Longitudinal characteristics and incremental validity of the Student Engagement Instrument (SEI)
Lovelace, Matthew Davis
MetadataShow full item record
The two studies presented here investigated the longitudinal properties of the Student Engagement Instrument (SEI), a self-report measure of cognitive and affective engagement in school. First, change and stability were studied in SEI scores through the middle and high school years, including annual retest stability and mean level change. Next, the four-year predictive utility of the measure was assessed in terms of its ability to predict dropout and on-time graduation relative to data commonly available in school records, such as academic achievement and disciplinary data. Here, analyses of various ninth grade predictors were subjected to two-pronged tests of predictive power and yield and, ultimately, to an incremental validity analysis through a two-level multivariate logistic regression. Results indicated that stability and change in SEI scores over time fit well with expectations in terms of theory and prior empirical evidence in the engagement literature. Further, some factors on the SEI met several rigorous tests of predictive validity in relation to dropout, even when controlling for other powerful predictors of the outcome. Implications and limitations of findings are discussed, as well as future directions for research.