Testing Measurement Invariance of the Students’ Affective Characteristics Model across Gender Sub-Groups
Ergül DemirDepartment of Measurement and Evaluation, Ankara University, Cebeci Ankara 06590 Turkey
In this study, the aim was to construct a significant structural measurement model comparing students’ affective characteristics with their mathematic achievement. According to this model, the aim was to test the measurement invariances between gender sub-groups hierarchically. This study was conducted as basic and descriptive research. Secondary level analyses were conducted on the Program for International Student Assessment (PISA) 2012 Turkish student questionnaire data. The sample consisted of 4,848 fifteen-year-old students from 170 schools and 12 statistical territories. For analysis techniques, regression analysis, exploratory, and confirmatory factor analyses were executed in order to construct a significant initial measurement model. Multi-group confirmatory factor analysis was executed to analyse the invariance between gender sub-groups. According to the results, taking into consideration the limitations of the model which was constructed in this study, it was observed that strong invariance between gender sub-groups was provided. This finding indicates that there are similarities of affective characteristics for fifteen-year-old Turkish students across gender sub-groups. The results can be evaluated as evidence that the possibility of bias or prejudice in students’ affective characteristics toward mathematics is not high.