Investigating the Measurement Invariance of Large-Scale International Assessments Using Generalizability Theory

• In Event: 16.057 - Roundtable Session 5
  In Roundtable Session: 16.057-4 - Missing Data and Investigations of Statistical Assumptions

Abstract

The importance of examining the measurement invariance (MI) of measures before making international comparisons has gained increasing recognition. We demonstrated the application of Generalizability (G) theory to evaluate the MI of large-scale international assessments through two examples using data from the 2009 Programme for International Student Assessment. The two examples were used to illustrate the procedure of investigating MI using G theory when the facets of differentiation were the jurisdictions and the students respectively. Understanding the sources of observed score variance is helpful for the comparative research as well as the development of international assessments. This paper attempted to help researchers, educators and policymakers understand how to apply G theory to investigate MI and how to interpret the analysis results.

Authors

• Yue (Michelle) Chen, The University of British Columbia

• Wen Zhang

• Eric K. H. Chan, The University of British Columbia

• Bruno D. Zumbo, The University of British Columbia