The longitudinal approach to the study of information seeking has received new impetus through a recent integration of constructivist personality theory and longitudinal, in-depth study of students involved in course-integrated writing projects (Kuhlthau, 1993). Longitudinal monitoring of the affective reactions of novice searchers makes available the empirical data necessary to discover patterns of information seeking behavior. For instance, Kuhlthau's data point to six stages of affective behavior through the course of a single writing project: uncertainty, optimism confusion, clarity, confidence, and satisfaction. This kind of data is possible only with repeated probes throughout the behavioral unit being studied, such as a semester, a training period, or a search session or episode.
An important role of the affective domain in the information behavior of novices was discovered when it was found that search style and acceptance of instructions were both influenced by self-efficacy judgments expressed prior to the search (Nahl, 1993a,b; Nahl & Harada, 1996; Nahl, 1995). In other words, novices who expect to be successful at a search task are more efficient and adaptive searchers than those who express doubt and a lack of self-confidence in their ability to carry out successful searches. Earlier studies on computerphobia showed a similar effect (Busch, 1995).
According to Seligman, (1991) "explanatory style" is the most salient distinction between optimists and pessimists, or between individuals with high vs. low self-esteem. For example, an individual faced with some unsolved problem can think "I'm stupid" (the hopeless style), or "I'm hung over" (the hopeful style). While searching, a novice can think, "I'm not good at this" or "I don't do well at this." These beliefs may have maladaptive consequences at both the affective and cognitive levels, leading to less functional overt actions such as abandoning a search too soon, avoiding getting help, failing to take notes, or keying inaccurately. Research shows that the strength of one's self-efficacy judgment prior to learning a task has an influence on how much effort one then expends or how persistent at the task one is (Barling & Beattie, 1983; Wood & Bandura, 1989). The negative consequences of low self-efficacy judgments have been documented in recent research on computer avoidance and technophobia Olivier, 1993; Jorde-Bloom, 1988; Rosen, 1993.
The effect of self-efficacy on performance is thought to operate through "outcome expectations" influencing effort, persistence, and mastery. Simply put, if you expect less of yourself, you expend less effort. According to Bandura, (1986, 391), self-efficacy "is concerned not with the skills one has but with judgments of what one can do with whatever skills one possesses." According to Compeau & Higgins, (1995, 192), "Individuals with a weak sense of self-efficacy will be frustrated more easily by obstacles to their performance and will respond by lowering their perceptions of their capability." Other motivational concepts related to self-efficacy include self-confidence, self-esteem, effort-performance expectancy, and locus of control (Compeau & Higgins, 1995; Nahl, 1993). In a prior study, novices were asked to rate the likelihood of their search success prior to a CD-ROM search task. The group was divided at the median: those who had higher self-efficacy judgments significantly outscored the pessimists on all dependent measures including, efficiency, success, satisfaction, interactivity, and valuing documentation (Nahl, 1993). The results were replicated in two simulated search experiments in which college students (Nahl, 1995) and high school students (Nahl & Harada, 1996) predicted the correctness of their answers on a quiz that tested concept analysis, search term selection, and Boolean reasoning. The current study examines the influence of self-efficacy judgments of students enrolled in a psychology Internet course.
A recent study on the needs of novice Internet learners reminds trainers to "remember the affective elements of training" to insure "productive learning experiences" (Hert, Rosenbaum, Skutnik & Backs, 1995, 67). The affective domain is hypothesized to influence both the acquisition of component skills (motivation). and their integration during execution of the overall task or activity. Self-efficacy judgments are considered affective reactions, and since they take on a variety of situational characteristics, it was necessary to use a probe technique with several affective rating scales. Prior experience with Internet learners influenced the selection of six affective components to measure: (1). degree of difficulty experienced; (2). amount of negative emotions experienced; (3). how valuable are the skills acquired; (4). likelihood of getting good at it; (5). degree of satisfaction; and (6). degree of effort (how hard I tried).
Method
The data for this experiment are based on the weekly self-reports submitted by senior college students enrolled in a seminar on "The Social Psychology of Learning the Internet."[1] As part of their weekly report, students filled out a questionnaire with six varieties of affective responses on a 10-point scale. For instance:
How difficult was this week's task (lumping all the subtasks together)? Circle one.
