Designing user interfaces to knowledge-based systems: an empirical investigation of information presentation and mental representations in human-computer problem solving

Human beings are characterized by individual differences in how they perceive, evaluate, and process computer-displayed information for problem solving. Knowledge-based systems (KBS) permit various internal representations of external, problem-specific knowledge. However, KBS are relatively inflexible in presenting information to users in ways which capitalize on individual differences in the use of mental model imagery and memory structures. An experiment is conducted which examines whether tree diagram and/or verbal KBS user interface formats interact with spatial ability in impacting problem solving speed and accuracy, satisfaction, and confidence. A KBS hypertext user interface medium is also tested. The Group Embedded Figures Test (GEFT) is used to measure individual spatial ability. A within-subjects design is utilized to assess dependent variable measures for each subject while using the three interfaces.