Modeling Officer Selection for Naval Special Warfare

Every year, the SEAL Officer Community Management (OCM) receives approximately 300 applications from candidates who want to attend Basic Underwater Demolition/SEAL (BUD/S) training to become Navy SEAL Officers. The applications include multiple data elements such as university attendance, grade point average (GPA) and physical fitness test scores. From this data, the SEAL OCM selects approximately 80% of the candidates to participate in SEAL Officer Assessment and Selection (SOAS). At SOAS, the candidates are assessed by the cadre at Naval Special Warfare (NSW) Basic Training Command (BTC), who provide evaluation data and recommendations to the SEAL OCM for candidate selection to BUD/S. In total, the process of assessing, selecting and training candidates to become SEAL Officers is resource-intensive, incurring a financial cost to the Navy, manning challenges for NSW and the time and energy of the candidates themselves. Removing candidates who have a low probability of success at BUD/S early in the process reduces the costs and allows them to be reassigned within the Navy to a community more appropriate to their abilities. This thesis aims to analyze data collected on the candidates to train statistical models capable of predicting a candidate's probability of success in the first phase of BUD/S and inform data collection to capture information on candidates for future data analysis.