that perceived ease of use had a significant influence on perceived usefulness, with adjusted R-squared value at .63 and p<.01. Thus Hypothesis H1 was supported.
Hypotheses H2 and H3 were tested by regressing perceived ease of use and perceived usefulness on attitude toward using the site. Perceived usefulness had a significant influence on attitude, with p<.01, while perceived ease of use did not turn out to be a significant factor in attitude, with p>.05. This model explained 62% of the variance (adjusted R-squared) in attitude toward using the companion site. The R-squared and F values indicate a good fit of the model. The variable inflation factor (VIF) values for the two independent variables in this model were both at 2.772, less than the suggested cutoff point of 5 (Studenmund, 2001) or 10 (Neter, Kutner, Nachtsheim, & Wasserman, 1996), indicating no concern for multicollinearity. Standardized residuals were normal, and scatter plots of standardized residual versus predicted value showed no particular pattern, indicting no heteroscedasticity. Thus Hypothesis H3 was supported, and Hypothesis H2 was not supported.
Hypotheses H4 and H5 were tested through a third regression model, with perceived usefulness and attitude regressing upon intention to use. To control for the indirect influence of perceived usefulness on intention to use through attitude, a stepwise regression was performed in which perceived usefulness was entered in step one and attitude was entered in step two. Results show that both variables are significant predictors of intention to use, with perceived usefulness being significant at p<.01, and attitude toward using being significant at p<.05. The model explained about 59% (adjusted R-squared) of variance in intention to use. Once again the model had good fit as reflected in its R-squared and F values. VIFs were at 2.675 for the two independent variables in this model, lower than the suggested cutoff point of 5 (Studenmund, 2001) or 10 (Neter et al., 1996), indicating no concern for multicollinearity. Standardized residuals were normal, and scatter plots of standardized residual versus predicted value gave no indication of heteroscedasticity in the data. Thus both hypotheses H4 and H5 were supported in this study.
Hypothesis H6 was tested by regressing intention to use on the collated self-reported actual usage data collected in a follow-up survey from participants who were then taking the course. Of the 56 originally surveyed, 39 were able to report their actual usage data. Thus this model was performed on 39 data points. Results show that 23% (adjusted R-squared) of the variance in actual usage was explained by intention to use, and that intention to use had a significant influence (p<.01) on actual use. Thus Hypothesis H6 received support in this study. This result might be of most practical implication among all hypotheses tested. It confirmed in our current context what has been long suggested in the theory of reason actions (TRA) that behavioral intentions have a strong effect on actual behavior. This result was also consistent with prior research in TAM applications. Table 3 summarizes the results of our hypotheses testing.
Results of this study suggest that TAM offers a useful tool in evaluating hypermedia-based course companion websites. All but one of the hypotheses proposed in this study, based on the relationships in the original TAM model, were supported. The only hypothesis that was not supported in this study was the relationship between perceived ease of use and attitude toward using, and thus a discussion of its implications is warranted.
TAM treats perceived usefulness and perceived ease of use as two distinct antecedents of attitude toward the use of a technology (Davis, 1989; Davis et al., 1989). However, findings from the first two applications of TAM (Davis; Davis et al.) showed that perceived usefulness was a significantly stronger factor than perceived ease of use. Davis et al. observed that when users learned to effectively use the system, the direct effect of ease of use on attitude and behavioral intentions disappeared. After two studies, Davis concluded that "no amount of ease of use can compensate for a system that does not perform a useful function" (p.333). Regression results showed that perceived ease of use might be an antecedent to perceived usefulness, rather than a construct parallel to perceived usefulness, and such that perceived usefulness mediates perceived ease of use (Davis). Adapting TAM to websites, Koufaris (2002) found a similarly much stronger significance of perceived usefulness over perceived ease of use on customers' intention to return to a website, where he found p<.01 for perceived usefulness and not significant for perceived ease of use on attitude.
In the current study, most of the participants were very familiar with the web browser and most of them spent a lot of time using the Internet for information and entertainment. The simple and user-friendly interface of a web browser significantly diminishes the direct impact of ease of use on user attitude. Since the web-based course companion site was viewed through this familiar browser interface, it is not surprising that we did not find the relationship between perceived ease of use and attitude significant. It is reasonable to suspect that when the interface is more complex and the stimuli more diverse, for example, multiple sites with diverse levels of ease of use, the perceived ease of use factor may be found to play a more significant role than that was found in this study.
Student sample was deemed appropriate in this study since the goal of this study was to examine the influences of perceptions on the use of hypermedia-based course support in an educational environment. However, educational hypermedia can be used in a wide range of institutions and how other populations might respond to questions on this survey is unknown and should be studied in future research. Another limitation of this study was the use of self-reported usage data, a measure often used in IS and other research, especially in cases where system log files on actual usage are not available. In this study, that objective data would be very difficult to obtain, if not impossible. Whenever possible, future research should strive to obtain objective usage data for enhanced accuracy. Future research should also apply TAM to other educational hypermedia than course companion sites to test its predictive power.
This study proposes and finds that TAM can be a valuable tool in the evaluation and selection of hypermedia-based educational products such as a course companion website. Prior research has shown that TAM provides parsimonious power in predicting user acceptance of technology usage. Though TAM is not a descriptive model, that is, it does not provide diagnostic capability for specific flaws in technology or systems, it can serve the purpose of evaluating competing products such as textbooks with companion sites from different publishers and predict system or product acceptability. TAM suggests that such evaluations can be done very cost-effectively because user perceptions of a technology are formed very early just after his or her initial exposure to the system. It can thus provide a valuable tool to educators who aim to incorporate hypermedia-based support in teaching and learning.
