📄 中文摘要
包容性、公平性和可及性在人工智能和教育领域受到广泛重视，但通常通过粗略的样本描述或事后自我报告进行评估，这些方法无法捕捉到在协作问题解决过程中包容性如何在每个瞬间被塑造。提出了一种名为包容性分析的框架，旨在将包容性视为协作问题解决中的动态互动过程。包容性被概念化为三个互补维度——参与公平、情感气候和认知公平，并展示了如何通过可扩展的互动级别测量使这些构念在分析上可见。通过模拟对话和来自人机团队实验的实证数据，展示了这些分析方法的有效性。
交互中的包容性测量：人机协作学习的包容性分析
📄 English Summary
Measuring Inclusion in Interaction: Inclusion Analytics for Human-AI Collaborative Learning
Inclusion, equity, and access are highly valued in AI and education but are often evaluated through coarse sample descriptors or post-hoc self-reports that fail to capture how inclusion is shaped moment by moment in collaborative problem solving (CPS). A discourse-based framework called inclusion analytics is proposed to examine inclusion as a dynamic, interactional process in CPS. Inclusion is conceptualized along three complementary dimensions: participation equity, affective climate, and epistemic equity. The paper demonstrates how these constructs can be analytically visible using scalable, interaction-level measures. Using both simulated conversations and empirical data from human-AI teaming experiments, the effectiveness of these analytical methods is illustrated.