Predicting Midshipmen's Outcomes At The United States Naval Academy
Abstract
This research examines, describes, and analyzes factors that are associated with midshipmens outcomes at the United States Naval Academy. Specifically, we identify factors that help predict which midshipmen will graduate in the top 10%, bottom 10%, or undergo attrition. The goal is to identify a list of factors which company officers and senior enlisted leaders can use to help develop midshipmen morally, mentally, and physically. We used logistic regression, classification trees, and random forests to seek the most effective prediction model for midshipmen's outcomes. The results of our logistic regression model accurately identify 71.4% of midshipman who are predicted to graduate in the top 10%, and 66.7% of midshipmen who are predicted to graduate in the bottom 10%. Additionally, whole person multiple, math SAT scores, participation in extracurricular activities, Myers-Briggs Type Indicator results, and mile times are key factors for predicting the top 10%. For the bottom 10%, the key factors are whole person multiple, math SAT scores, race/ethnicity, and prior enlistment. Due to a lack of specific attrition data, attrition models were unsuccessful. This study summarizes results, makes recommendations to the United States Naval Academy, and lists potential future work for Naval Postgraduate students.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2021
- Accession Number
- AD1151022
Entities
People
- George Jamison
Organizations
- Naval Postgraduate School