Very easy 1 2 3 4 5 6 7 8 9 10 Very hard
Six affective areas were probed. Besides difficulty, the following were assessed:
How much negative emotion did it cost you, all in all? (Very little/Very much)
How valuable for later use is this knowledge or skill going to be for you? (Not useful/Very useful)
How likely is it that you'll be getting good at these tasks? (Not likely/Quite likely)
How satisfied are you with the computer and the Internet systems? (Not satisfied/Very satisfied)
How hard did you try to get through this week's Internet tasks? (Gave up easily/Refused to give up)
Students also wrote explanations for their ratings. The data were collected weekly during the Internet training period ending after week 8. The course continued with its topics and Internet tasks through week 16. During the final week, students filled out the same questionnaire; this time globally applying it to "this semester's work" instead of "this week's tasks." Thus, their average weekly ratings during the eight-week training period were compared with the end-of-course ratings.
Eighteen students were present in the first class (Day 1). during which they filled-out a questionnaire designed to probe the strength of their self-efficacy judgments, or feelings of self-confidence as a searcher (Nahl, 1993a,b; Nahl, 1995; Nahl & Harada, 1996). Apparently, strong positive self-perceptions and feelings of self-confidence are important affective skills which empower searchers to become more motivated, more persistent, more accurate, more efficient, better able to learn search skills and effective problem solving skills. In other words, individuals who avail themselves of these affective resources are more likely to succeed than those who don't. In this case, three affective probes were constructed to obtain an objective index of their self-efficacy strength. These were:
How likely is it that this semester
(a). I can become good at surfing the Internet?
Not likely 1 2 3 4 5 6 7 8 9 10 Quite likely
(b). I can find it easy and fun to surf the Internet?
Not likely 1 2 3 4 5 6 7 8 9 10 Quite likely
(c). I can competently use Internet facilities such as electronic mail, file transfer protocol, telnet, and World Wide Web HTML scripting?
Not likely 1 2 3 4 5 6 7 8 9 10 Quite likely
These data were obtained at the end of the first class, during which students were given a precise description of the skills they were expected to acquire in eight weeks. This no doubt involved them in thinking about the effort that they would have to expend to meet these criteria. At the same time, students were encouraged and given hope by the assertion, repeated several times, that no one had failed in the past who stuck with it through the eight weeks. This was meant as a challenge to their inner motivation to overcome obstacles.
A questionnaire given at the beginning of class on Day 1 indicated that none of the students in the class had any Internet experience. All were familiar with word processors and online catalogs. The goal of the eight-week training period was to have students acquire the ability to upload their weekly typed reports and post them on their own self-constructed World Wide Web home pages. The sub-component skills involved learning UNIX file management, learning to upload files from a diskette, and learning how to insert HTML code using an online editor to edit and test live hyperlinks. The sequence of weekly assignments was as follows:
Week 1: Using Netscape Navigator for the First Time
Week 2: Using Netscape for the Second Time
Week 3: Logging-in to UNIX and sending e-mail with Pine for the First Time
Week 4: Using FTP and TELNET for the First Time
Week 5: Putting up my Home Page on the World Wide Web
Week 6: Beautifying my Home Page and Linking it to the World
Week 7: Exploring and Evaluating the Home Pages of my Classmates
Week 8: Constructing a Hypertext Index for my Home Page Site
By tabulating the data for each week within each affective area, one obtains a visible picture of the longitudinal pattern of affect which these novice learners experienced. Tracking affective changes over time can help identify the learning dynamics involved.
Eighteen students began the course and 12 finished it. All those who completed the eight-week training segment went on to the end (Week 16). and obtained a grade of A or B. The six students who did not complete the training period failed the course and none of them continued to attend class after week 5. The results on self-efficacy include the 6 drop outs (total N=18). while the results on longitudinal weekly probes do not (total N=12).
Results
Data on Self-efficacy Judgments
Table 1 shows that all three affective probes discriminate well and predict accurately who would be successful and who would fail in Internet training. In each case, as well as combined, these self-efficacy predictions are significantly lower for those students who eventually dropped out in comparison to those who went on to complete the training segment and the course. There was no significant difference between the two groups in computer experience, which was at the novice level for everyone in this course.
Table 2 explores the self-efficacy probes in another way using Chi Square. The 18 students who filled out the self-efficacy predictions on Day 1 were divided into upper and lower half of scores separately for each scale, and then for the combined, as shown in Table 2. The upper half ("high" group). and the lower half ("low" group). in each case divided at score 7.00 (7.33 for the combined). on the 10-point scale. Table 2 shows that the vast majority of those who were successful and completed the course had a relatively higher self-efficacy judgment. All those who had a relatively low self-efficacy judgment on Day 1 were not successful and became drop outs.
