Understanding kiosk navigation: Using log files to capture hypermedia searches

The purposes of these studies were to investigate: a) the role of different intentions for constraining kiosk searches; b) whether selected variables related to reading comprehension (e.g., prior content knowledge) were also related to document search; and c) the potential of log files (timed-stamped records of students' navigational choices in the kiosk) as a non-intrusive tool for seamlessly capturing navigational performance. In Study 1, 64 college students were recruited and randomly assigned to one of two conditions. Both groups used an identical computer kiosk. However, one condition used the kiosk to address a specific problem, while the other condition used the kiosk based on their own interests-no specific task was provided. A discriminant function analysis was able to improve prediction about the group membership by 79% (adjusted with Cohen's Kappa) over and above chance. In Study 2, 54 participants were randomly assigned a simple or complex task. Participants were pretested on prior topic knowledge and solutions were rated based on a composite of navigational scores, Although prior content knowledge and the interaction were not significant, the type of task assigned was related to the quality of solution, These results provide further reason to distinguish between document search and reading comprehension. Further, they suggest new tools for assessment (log files) that have implications for many computerized applications, including hypertext, hypermedia, and the World Wide Web.