This journal article discusses findings from a study in which researchers developed a machine learning classifier to predict nontraditional student dropout.
Researchers at the Regional Educational Laboratory Southwest, operated by AIR, in collaboration with the Texas Hispanic STEM Research Alliance, conducted a study to identify associations between predictive indicators and postsecondary science, technology, engineering, and math (STEM) success among Hispanic students in Texas. The goals of the study were to identify factors that predict positive STEM-related postsecondary outcomes for students in Texas and to determine whether the association between predictive factors and outcomes differs between Hispanic and non-Hispanic White students.
AIR, in collaboration with IMPAQ, is designing and building a borrower-based dynamic microsimulation model of the repayment of federal student loans for the Cost Estimation and Analysis Division of the U.S. Department of Education's Office of Budget Services. This work will help the Department better estimate the costs and consequences of student loan debt for a wide array of student populations, as well as understand the impact of potential policy changes on loan program costs and student outcomes.
The Regional Educational Laboratory Southwest, operated by AIR, in collaboration with the Southwest College and Career Readiness Research Partnership, studied the ability of these indicators to predict postsecondary readiness (ACT score of 19 or above) and success (college enrollment and persistence within eight years of beginning grade 6) for Arkansas students who entered grade 6 in 2008/09 or 2009/10. The study’s findings can help state and local education agencies, both in Arkansas and across the nation, identify and support middle and high school students who are on and off track for attaining postsecondary readiness and success.