Investigation of measurement invariance of TIMSS 2019 Mathematics Achievement Test and mathematical attitudinal characteristics

TIMSS implementations reveal the knowledge and skills of students in mathematics and science and support countries in improving their education in these fields and predicting their achievements. With the results obtained from TIMSS implementations, the achievements and various characteristics of students in mathematics and science in different countries are compared. At this point, in order for the interpretations regarding the comparisons to be accurate, it is necessary to collect evidence on the measurement invariance of the achievement tests and scales used in TIMSS implementations. Based on this need, this study aimed to test the measurement invariance of the 8th grade mathematics achievement test in the TIMSS 2019 implementation and the model created from the scales of "Students Like Learning Mathematics", "Students Confident in Mathematics" and "Students Value Mathematics" related to affective traits and each scale separately according to the countries of Turkey, South Korea and Finland. First, exploratory factor analyses and then confirmatory factor analyses were conducted for the achievement test, the model of affective traits and each scale. In addition, after the validation process for each country, measurement invariance was tested with Multiple Group Confirmatory Factor Analysis based on Structural Equation Modeling by using data from three countries together. As a result of the research, the evidence for the "scale invariance" stage was not provided for the achievement test, the model for affective traits and each scale, while the "metric invariance" stage was provided for the model for affective traits and each scale. Therefore, it would not be correct to state that the differences that emerged are due to the countries. Therefore, this situation should not be neglected when making comments about the countries.