Tests of factorial structure and measurement invariance for the Student Engagement Instrument: Evidence from middle and high school students in Portugal

Abstract

While past research has shown the Student Engagement Instrument (SEI) to be an adequate measure of school engagement, this self-report questionnaire currently exists as a number of different versions, each comprising various subsets of the original 35 items (Appleton et al., 2006). The first objective of this study was to use confirmatory factor analysis to assess the fit of different factorial structures with data acquired from 1,229 Portuguese adolescents in the 7th to 12th grades. Our second objective was then to test measurement invariance across gender and school year and to examine the psychometric properties of the factorial structure of the SEI with the best fit to our data. Our analyses revealed that the 15-item Brief-SEI structure fit our data best, and continued to fit the data well with the addition of two-higher order factors representing Cognitive and Psychological Engagement. This higher-order model showed strict measurement invariance across gender and strong measurement invariance across school year. Internal consistency of the subscales was good and the subscales were correlated (albeit weakly) with academic performance. This study therefore provides a synthesis of current research on the SEI, and offers a validated instrument which is consistent with Appleton’s theoretical conceptualization of student engagement.