The Effect of Alternative Procedural Explanation Types on Procedural Knowledge Acquisition During Knowledge-Based Systems Use

Georgia Smedley
University of Nevada, Las Vegas

Steve G. Sutton
University of Central Florida

 
ABSTRACT:

This study applies ACT-R, a theory of cognitive skill acquisition, to identify two techniques theorized to provide learners with a simplified and situation-responsive set of production rules to use in a problem-solving context.  The two techniques are abstraction (an optimization technique that produces a generalized rule set) and goal structuring (another optimization technique that produces a differentiated rule set).  Accordingly, abstraction and goal-structuring explanations were provided to users through a KBS. Due to cognitive effort constraints on procedural learning, a subset of volunteer participants was extracted for analysis based on an exhibition of attentive learning behavior. Results of the study found that while intermediary stages of development were not detectable, participants receiving goal structuring explanations exhibited better problem solving performance and the joint presentation of abstraction explanations led to further problem solving improvements. Abstraction explanations did not lead to improved problem solving, however, when provided in absence of goal structuring explanations.  These findings extend ACT-R to a new venue, increase understanding of ACT-R theory, and provide developers of KBS with more substantive knowledge on optimization of KBS explanation design when knowledge transfer to less expert users is an objective. 

Key Words: knowledge based systems, intelligent systems, expert systems, intelligent decision aids, intelligent decision support systems, explanation facilities, knowledge transfer, learning theory, procedural knowledge, abstraction, goal structuring.

Data availability: Data from this study along with programs and case materials can be attained through the first author within the limitations of human-subjects guidelines.

Back