From document clues to descriptive metadata: Document characteristics used by graduate students in judging the usefulness of Web documents

This study intends to explore what would be an effective means of identifying and describing the Web document encountered so that it can be evaluated more easily. The primary objective is to identify the salient document characteristics that enable users to better predict the usefulness of that document. The second objective is to examine if discipline-oriented metadata are needed in order to satisfy the needs of users across various disciplines.; This study assumes a user-centered approach. Thirty-eight graduate students were recruited from three distinct disciplines (English, Physics, and Sociology). Each participant was asked to conduct a Web search in response to his/her current information need. During the searching and evaluation process, participants were asked to highlight any text portion that affected their usefulness judgments and to talk about their thoughts underlying the actions concurrently. The entire process was audiotaped and videotaped. The audiotaped data were transcribed and content analysis was employed to analyze and code the data.; The results of this study revealed that participants employed their own reviewing sequences in examining the retrieved Web documents. They also had their own ways of interpreting the text portions identified within the documents. Participants could identify a document clue from anywhere within a document. Generally, the text of a document is the primary source of clues that help participants judge whether the document is useful.; Document characteristics identified by participants were grouped into eight broad trait groups or 46 trait sub-classes. Topic/Subject Area, Document Type, Scope/Coverage, Publishing Organization, and Focus/Key Issue of a document were the five document traits most frequently mentioned by participants. Still, any aspect of a document or any text portion identified within a document could be used by the participant in judging the usefulness of the document. The participant himself/herself is the best judge to evaluate the usage of the identified clues.; The results also showed that discipline had an effect on the selection and use of document clues. The frequency distribution of the mention of these eight trait groups across the three participant groups revealed variations. Additionally, some document traits might be mentioned more often by one participant group than by the other two groups.; The results of this study identified a range of document traits that users employed as clues to or indicators of the usefulness of the retrieved documents. It also empirically demonstrates that this is a feasible approach to explore what document characteristics can be used as metadata elements.