The investigation of measurement invariance of TIMSS-2015 mathematics and science affective characteristics models according to culture, gender and statistical region

The purpose of this study was to investigate measurement invariance of the mathematics and science model constructed by using TIMSS 2015 8th grade student survey, according to cultures (Turkey, Saudi Arabia and Singapore), gender and regions. The research was conducted by means of 15954 students' data from Turkey, Singapore and Saudi Arabia sample participated TIMSS 2015 survey. In the research, mathematics and science affective characteristics models are constructed as students like learning mathematics and science, students value mathematics and science, student confident in mathematics and science and students engaged in mathematics and science lessons. Measurement invariance of the model constructed by using EFA and CFA, according to culture, gender and regions are analyzed. Measurement invariance is examined by multiple group confirmatory factor analysis. As a result, the mathematics and science models support scalar invariance between regions and cultures and support strict invariance between genders. It can be concluded that since strict invariance is supported, all comparisons between genders will be meaningful and since scalar invariance is supported according to cultures and regions, cross-group comparison of latent variables means on the latent variable of interest.