Can students build cognitive models that reflect their own cognitive information processing? Results of preliminary class practice

Abstract

We have developed a learning environment to enable students to build rule-based cognitive models. Participants were required to build cognitive models that solve a cryptarithmetic task To solve this task, multiple types of procedural knowledge were used during the solution processes; participants were required to extemalize their knowledge when constructing a model. Twenty-two participants successfully constructed sophisticated models that were able to solve the given task in approximately 21 steps. We compared the participants' and their models' problem-solving behaviors. As a result, only 9 of the 22 (41%) models trance the participants' problem-solving paths. This implies that it was relatively difficult for the participants to construct a model that reflects their own cognitive information processing.