Exploring the use of non-cognitive factors in predicting college academic outcomes

The present study evaluated the incremental utility of a blended model of cognitive and non-cognitive variables versus a cognitive-only set of variables to predict academic success among new college students. Traditional cognitive predictors include high school grade point average and ACT scores. The non-cognitive factors of interest were grit, goal orientation, and academic self-efficacy. It was hypothesized that the blended model would demonstrate stronger predictive validity than cognitive predictors alone, grit would be associated with goal orientations and academic self-efficacy, and that academic self-efficacy would mediate the relationship between other non-cognitive predictors and academic success outcomes. The results from archival data (N = 8,742) and online survey collection (N = 624) suggest non-cognitive factors improve traditional prediction models, particularly through grit and academic self-efficacy. Self-efficacy also mediated the non-cognitive to academic success relationship. Higher Educational institutions could consider the implementation of blended models, using key non-cognitive predictors