Opinion Survey of Course-Related Web Site Please circle the number that best indicates your agreement or disagreement with each statement. (Items were separated and mixed on actual survey.) Definitely Definitely Disagree Agree Perceived Ease of Use 1) I found the site easy to use. 1 2 3 4 5 6 7 2) Learning to use this site would be easy 1 2 3 4 5 6 7 for me. 3) My interaction with the site was clear 1 2 3 4 5 6 7 and understandable. 4) It would be easy for me to find 1 2 3 4 5 6 7 information at the site. Perceived Usefulness 5) Using this site would enhance my 1 2 3 4 5 6 7 effectiveness in learning. 6) Using this site would improve my course 1 2 3 4 5 6 7 performance. 7) Using this site would increase my 1 2 3 4 5 6 7 productivity in my course work. 8) I found the course website useful. 1 2 3 4 5 6 7 Attitude toward Using 9) I dislike the idea of using this course 1 2 3 4 5 6 7 website. (R) 10) I have a generally favorable attitude 1 2 3 4 5 6 7 toward using this site. 11) I believe it is (would be) a good idea 1 2 3 4 5 6 7 to use this site for my coursework. 12) Using this course website is a foolish 1 2 3 4 5 6 7 idea. (R) Intention to Use 13) I intend to use the site during the 1 2 3 4 5 6 7 semester (or when I take this course). 14) I will (would when I take this course) 1 2 3 4 5 6 7 return to this site often. 15) I intent to visit this site frequently 1 2 3 4 5 6 7 for my course work. * R - reversed item. Table 1 Scale Reliability Scale Cronbach's Alpha Perceived Ease of Use .8585 Perceived Usefulness .9022 Attitude toward Using .8170 Intention to Use .8382 Table 2 Factor Analysis Component 1 2 3 4 PRFRMNCE .773 -.659 .552 .371 EFFECTIVE .859 -.616 .393 .474 USEFUL .886 -.523 .466 .491 PRDCTIVE .883 -.567 .507 .577 EASYUSE .351 -.850 .503 .249 EZLEARN .409 -.894 .349 .292 EZINTRACT .682 -.790 .428 .321 EZFINDINFO .667 -.864 .474 .425 WILLRETRN .279 -.461 .845 .352 INTENTO .565 -.504 .789 .556 FREQUENT .688 -.477 .811 .588 NTFOOLISH .149 -.334 .625 .815 NTDISLIKE .458 -.276 .317 .880 GOODIDEA .741 -.490 .485 .849 FAVORABLE .659 -.492 .479 .704 Extraction Method: Principle Component Analysis Rotation Method: Oblimin with Kaiser Normalization Table 3 Summary of Hypotheses Testing Hypothesis Relationship Tested Results H1 H1: Perceived ease of use is Supported (p<.01) positively related to perceived usefulness. H2 H2: Perceived ease of use is Not Supported positively related to attitude (p>.05) toward using. H3 H3: Perceived usefulness is Supported (p<.01) positively related to attitude toward using. H4 H4: Perceived usefulness is Supported (p<.01) positively related to intention to use. H5 H5: Attitude toward using is Supported (p<.05) positively related to intention to use. H6 H6: Intention to use is positively Supported (p<.01) related to actual use.
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice-Hall.
Davis, F.D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13, 319-340.
Davis, F.D., Bagozzi, R.P., & Warshaw, P.R. (1989). User acceptance of computer technology: Comparison of two theoretical models. Management Science, 35(8), 982-1003.
Fishbein, M., & Ajzen, I. (1975). Belief attitude, intention and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley.
Hu, P.J., Chau, P.Y.K., Sheng, O.R.L., & Tam, K.Y. (1999). Examining the technology acceptance model using physical acceptance of telemedicine technology. Journal of management Information Systems, 16(2), 91-112.
Koufaris, M. (2002). Applying the technology acceptance model and flow theory to online consumer behavior. Information Systems Research, 13(2), 205-223.
Mathieson, K. (1991). Predicting user intentions: Comparing the technology acceptance model with the theory of planned behavior. Information Systems Research, 2(3), 173-191.
Morris, M.G., & Dillon, A. (1997). The influence of user perceptions on software utilization: Application and evaluation of a theoretical model of technology acceptance. IEEE Software, 14(4), 56-75.
Neter, J., Kutner, M.H., Nachtsheim, C.J., & Wasserman, W. (1996). Applied liner statistical models (4th ed.). Burr Ridge, IL: Irwin.
Nunnally, J.C. (1978). Psychometric theory. New York: McGraw Hill.
Pedhazur, E.J., & Schmelkin, L.P. (1991). Measurement, design, and analysis: An integrated approach. Hillsdale, NJ: Lawrence Erlbaum.
Psaromiligkos, Y., & Retalis, S. (2003). Re-evaluating the effectiveness of a web-based learning system: A comparative case study. Journal of Educational Multimedia and Hypermedia, 12(1), 5-20.
Studenmund, A.H. (2001). Using econometrics: A practical guide (4th ed.). Reading, PA: Addison-Wesley.
Szajna, B. (1996). Empirical evaluation of the revised technology acceptance model. Management Science, 42(1), 85-89.
Ramapo College of New Jersey