Diagnostic Models of Science Attribute Mastery: An International Multigroup Study

• In Event: Discussions on STEM Assessment Within an International Context

Abstract

Multiple studies employing latent class models for cognitive diagnosis have examined international large-scale assessments to better understand student mastery of educational content. Many of these studies have exclusively focused on investigating a student’s attribute mastery on mathematics, reading, or language examinations. The current study adds to this literature by examining the science attribute mastery of Singapore, Hong Kong, and the U.S. on the 2007 Trends in International Mathematics and Science Study (TIMSS). A Q-matrix is constructed by three science education experts using 25 items from the 2007 TIMSS, and science attribute mastery is examined using a multi-group deterministic input noisy “and” gate (MG-DINA) model and a multi-group general diagnostic model (MG-GDM).

Authors

• Dakota Wayne Cintron, Neag School of Education, University of Connecticut

• Young-Sun Lee, Teachers College, Columbia University