Assessing the "explanatory power" of school process variables constructed through multilevel exploratory factor analysis

This paper investigates the "explanatory power" of school process variables based on teacher data constructed through multilevel exploratory factor analysis, compared to school process variables developed using conventional exploratory factor analysis plus aggregation. Measures for school process variables are associated with student achievement and non-cognitive outcomes. The results are mixed: While for some measures we observe different results depending on the constructional approach, for other measures we find exactly the same results irrespective of the constructional approach. Nevertheless, these results underline the importance of using the correct approach to develop measures for school process variables. Since multilevel exploratory factor analysis has only scarcely been used in educational effectiveness research, this study has major relevance for educational effectiveness researchers.