Developing game-based models of cooperation, persistence and problem solving from collaborative gameplay

Abstract

Collaborative-problem solving (CPS) is an important 21st century skill and it continues to be a complex skill to model and assess. We approached this challenge by first looking at the individual level primary cognitive and social aspects of CPS. This paper demonstrates ongoing work of designing and developing game-based models of three CPS components: cooperation, problem-solving, and persistence. A study was conducted collecting data from the game-play of 11 groups (three middle school students in each group) tasked with solving challenges in Physics Playground. We employed evidence-centered design principles to develop behavioral indicators of cooperation, problem-solving and persistence. These were used to code each student’s behavior during three hours of video-recorded gameplay. For each CPS component, we applied hierarchical clustering on this video-coded data and qualitatively evaluated two generated clusters of students across groups. For cooperation, there was more communication with other students in working towards a solution for one group. For problem-solving, one group had more instances of talking about possible solutions. For persistence, one group had more attempts in a challenge and was more on-task. Implications of results, limitations and future work were discussed.