A Data-Mining Approach to Differentiate Predictors of Retention.

Abstract

Student retention is an important issue for all university administrators and faculty due to the potential negative impact of student attrition. Universities with high attrition rates face the substantial loss of tuition, fees, and potential alumni contributions (DeBerard, Spielmans, & Julka, 2004), while the students themselves also face negative consequences. According to the U.S. Department of Education, National Center for Education Statistics (NCES), students who leave college early are likely to earn less income over their lifetimes when compared to peers who have graduated (NCES, 1989). Despite the identified consequences of college dropout for universities and students, as well as concentrated efforts from all educational institutions on improving student retention, attrition rates remain relatively high across the United States. Data from the National Center for Public Policy and Higher Education (NCPPHE) reveal that only 73.6 percent of first-time, full-time freshmen (enrolled in 2002) returned for their second semester (2007). Looking at college completion data from 2005, only 39.5 percent of undergraduate students enrolled in public institutions completed their degrees within five years (ACT, 2005). In discussing retention statistics, it is important to explore the definition and methods for