An Investigation of Measurement Invariance of the TIMSS Fourth-Grade Mathematics Assessment Across OECD Member Countries

This study aimed to examine the measurement invariance of the factor structure of the scores obtained from the fourth-grade mathematics achievement test administered within the framework of TIMSS 2015 across OECD member countries. A total of 132226 students from 24 countries participated in TIMSS 2015. This correlational survey model research was conducted on the data of 9641 randomly selected students who took the 7th booklet. Within the scope of the study, normality was tested before the measurement invariance analyses. Normality is a basic assumption of measurement invariance analyses. Reliability coefficients were determined, and confirmatory factor analysis was performed for all groups to be tested for measurement invariance. Lastly, the equality of covariance matrices was tested. As a result of the analyses, it was determined that the assumptions were confirmed. Then, multi-group confirmatory factor analysis and measurement invariance analyses were performed on a single-factor design. The research determined that the best working model among the measurement invariance models was the strong factorial invariance model. Based on this finding, it was concluded that TIMSS 2015 fourth-grade mathematics achievement test scores did not provide measurement invariance among OECD member countries; in other words, the factorial design was not the same for the countries compared, thus raising doubts about making comparisons between these countries.