CogTool-Explorer: a model of goal-directed user exploration that considers information layout
Proceedings of the SIGCHI Conference on Human Factors in Computing Systems
| Hi-index | 0.00 |
This research has two important sets of purposes. The first set of purposes was to evaluate working memory load and persistent behavior of computer users navigating complex websites, design and test the effectiveness of feedback intended to reduce revisits to previously-visited (dead-end) states, and determine the extent to which reducing revisits improves overall web navigation performance on difficult search tasks. The second purpose was to contribute to the development of the Cognitive Walkthrough for the Web (CWW), verifying and refining the psychological validity of CWW predictions of search difficulty. Participants in the primary experiment completed 36 tasks that varied in difficulty on an experimental website simulating an actual online encyclopedia. The between-subjects independent variables were link color change (no link color change, cumulative link color change, task-specific link color change), heading feedback (heading feedback, no heading feedback), and order (original, revised). The within-subject independent variables were search difficulty (difficult, easy), and task set (first-half, second-half), the latter being a measure of learning effects. Revisits to subordinate links — a clear, direct measure of working memory errors and perseverance — was the main dependent variable.Results showed that task-specific link color change dramatically reduced revisits to subordinate links, and there was a significant interaction of task set with link color change. In the no-link-color condition, individuals varied dramatically in both mean revisits and maximum revisits, and certain individuals, possibly with low working memory capacity or tendency to perseverate, derived the most benefits from task-specific link color change. However, even task-specific feedback reduced clicks on links by only about a half click, and tasks that CWW predicts to be difficult are not significantly reduced in difficulty. The three independent variables that CWW uses to predict search difficulty and usability problems and predicted clicks on links turn out to equally valid for experimental websites with no link color change and for real-world websites that do provide link color change feedback.