Using school and child welfare data for Pennsylvania school districts, REL MidAtlantic finds that predictive models can be used to effectively identify at-risk students. They consider short-term academic outcomes including chronic absenteeism, suspensions, course failure, low grade point average, and low scores on state tests. The idea is to successfully identify near-term challenges so that administrators and school staff can provide additional support before a problem develops or a student considers dropping out. Interestingly, researchers found that models including out-of-school predictors from human services data did not enhance the performance of the models, suggesting models using only in-school data are sufficient. #education
You must be logged in to post a comment.