Collaborative Problem Solving Type Analysis Using Process Data in Simulation-based Assessment

This study aims to identify the types and characteristics of collaborative problem-solving skills among test takers in South Korea. For this, process data from the PISA 2015 collaborative problem-solving assessment was analyzed using the K-medoids clustering method. The analysis revealed three distinct types of collaborative problem-solving skills: a fast type that allocates less time to responses, a deliberative type, and a type that engages in a higher number of interactions over a longer period, even when achievement levels are similar. We also analyzed the distribution of correct answer rates and collaborative problem-solving competency scores by cluster and level, confirming that various achievement levels were represented across all clusters. Based on these differences, educational and evaluative intervention strategies were explored, and the implications for assessment were discussed.