Discussion of Findings

Research findings over the past several years have led investigators to recognize the importance of table of contents as one of the effective navigation tools in hypertext system as well as to identify the various types of table of contents that must be considered in investigating the variable. In addition, the controversy over frame usage among World Wide Web application designers has not been clarified yet. The results of this study would have bearing on the pursuit of hypertext developing activities because they would suggest to the interface designers how to employ these variables in the design of WWW applications and hypertext as well.

Hypothesis 1

The first hypothesis was that users' performance in information searching and browsing with expandable TOC would be more accurate than with traditional TOC. Results of an analysis to test this hypothesis did not support it. However, a significant difference was found between users' performance under expandable TOC treatment and traditional TOC treatment. Users under traditional TOC treatment performed significantly better than those under expandable TOC treatment.

There are two possible explanations for the lack of significant effects of expandable TOC on users' performance over traditional TOC.

1. The average time spent interacting with the expandable TOC navigation tool was less than one hour for virtually all subjects. This short treatment period did not provide adequate exposure to this navigation tool. All subjects have reported themselves as intermittent or frequent users of WWW applications which can indicate that they tie-in with traditional TOC style rather than expandable TOC style. Subjects may have already developed a mental model of a WWW application in which all they had to do for the most operations was point, click, and scroll. It takes time for users to become familiar with the use of expandable TOC. The unfamiliarity of users with treatment has washed out most of the treatment effects. However, training could solve the unfamiliar nature of the tool. Having gained some familiarity with the various dimensions of a computerized environment, the user would become more comfortable with the design features of the interface. The researcher would expect this to happen if the individual continues to use and gain mastery of the same interface over some length of time frame.

2. Users employing expandable TOC complained about loading time. There were so many occasions on which users tried to expand or collapse the TOC and waited for a return set. Response time is one of the most critical issues in the friendliness of a hypertext system (Perrott & Smith, 1992). Akscyn, McCracken and Yoder (1988) suggest that a slow response time, for the selection of a link, can cause disorientation. Disorientation, in turn, can cause cognitive overload, and consequently degrades users' performance. A fast response time, for the selection of a link, would be an asset for a hypertext system because users would be more inclined to follow links and become experienced in the system. In order to optimize the response time, more sophisticated algorithm and effective and efficient programming language must be employed.

Hypothesis 2

The second hypothesis was that the users under expandable TOC treatment would have higher speed than those under traditional TOC treatment. The same rationale was followed in hypothesizing this hypothesis as in the first hypothesis. The results of an analysis did not support this prediction. However, a significant difference was found between users' speed under expandable TOC and that under traditional TOC treatment. A mean score of users' speed under traditional TOC treatment was significantly higher than that under expandable TOC treatment.

Two possible explanations for unexpected results are similar to those proposed for the first hypothesis.

Hypothesis 3

The third hypothesis was that the navigation would be more satisfied using expandable TOC as opposed to traditional TOC. To test this hypothesis, two dependent variables were employed: perceived degree of disorientation and perceived ease of use.

Perceived degree of disorientation. This hypothesis was predicted because Rada and Murphy's (1992) study presented that desirability rating was higher for SuperBook employing an expandable TOC. Lai and Waugh's (1995) postulation was also one of the bases for this hypothesis. They postulated that attitude towards the program should be negatively correlated to a sense of getting lost. It was, therefore, hypothesized that users under expandable TOC treatment would perceive less disoriented than those under traditional TOC treatment. Results of data analysis did not support this hypothesis. A significant difference between users' perceived degree of disorientation under expandable TOC treatment and under traditional TOC treatment was found, however. Users under expandable TOC treatment reported significantly more disorientation than those under traditional TOC treatment.

Similar to the first and second hypotheses, two possible factors influenced an unexpected result were slow loading time and users' unfamiliarity with the navigational tool. The findings of this current study support the suggestion of Akscyn, McCracken and Yoder (1988) who suggest that a low response time can cause disorientation. In addition, they also support Lai and Waugh (1995) who state that the unfamiliar interface in the hypertext system could increase users' sense of disorientation.

