报告时间：2016年5月18日（周三）下午4:00

报告地点：京师学堂第七会议室

报 告 人：包钰，美国佐治亚大学教育心理系博士生。

报告题目：The General Diagnostic Classification Model Framework and Its Application 

报告摘要：A key purpose of educational assessment is to reliably and efficiently determine what students do and do not understand. Diagnostic classification models (DCMs) are a newer class of statistical tools well-suited to fulfill this purpose. DCMs classify students according to mastery levels of latent knowledge components, called attributes. Diagnostic tests are carefully designed where each item elicits one or more attributes. As a general DCM framework, the log-linear cognitive diagnosis model (LCDM) allows attribute behavior to vary across items within the same diagnostic test. In this talk, I will introduce LCDM and its relationship with the most commonly used DCMs, like DINA, DINO and CRUM models. I will then present a top-down approach by Mplus for a real data analysis. I will also presents the current developments on DCMs.