Identifying supportive contexts for mindset interventions: A two-model machine learning approach
share
Growth mindset interventions (which foster students' beliefs that their abilities can grow through effort) are more effective in some contexts than others; however, relatively few variables have been explored that could identify contexts in which growth mindset interventions are most effective. In this study, we utilized machine learning methods to predict growth mindset effectiveness in a nationwide experiment in the U.S. with over 10,000 students. These methods enable analysis of arbitrarily-complex interactions between combinations of student-level predictor variables and intervention outcome, defined as the improvement in grade point average (GPA) during the transition from high school. We utilized two separate machine learning models: one to control for complex relationships between 51 student-level predictors and GPA, and one to predict the change in GPA due to the intervention. We analyzed the trained models to discover which features influenced model predictions most, finding that prior academic achievement, intervention compliance, self-reported reasons for learning, and race/ethnicity were the most important predictors in the model for predicting intervention effectiveness. Unique to this study, we found that low intervention compliance (attempting to navigate through the intervention software without completing all steps) resulted in as much as -0.2 difference in predicted intervention effect on GPA. Our findings have implications for the design of computer-administered growth mindset interventions, especially for students who do not properly complete the intervention.
READ FULL TEXT
