Learning to diagnose collaboratively – Effects of adaptive collaboration scripts in agent-based medical simulations

Abstract We investigated how medical students' collaborative diagnostic reasoning, particularly evidence elicitation and sharing, can be facilitated effectively using agent-based simulations. Providing adaptive collaboration scripts has been suggested to increase effectiveness, but existing evidence is diverse and could be affected by unsystematic group constellations. Collaboration scripts have been criticized for undermining learners' agency. We investigate the effect of adaptive and static scripts on collaborative diagnostic reasoning and basic psychological needs. We randomly allocated 160 medical students to one of three groups: adaptive, static, or no collaboration script. We found that learning with adaptive collaboration scripts enhanced evidence sharing performance and transfer performance. Scripting did not affect learners’ perceived autonomy and social relatedness. Yet, compared to static scripts, adaptive scripts had positive effects on perceived competence. We conclude that for complex skills complementing agent-based simulations with adaptive scripts seems beneficial to help learners internalize collaboration scripts without negatively affecting basic psychological needs.