Analysis of Mid-Grade Naval Aviator Retention.
Abstract
Attrition of aviators is of major concern to the Navy because of the costs and numbers involved. The Navy currently forecasts aviator retention and attrition by extrapolating historical trends. This thesis recommends that the Navy replace the current method with two alternative statistical techniques: logistic regression and classification trees. They are recommended for two reasons. First, the proposed techniques make significantly more accurate forecasts of aviator retention than the current method. Second, the proposed techniques, unlike the current method, can identify the significant variables affecting aviator retention. Use of the proposed techniques can therefore lead to the formulation of better aviator retention policies by the Navy. These arguments are demonstrated with a case study of an existing retention database. The variable identified as most significant for aviator retention in this analysis were the geographic location of an aviator' duty station, assignment to non-flying billets, and grade. Policy implications of these findings are discussed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1998
- Accession Number
- ADA356023
Entities
People
- Scott H. Poindexter
Organizations
- Naval Postgraduate School