Delayed Entry Program attrition: a multivariate analysis

Abstract

This thesis uses binary logit models to examine the effects of personal background characteristics and local area economic conditions on an individual's likelihood to leave the Delayed Entry Program (DEP). The DEP allows individuals to contract for enlistment up to one year prior to starting basic training. Between fiscal years 1991 and 1996, over 1.1 million individuals joined the DEP for all four military branches, yet 167,134 (15 percent) never entered basic training. The Defense Manpower Data Center (DMDC) provided a data file that included all individuals who joined the DEP between October 1989 and June 1996. The DMDC raw data file was merged with county-level unemployment data provided by the Bureau of Labor Statistics. DEP attrition was modeled as a function of gender, educational level, dependent status, Armed Forces Qualification Test (AFQT) score, race, ethnicity, moral waiver status, and county-level unemployment rates. Results show that a person's likelihood of leaving the DEP is affected most strongly by gender and educational level. Women and high school seniors are more likely than men and high school graduates to leave the DEP. County-level unemployment rates have a significant, but small, negative effect on DEP attrition.