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.

我们研究了如何使用基于代理的模拟有效地促进医学生的协作诊断推理，尤其是证据的引出和共享。已建议提供自适应协作脚本来提高有效性，但是现有证据多种多样，并且可能会受到非系统群体的影响。人们批评协作脚本破坏了学习者的能力。我们调查了自适应和静态脚本对协作诊断推理和基本心理需求的影响。我们将160名医学生随机分配到以下三个组之一：自适应，静态或无协作脚本。我们发现，使用自适应协作脚本进行学习可以增强证据共享性能和转移性能。脚本编写不会影响学习者的自主性和社会关联性。但是，与静态脚本相比，自适应脚本对感知能力具有积极的影响。我们得出结论，对于复杂技能，用适应性脚本补充基于代理的模拟似乎对帮助学习者内部化协作脚本而不会对基本的心理需求产生负面影响是有益的。