A multiple linear regression model for predicting zone A retention by military occupational specialty.

Abstract

The Selective Reenlistment Bonus (SRB) program is designed to offer an attractive reenlistment incentive to improve manning in critical skills. To efficiently manage the SRB program, a requirement exists to maintain MOS level estimating factors for use in projecting retention rate improvement as a function of SRB award level. This thesis formulates and solves a mathematical model which explains the variation in zone A retention rates as a function of SRB award level and other factors believed significant in the reenlistment decision. To allow for comparison of the estimating factors associated with the SRB variable across MOS, an overall projection model was developed. Stepwise multiple linear regression analysis techniques were used on a subset of the enlisted MOS inventory in the model development phase of this analysis. The proposed overall model was then fitted to a second subset of MOS to validate the assumptions and effectiveness of the proposed linear model.