Multi-Group Confirmatory Factor Analysis via Partial Measurement Invariance Methods

In this study, the data obtained from the 6-item survey applied to the participants of the Trends in International Mathematics and Science Study (TIMSS) 2019 Science course were used. The data of the TIMSS 2019 United States of America (USA) sample were compared with1 the data of countries that have different cultural structures such as Turkey (TUR), Japan (JPN), and Norway (NOR). This study aims to investigate the measurement invariance by constructing structural equation models for the comparisons to be meaningful and reliable. In the study, Exploratory Factor Analysis (EFA) is performed, and Structural Equation Model (SEM) is developed. The measurement invariance was examined by performing confirmatory factor analysis of the measurement tool amongst countries with different cultural structures. The results of data comparison between the USA and these other three countries (TUR, JPN, and NOR) are as follow; between USA and TUR: configural and weak; partially scalar; and partially strict invariance, between USA and JPN: configural and weak; scalar and strict invariance, while between USA and NOR: configural and weak; scalar and strict invariance.