Data on Six Affective Probes for Weeks 1 to 8
The 12 students who completed the eight-week training period exhibit weekly changes that correspond to the complexity of the tasks for that week. As indicated above, the weekly tasks varied in difficulty. There were four blocks of two weeks each: Block 1 (weeks 1 & 2). was relatively simple as it involved merely using Netscape at the computer lab simply by double clicking on its icon. Block 2 (weeks 3 & 4). was more complex requiring students to logon to UNIX, send e-mail, and learn telnet and ftp software. Block 3 (weeks 5 & 6). was also complex asking students to create their own home pages using models that were provided, and which they were allowed to copy and adjust by adding their own content. Finally, Block 4 (weeks 7 & 8). was relatively simple asking students to explore and comment on the home pages of fellow students and to expand their own.
Figures 1 through 6 exhibit the results using a repeated measures, one-way ANOVA to assess the significance of the changes over the eight-week training period for each of the six affective probes. The degree of difficulty the novices felt each week seems to be a reaction to timing and content of the activity being learned. There were four types of activities in this program. The first two weeks were planned to be relatively easy, namely, using Netscape Navigator to browse the World Wide Web. Since all settings were pre-installed in the computer lab, all this required was to double click on the big N icon to go online. Still, students found this to be moderately difficult since it involved going to a novel technological place and sitting by oneself, figuring out what to do. The next two weeks (3 & 4). were much more difficult in comparison, as students had to come to grips with the heart of being an online user as a Web author and publisher: logging on to UNIX, manipulating directories and files, doing e-mail with Pine, uploading a file with ftp software, and using telnet to access one's home page directory. Felt difficulty doubled during this period and reached a maximum.
The next block (weeks 5 & 6). was relatively easy again even though it involved the task of putting up a home page. No doubt things were made safer for students by allowing them to copy available models and personalize it with their own content. Finally, the last block (weeks 7 & 8). was the easiest even though it involved learning some of the complexities of HTML code, including making hypertext links that worked, learning text formatting and how to achieve visually pleasing effects with backgrounds and icons. In addition, everything that was done in earlier weeks had to be repeated and continued each week, while performing new tasks for that week. Despite this, there was a steady decrease of felt difficulty over time and cumulative practice. It appears that, for students with initially high self-efficacy judgments, confidence builds over the weeks and they are able to accomplish ever more complex tasks, while continuing to find them less and less difficult.
The amount or intensity of negative emotions reported for each week paralleled closely the amount of felt difficulty, as can be seen by comparing the two curves in Figs. 1 and 2. In both cases, the changes over the eight weeks were significant. It is clear that, in this context, negative affect associated with learning the Internet rose at first to a peak, then gradually but significantly decreased to a globally low level.
There seems to be a uniformly positive value attached to the skills acquired each week as shown in Fig. 3 (no significant changes according to the ANOVA). The positive value expressed remains near the maximum each week, irrespective of the negative affect experienced due to difficulties. Part of this success was no doubt due to weekly class meetings during which students were able to commiserate and encourage each other, and during which the overall value of the learning was continually reinforced by the instructor. Especially influential, according to the instructor, was his emphasis on a generational framework in which students can see the work of prior semesters and emulate it, knowing that they all start at the same novice level.
Students were generally optimistic about the likelihood of achieving mastery with the Internet (Fig. 4). However, they experienced small but significant set backs during the second and third blocks (weeks 3 through 5), during which they experienced the most difficulty and negative emotions as they were trying to cope with the nitty gritty of telnet, ftp, and UNIX and HTML. The set back in self-confidence was temporary, and by the end of the training period, their self-confidence in "getting good at it" was the highest ever.
Students reported an overall increase in their satisfaction with the Internet as a system, from beginning to end of the training period (Fig. 5). Satisfaction was significantly stronger in the second half of the eight-week period, though the difference is modest, being uniformly high. Small differences can nevertheless be detected. In the weeks of greatest difficulty and least self-confidence (3, 4 and 5), satisfaction with the Internet as a system was somewhat less. There is obviously a congruence between the various indicators of affect in this learning situation. This evidence supports the notion that the cluster of affective elements that are activated in a particular situation are not independent but interdependent, being arranged in a hierarchy of levels and categories. Taxonomic models of affect are explicit representations of such situation-based arrangements of feelings people normally experience (Jakobovits & Nahl-Jakobovits, 1987; Jakobovits & Nahl-Jakobovits, 1990; Nahl, 1993; Nahl, 1996; Nahl & Tenopir, 1996; Nahl-Jakobovits & Jakobovits, 1993).
