Data Mining to Explore Trends in Undergraduate Retention and Graduation Rates at NKU

• Activity Number: 552
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Mark Lancaster* and Kacie Kotnik and Nathaniel Hudson and Amy Becknell and Madhura Kulkarni and Maureen Doyle and Kristi Haik and Joseph Nolan
• Companies: Northern Kentucky University and Northern Kentucky University and Northern Kentucky University and Northern Kentucky University and Northern Kentucky University and Northern Kentucky University and Northern Kentucky University and Northern Kentucky University
• Keywords: Data Mining ; Statistical Learning ; Markov Modeling ; Regression Modeling

Abstract

Both nationally and at Northern Kentucky University (NKU), factors affecting retention rates of undergraduate students are of significant interest. We are working in conjunction with NKU's Office of Institutional Research to identify specific patterns and factors that influence student success. Highlighting the end-to-end process, we (1) utilize exploratory data analysis techniques to correct naturally occurring inconsistencies and assess the utility of the data, and (2) use statistical analysis (regression and classification models) and Markov modeling to allow for the identification of trends in student retention as well as graduation rates for various demographics. (3) Additional ancillary data, such as regional economic information from the Bureau of Labor Statistics are being considered as part of the external factors that should be appropriately modeled. (4) Using R's graphing capabilities, maps displaying these results regionally for key areas of interest in Kentucky are provided. The ultimate goal is to disseminate actionable information that will equip NKU (and perhaps other universities) to make strategic decisions to better support its students to graduation.