Student-student relations from the teacher versus student perspective: A multi-level confirmatory factor analysis

Educational research emphasizes the advantages of multimethod designs. However, if the design comprises different perspectives, the question of construct validity emerges. We related this question to student and teacher ratings of student-student relations, which are of high interest in research on physical education. In our study, 2,160 students and their 114 physical education teachers answered parallel items. Data was analyzed by applying the multi-level correlated trait-correlated method minus one approach, which makes it possible to consider multiple indicators per trait and to distinguish between trait, method, and error components. The student-student relations subscales student conflicts, class cohesion, and pleasure in demonstrating were reliable and valid regarding the two perspectives. Moreover, teacher and student ratings of student-student relations were consistent to a certain extent. The administered questionnaire was suitable for both perspectives. The applied model provides a good basis for similar research questions regarding construct validity.