Preliminary Study on Learning by Constructing a Cognitive Model Based on Problem-Solving Processes

Abstract

Construction of models is promising as a learning activity, and computational environments are useful for that. However, it can be a heavy task for novice learners to construct computational models because it requires considerable instruction and practice of programming languages. We designed a basic framework for learning by experiencing construction of models on a production system in the domain of cognitive science. In this framework, a model abstractly describing human problem-solving processes and its computer model implemented on the production system is prepared by an instructor in advance. A learner is given the abstract model and processes of problem solving produced by executing the implementation model, and then engaged in instantiating the abstract model into an implementation model. This activity is expected to deepen learner understanding of mental processes embedded in the abstract model. We preliminary studied the effect of learning a model which simulates subtraction requiring regrouping in the framework. The results confirm the possibility that such experience can improve learner understanding of mental processes behind the model, and necessity to expand learning activities in the framework.