Effect of Intelligent Instruction on Students' Mental Models and Performance

Abstract

The effects of two types of Intelligent Computer Aided Instruction (ICAI) and a non-intelligent tutoring system on the development of students' mental models and on performance was explored. One of the two ICAI systems provided both principle-based and task specific advice while the other provided task specific advice only. The non-intelligent CAI provided no advice. The problems presented to students increased in difficulty as the student progressed. It was hypothesized that the mental models of students who interacted with the ICAI that provided both principle-based and task specific advice would be most similar in structure to a “gold standard” model, and that these students would also exhibit the best performance. It was also hypothesized that students who were exposed to the ICAI that provided only task specific advice would perform better, and exhibit more accurate and detailed mental models than would students who interacted with the CAI that provided no advice. Study results indicated that students in both ICAI conditions achieved a higher similarity index, fewer errors, and faster task time, than did students in the no advice group. However, there was no significant difference between the two ICAIs for any of these measures. There was a significant interaction between version and problem difficulty for task time. As problem difficulty increased, task time increased for both of the two ICAI conditions, but not for the no-advice condition. These results are discussed in terms of their implications for the predictive validity of the Pathfinder measure, and in terms of the appropriate use and design of ICAI.