Evaluative metadata in educational digital libraries: How users use evaluative metadata in the process of document selection.

摘要

The purpose of this research was to contribute to an important research and practice issue, how users use evaluative metadata (i.e., end-user ratings, annotations, opinions, usage, and experts' recommendations) when searching for documents in educational digital libraries. The dissertation built a model of the predictive judgment process to develop design principles for digital library systems with evaluative metadata.;Multiple methods were employed in the research design: (1) a questionnaire gathered demographic information, the subjects' computer/Internet/web knowledge, information needs (tasks), search keywords, and topic knowledge; (2) a think-aloud protocol was used to collect data about cognitive processes, in this case document selection; (3) semi-structured interviews gathered information about subjects' thoughts about their behavior; (4) the researcher wrote field notes (observation); and (5) Camtasia software was used to record the trace of each subject's metadata usage.;The participants were 14 graduate students from School of Library and Information Studies in Florida State University. Four tasks were given to them in a laboratory setting. Tasks were two types 1) for finding teaching materials as an instructor, and 2) study materials as a student. The analyses were based on data from surveys, searching experiments, and post interviews. The data included 366 reviewed items in the search results of 60 searches. The results showed the user's positive inclination toward the evaluative metadata during searching in MERLOT. The research identified the reasons for scanning and examining evaluative metadata elements in the predictive judgment process. In addition, three ways to use evaluative metadata in the predictive judgment process were showed. The research also found the factors which influence the usage of the evaluative metadata: Users' cognitive, Function of the Digital Libraries, Situational, and Task. Finally, the revised predictive process model was presented including three stages of the predictive judgment (Scanning/Examining, Judgment, and Decision), the factors which affect the usage of evaluative metadata, and relevance criteria when the participants searched documents in MERLOT. The findings of the research provide implications for advancing empirical research for evaluative metadata and user-centered relevance study, and improving the design of the educational digital libraries.