Examination of High Quality Enlistees on a Recruiting District Level

摘要

In this thesis two distinct analytical techniques were used to develop models to forecast the number of high quality nonprior service males enlisting monthly in the U.S. Navy on a recruiting district level. The Box-Jenkins methodology was applied to a monthly time series of enlistments for the period October 1978 to September 1983. A multiple regression causal model was developed based on the explanitory variables: numbers of unemployed, change in the rate of unemployment and military/civilian pay ratio. A combined time series/causal model was developed by applying the Box-Jenkins technique to the residuals of the multiple regression. These models were compared for predictive validity. Recommendations for further development of models containing explicit time series elements are presented.