The last of the affective measures assesses students' overall motivational resource as learners during each week, according to their own feelings. As Fig. 6 shows, students experienced a uniformly strong feeling that they were trying as hard as they could. There was no significant decrease of this feeling throughout the eight weeks of learning. When this pattern is considered along with the other affective measures, the dynamics of feelings that is part of learning the Internet is quite complex. During the difficult and relatively negative weeks (3, 4 & 5), when self-confidence was threatened, satisfaction with the system was lower, but perceived effort did not slacken, and value remained near the maximum.
End-of-semester Ratings
The 12 students who completed the training period and the course were given the same questionnaire to fill out on the last day of the course (Week 16). These results are presented in Table 3. The interest here is to contrast the average affective rating for the training period (weeks 1 to 8). with the overall assessment for the entire course given at the end (weeks 1 to 16). The students rate the overall course assessment more favorably than the training period for all six measures. However, only three of the comparisons are statistically significant, while in three other cases the trend is more or less ambiguous. Further research is needed to establish whether there is indeed a tendency for learners to mitigate the importance of negative experiences during the training period, and to see them as less severe in hindsight. The purpose of this study was to create a condition for identifying the repertoire of affective and cognitive behaviors of novice users learned to adapt to Internet use. Therefore, no attempt is made to generalize from this small sample to the population in terms of specific effects or frequency of occurrence of certain events. As a result, future research will have to determine the distribution and occurrence of the patterns identified in this study. For instance, exploring whether particular user groups show different patterns.
Discussion
The affective dynamics uncovered through these weekly self-witnessing reports is consistent with the cognitive and sensorimotor outcome of the training course. At the end of eight weeks every one of the 12 students achieved the minimal standards of objectives set for the course despite the fact that none of them had any computer or telecommunication experience beyond word processing and online library catalogs. The skills acquired by every student at the end of the eight weeks included:
* direct and telnet logon
* UNIX file management and directory navigation
* file uploading with ftp software
* using an online editor to enter HTML code
* maintaining a home page Site on the World Wide Web
* navigating the Internet and using search engines to find
specific information
There were individual differences among the students in terms how complex and creative their output was beyond these minimum skills. This was especially apparent with those who confessed to becoming enthralled with, or in a few cases, "addicted" to the Internet. Those who became involved in this intense way, reported devoting upwards of 30 hours a week in comparison to the majority, who put in the minimum expected 9 hours. Two patterns of longitudinal adaptation to the Internet are apparent in the results. Figures 1 and 2 show an inverted-U effect similar to error curves in learning trials. In this case, experiencing difficulty and negative emotions increased to a peak during the third week of training, then gradually decreased till the end. It is a picture of adapting to the learning situation. Figures 3 through 6, on the other hand, are less variable and remain uniformly high throughout the training phase. These measures relate to effort, satisfaction, and outcome expectation. Since these patterns reveal the psychological dynamics of successful Internet learners, they can be useful to managers of information environments. For instance, it is normal for successful Internet learners to experience increased difficulty and negative emotions for the first few weeks, until the challenging peak, after which the trend reverses. For successful learners, the value and satisfaction ratings remain uniformly positive, even during highest difficulty and uncertainty. At the same time, self-efficacy judgments start and remain positive throughout the training experience.
Improving the Affective Learning Environment
Knowing about psychlogical patterns can help information managers and instructors to design better programs and more user-centered environments. The importance of maintaining positive self-efficacy judgments is shown in two ways. First, those who have a less positive initial self-efficacy perception can be overwhelmed and end up dopping out. Second, those who have a more positive initial self-efficacy perception maintain this perception throughout the program, all the way to success. Nahl-Jakobovits and Jakobovits (1985) describe several principles for the micro-management of the affective environment of information users, one of which is social facilitation. In the present study, the instructor encouraged social facilitation effects in various ways, including weekly group discussions on overcoming information anxiety, weekly self-monitoring of feelings, and a system of rotating partners in which pairs of students worked together on team tasks, teaching each other.
Information managers and instructors can devise appropriate social facilitation procedures that include two important features suggested by the data in this study: (1). treating learners as a group, and (2). encouraging learners to monitor their affective reactions during training. By treating learners as a group, individuals are exposed to each other's reactions and problems, which helps them in their effort to overcome self-doubt and pessimistic self-judgments in the face of initial failure. By encouraging learners to do regular affective self-monitoring, individuals can treat their self-perceptions as objective data. If they choose to share some of these self-perceptions in a group, as was done in this class, individuals are exposed to positive feedback from the instructor and encouragement and help from the rest of the group.
