LAUSR

Abstract In recent years, there has been a bourgeoning scholarly interest in teacher engagement that has stemmed, in part, from broader policy discussions around teacher effectiveness. However, theory and research on the dimensional structure of teacher engagement data remains underdeveloped, potentially limiting the interpretational utility of inferences drawn from teacher engagement research. The present study examined the dimensionality of teacher work engagement data using an analytic framework designed to account for multiple sources of construct-relevant multidimensionality. In addition, the measurement and structural invariance across teaching level, the longitudinal invariance and stability, and criterion-related validity with respect to job satisfaction were examined. Results showed that a bifactor-ESEM model, characterized by factors representing general engagement and specific energetic investments, provided the best representation of teacher engagement data. This measurement structure was found to be invariant across teaching level; however, plausible latent mean differences in the engagement dimensions were found across primary and secondary teachers, such that primary teachers reported significantly higher general engagement and specific engagement with students than secondary teachers. Results also supported the full longitudinal measurement invariance of the retained measurement structure and considerable rank-order stability of the general and specific engagement dimensions over a four-month period. Finally, evidence of the criterion-related validity of the engagement scores with respect to job satisfaction was obtained. Taken together, the study yields crucial evidence supporting a novel multidimensional, hierarchical representation of teacher engagement.