Adaptive Hypermedia Instructional Systems: Possibilities for Learner Modeling

A possible method of solving problems of knowledge representation and system adaptation in a hypermedia environment was examined. The method, based on the notion of semantic networks, uses the technique of knowledge mapping. In the process of reading a hypermedia document, the learner constructs a knowledge map (KM) specifying the relationships between concepts in the domain. The resulting map can be used to recognize learner misconceptions and identify potential instructional opportunities. The prototype KM system was tested with 36 undergraduates assigned to 1 of 3 groups using different versions of the system (browse, map, and relate), which allowed the learner to use a KM, construct a KM without feedback, or construct a KM with feedback. Subjects also completed paper and pencil pretests and posttests to identify the relationships between concepts introduced. All participants gained a significant amount of knowledge about the content. The browse group performed better on the short answer test, perhaps because participants in the other conditions were so involved in processing their KMs they did not study the concepts as well. A hypermedia system of this sort may provide an effective means of judging when a learner has attained some level of understanding of the domain. The system could then provide a new and appropriate level of detail for the learner. Four tables and five figures present study data. A 25-item list of references and 4 appendixes illustrating KMs and analysis procedures are included. (SLD)