ABSTRACT:

The growing use of expert systems (ESs) to improve decision making has raised significant research interest in the epistemological issue of these systems limiting the novice user's ability to learn how to make analogous or related decisions without the aid of the expert system. Prior research into this de-skilling has demonstrated that expert systems do possess the potential to serve as a teaching aid as well as a decision aid. The question then becomes one of how the user interface can be designed to facilitate the transfer of expertise to novices to avoid de-skilling. This study reports the results of a laboratory experiment that examined the impact of differences in elaboration- based explanations.

Three levels of elaboration type and two alternative elaboration placements were examined. Elaboration type was the amount of elaboration provided in the feedback, and elaboration placement was the point where the feedback was provided. Six different treatment groups each used one version of the expert system to evaluate the strength of internal control in a payroll system. Two types of knowledge, declarative and procedural, were measured to determine whether leaming had occurred. The results indicated that different elaboration placements and elaboration types significantly differ in the effect on the development of declarative knowledge. As the amount of elaboration provided increased, the knowledge transfer decreased. This effect was most prominent with the ESs that provided explanations at the end of the session, the traditional placement. With the ESs that provided explanations continu- ously, increased elaboration facilitated knowledge transfer initially up to a point of possible information overload. No effect was found for procedural knowledge.