Where this kind of group procedure is not appropriate or convenient, one could consider online and archival social facilitation procedures which might include listservs, MUDS or MOOS, FAQ files, generational records of prior classes or groups, affective instructions containing advice and reassurance, an inventory of common errors with strategies for avoiding and correcting them, and information self-counseling hints. Indeed, Internet system design is moving in the direction of providing more interactive facilities that can be incorporated into training. The frontier area of affective computing intends to develop a wearable computer that "attends to you during your waking hours, could notice what you eat, what you do, what you look at, and what emotions you express." and use this data to help you cope on your daily round (Picard, 1995, 14). The system would sense a variety of biological and behavioral data in the wearer, including affective data, such as interest level, arousal, mood, etc., combining biological information with a profiling scheme that is continuously updated by incoming data. One can imagine how such systems might provide coaching services to assist novices learning complex systems such as the Internet, e.g., helping the learner to stay on-task; providing support to overcome barriers; providing situational help with procedures, strategy, and design; helping avoid over-dramatizing routine problems, helping build confidence, helping achieve a sense of accomplishment and mastery; etc. In the current environment, group processes and self-monitoring of progress hold promise for humanizing information technology instruction.
REFERENCES
Bandura, A. (1986). Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, NJ: Prentice- Hall.
Barling, J., & Beattie, R. (1983). Self-efficacy Beliefs and Sales Performance. Journal of Organizational Behavior Management, 5, 41-51.
Busch, T. (1995). Gender Differences in Self-Efficacy and Attitudes Toward Computers. Journal of Educational Computing Research, 12(2), 147-58.
Compeau, D. R., & Higgins, C. A. (1995). Computer Self-Efficacy: Development of a Measure and Initial Test. MIS Quarterly, June, 189-211.
Hert, C. A., Rosenbaum, H., Skutnik, S. C., & Backs, S. M. (1995). Information Needs and Uses During Internet Training. ASIS '95 Proceedings of the 58th ASIS Annual Meeting, 32, 61-68.
Jakobovits, L. A., & Nahl-Jakobovits, D. (1987). Learning the Library: Taxonomy of Skills and Errors. College & Research Libraries, 48(3), 203-214.
Jakobovits, L. A., & Nahl-Jakobovits, D. (1990). Measuring Information Searching Competence. College & Research Libraries, 51(5), 448-462.
Jorde-Bloom, P. (1988). Self-Efficacy Expectations as a Predictor of Computer Use: A Look at Early Childhood Administrators. Computers in the Schools, 5(1), 45-65.
Kuhlthau, C. C. (1993). Seeking Meaning: A Process Approach to Library and Information Services. Norwood, New Jersey: Ablex.
Nahl, D. (1993). CD-ROM Point-of-Use Instructions for Novice Searchers: A Comparison of User-Centered Affectively Elaborated and System-Centered Unelaborated Text. Unpublished Ph.D. Dissertation, University of Hawaii, Honolulu Full text available on the Word Wide Web at http://www2.hawaii.edu/slis/nahl/nahl.html.
Nahl, D. (1995). Affective Elaborations in Boolean Search Instructions for Novices: Effects on Comprehension, Self-Confidence, and Error Type. ASIS '95 Proceedings of the 58th ASIS Annual Meeting, 32, 69-76.
Nahl, D., & Harada, V. (1996). Composing Boolean Search Statements: Self-Confidence, Concept Analysis, Search Logic, and Errors. School Library Media Quarterly, (in press).
Nahl, D., & Tenopir, C. (1996). Affective and Cognitive Searching Behavior of Novice End-Users of a Full-Text Database. Journal of the American Society for Information Science, 47(4), (in press).
Nahl-Jakobovits, D., & Jakobovits, L. A. (1985). Managing the Affective Micro-Information Environment. Research Strategies, 3(1), 17-28.
Nahl-Jakobovits, D., & Jakobovits, L. A. (1993). Bibliographic Instructional Design for Information Literacy: Integrating Affective and Cognitive Objectives. Research Strategies, 11(2), 73-88.
Olivier, T. A. & Shapiro, F. (1993). Self-Efficacy and Computers. Journal of Computer-Based Instruction, 20(3), 81-85.
Seligman, M. E. P. (1991). Learned Optimism. New York: Alfred A. Knopf.
Wood, R., & Bandura, A. (1989). Social Cognitive Theory of Organizational Management. Academy of Management Review, 14(3), 361-384.
