A Cognitive Model for Problem Solving in Computer Science

According to industry representatives, computer science education needs to emphasize the processes involved in solving computing problems rather than their solutions. Most of the current assessment tools used by universities and computer science departments analyze student answers to problems rather than investigating the processes involved in solving them. Approaching assessment from this perspective would reveal potential errors leading to incorrect solutions. This dissertation proposes a model describing how people solve computational problems by storing, retrieving, and manipulating information and knowledge. It describes how metacognition interacts with schemata representing conceptual and procedural knowledge, as well as with the external sources of information that might be needed to arrive at a solution. Metacognition includes higher-order, executive processes responsible for controlling and monitoring schemata, which in turn represent the algorithmic knowledge needed for organizing and adapting concepts to a specificc domain. The model illustrates how metacognitive processes interact with the knowledge represented by schemata as well as the information from external sources. This research investigates the didifferences in the way computer science novices use their metacognition and schemata to solve a computer programming problem. After J. Parham and L. Gugerty reached an 85% reliability for six metacognitive processes and six domain-specific schemata for writing a computer program, the resulting vocabulary provided the foundation for supporting the existence of and the interaction between metacognition, schemata, and external sources of information in computer programming. Overall, the participants in this research used their schemata 6% more than their metacognition and their metacognitive processes to control and monitor their schemata used to write a computer program. This research has potential implications in computer science education and software development through its understanding of the cognitive behavior used to solve computational problems