Examining the Measurement Invariance of TIMSS Science Self-Concept Scales Across Genders, Cultures, and Languages

This study employed item response theory (IRT), exploratory factor analysis (EFA), and confirmatory factor analysis (CFA) to examine the measurement invariance of TIMSS science self-concept (SSC) scales for grades 4 and 8 across genders, cultures, and languages in six countries and regions. IRT, and EFA and CFA were utilized to examine the reliability and validity of the SSC scales, respectively. The IRT results revealed substantial variation in each item and test information function values across the six societies. The EFA and CFA results identified two dimensions of SSC scales, that is, positive perception of SSC (SSCp) and negative perception of SSC (SSCn), with satisfactory fit indices. Measurement invariance was confirmed across genders, and cultural and linguistic groups. Recommendations are put forward for TIMSS program developers and researchers regarding the revision of the SSC scales.