Identifying At-Risk Factors of First-Year Attrition: A Neural Network Approach

• In Event: Marriott Roundtable Session 10
  In Roundtable Session: (Re)Defining Risks, Acknowledging Cultural Wealth, and Improving Success Odds

Abstract

College attrition has been and might still be one of the serious issues facing the nation and the higher education system. Utilizing a neural network approach and drawing on the enrollment census data of fall 2014 through fall 2017, this study identifies at-risk factors of first-year attrition. Results reveal that the neural network produces 79.28% of prediction accuracy. Results also confirms that credit hours attempted in first fall term and FYS grades are the most important predictors of first-year attrition/retention. Therefore, first-year seminars engagement and initial credits hours attempted could serve as the early alert that provide an opportunity for instructors and advisors to use the initial credit load and earlier FYS performance to identify students for targeted, proactive interventions.

Authors

• Qingmin Shi, University of Nevada - Las Vegas

• Tina Nelson, University of Nevada - Las Vegas

• Shiloh Bradley, University of Nevada - Las Vegas