Perceived ease of use. In an attempt to exam users' satisfaction with the system, the users' perception of ease of use was employed. Davis, Bagozzi and Warshaw' s (1989) statement was one of the basis of this hypothesis. Davis et al. (1989) state that if the system is perceived as easy to use, it leads to the creation of positive attitudes about it. A system that is considered easy to use may be perceived as likely to save time and effort, thus increasing its perceived usefulness as well. It was, therefore, hypothesized that users under expandable TOC treatment will perceive easier to use the navigation tool than those under traditional TOC treatment. Once again, results of the analysis on users' perceived ease of use did not support this hypothesis. However, a significant difference was found between users' perceived ease of use under expandable TOC treatment and that under traditional TOC treatment. Users under traditional TOC treatment reported perceived it significantly easier to use the navigation tool than those under expandable TOC treatment.

Similarly, two important factors that caused users' disorientation were the system's loading time and users' unfamiliarity with a navigational tool, increasing the difficulties to use a navigational tool.

Hypothesis 4

The fourth hypothesis predicted that users' performance in searching and browsing for information with multiple-frame would be more accurate than with single-frame. Results of an analysis to test this hypothesis did not support it. The rationale behind this hypothesis was based on the concepts of external memory and mental models. The window or frame may serve as an external memory. This external memory function reduces the need on the user's cognitive system caused by the increasing complexity of computer screen interfaces (Billingsly, 1988). Windows or frames can mitigate this overload by keeping important and frequently used information. In this current study, frames were organized so that the table of contents would be visible simultaneously with the text of the selected node. This external memory not only reduces the need for the user's actions to move back and forth between the text frame and a table of contents frame, but also keep the user reminded about the conceptual structure of a hypertext system presented in an hierarchical structure of a table of contents. That is, it is not enough simply to try to show the user how the system is functioning beneath its opaque surfaces. A useful representation must be cognitively transparent in the sense of facilitating the user's ability to grow a productive mental model of relevant aspects of the system (Brown, 1992). Previous research has shown that more advanced mental models benefit a user's search (Borgman, 1986; Elkerton & Williges, 1984). These sophisticated mental models have been adapted from the structure of the system being searched. Increased similarity between a user's cognitive structure and the hypertext structure should aid search performance (Brannon, 1993). Edwards and Hardman (1989) also note that a better cognitive conceptualization of the hypertext document allows one to search better.

The results of an analysis show that there was no significant main effect of frame. In addition, users under single-frame treatment outperformed those under multiple-frame treatment. This indicates opposite directed outcome to the predicted hypothesis. However, the results of an ANOVA test for significance indicate an overall interaction between table of contents and frame. The significance of the interaction of these two main effects suggests that the provision of multiple-frame with traditional TOC can make users' performance better than the provision of multiple-frame with expandable TOC. Besides, the users under multiple-frame and traditional TOC treatment performed better than all the other three groups. So, it is reasonable to conclude that the effect of multiple-frame has been masked and indeed the task performance was poorer under the provision of expandable table of contents.

Hypothesis 5

The fifth hypothesis predicted that users under multiple-frame treatment would have more speed than those under single-frame treatment. The same rationale was followed in hypothesizing this hypothesis as in the fourth hypothesis. There was no difference between these two treatment groups. Thus, this hypothesis was not supported.

The results can be explained by the same reasons used to interpret the result obtained in the fourth hypothesis.

Hypothesis 6

The sixth hypothesis was that the navigation would be more satisfied with multiple-frame as opposed to single-frame. Two dependent variables were employed to test this hypothesis: users' perceived degree of disorientation and users' perceived ease of use.

Perceived degree of disorientation. The same rationales were followed in hypothesizing this hypothesis as those in the third and fourth hypothesis. It was, therefore, hypothesized that users would perceive less degree of disorientation with multiple-frame than with single-frame. The results of an analysis did not support this hypothesis. Users under multiple-frame treatment reported perceiving less degree of disorientation than those under single-frame treatment, but the difference failed to reach a significant level. However, a marginal significance was found between the interaction of table of contents and frame (p = 0.1395). The unexpected result can be explained by the same reasons used to interpret the results obtained in the fourth hypothesis.

Perceived ease of use. The rationales supported the creation of this hypothesis were similar to those of the third and fourth hypothesis. It was, therefore, hypothesized that users under multiple-frame treatment would perceive the use a navigation tool as easier than those under single-frame treatment. Similarly, the users under multiple-frame treatment reported perceiving it easier to use the navigational tool than those under single-frame treatment, but the difference was not significant. Thus, the analysis to test this hypothesis did not support it. Once again, the same reasons used to interpret the results in the fourth hypothesis can be employed for the explanation to this unpredicted result, as well.