Assessing the dimensionality of student school engagement survey : support for a multidimensional bifactor model

Abstract

The Student School Engagement Survey (SSES) is used to evaluate student engagement interventions run by the National Center for Student Engagement in the U.S. It was designed to measure the behavioral, emotional, and cognitive components of engagement, but its factorial structure has not been validated. To address this limitation, we tested the factorial structure of the Portuguese version of the SSES using a representative sample of 4,866 adolescents. An exploratory factor analysis revealed five theoretically meaningful factors describing subtypes of emotional and behavioral engagement, and teacher support for learning. A confirmatory factor analysis supported modelling a shortened version of the SSES with a bifactor model. Bifactor indices indicated total SSES scores are interpretable as a measure of a single student engagement construct. Finally, as evidence of concurrent validity, the scale had a strong positive correlation with an established measure of student engagement. The proposed version of the SSES is a psychometrically adequate measure of student engagement, although cannot be said to measure cognitive engagement.