Students' multidimensional profiles of math engagement: Predictors and outcomes from a self-system motivational perspective.

Abstract

Math engagement research has been largely limited to examining the unique and additive relations of engagement dimensions with outcomes. However, an emerging perspective is that students may simultaneously invest varying degrees of their distinct energetic resources (e.g., cognitive vs. emotional) in their interactions with the math learning environment. Adopting a person-centred perspective, we examined unique latent subpopulations of adolescents' multidimensional math engagement. Importantly, we did so while accounting for generality and specificity in engagement data, including general engagement and specific cognitive, emotional, and social engagement dimensions. Additionally, we examined students' math self-efficacy, outcome expectations, and gender as predictors, and math achievement indices as outcomes, of profile membership. The sample comprised 400 Australian school students taking mandatory math classes. Data on students' multidimensional engagement, math self-efficacy, math outcome expectations, and demographic characteristics were collected at the beginning of the academic semester. Standardized test scores and class grades were retrieved at the end of the semester. Latent profile analyses, based on preliminary bifactor exploratory structural equation models intended to tease apart generality from specificity in engagement data, revealed 'Minimally Engaged', 'Emotionally Disengaged', and 'Moderately-to-Highly Engaged' profiles. Additionally, math self-efficacy, outcome expectations, and gender were found to predict the likelihood of profile membership. Finally, class grades, but not standardized test scores, were found to significantly differ across the profiles, accounting for prior achievement, gender, and grade level. The findings replicate previous work that has shown profiles of student engagement and extend this work by (1) accounting for known generality and specificity in engagement data and (2) obtaining evidence for relations of profile membership with gender, self-beliefs, and achievement.