Model tracing systems analyse problem solving episodes and maintain a model of problem solving which is traced against the student’s activities. This approach was used by LISP TUTOR (Anderson and Reiser 1985).
At a given state of problem solving, rules which model the problem solving activities within the domain are selected and fired to predict the next state. Expert and bug rules are fired with the goal of matching the student’s new state. The student is assumed to be using the rule that predicted the new state and the student model updated to reflect this. Any deviation between the student’s solution and the systems can be acted upon and suitable action taken.
Authored by Serengul Smith
E-mail to:
serengul1@mdx.ac.uk
School of Computing Science Middlesex University
Revised: September 1998