A Preliminary Development of a Nuclear Propulsion Officer Enlistment Supply Model.
Abstract
This study constructs linear forecasting models, for each of the six Navy Recruiting Areas and the Recruiting Command, that attempt to predict future United States Navy Nuclear Propulsion Officer contracts signed in any one of four fiscal year quarters, given estimates of independent supply variables included in the models. The models are developed using stepwise multiple regression analysis with ordinary least squares and are supported by historical data from fiscal years 1981 through 1985. In developing the models, the thesis examines the relationship between the contracts signed in a given quarter and the following supply variables: number of recruiters, annual goals of number of contracts to be signed, military-to-civilian pay ratio, unemployment rate, size of target population (in the form of market share), advertising and marketing costs and seasonal effects, represented by proxy variables. The assumptions of using multiple regression analysis and linear models are examined through a graphical study of the residuals and do not seem to be refuted. Each of the models are corrected for first order autocorrelation. Validation of the forecasting models was attempted by the comparison of predicted contracts signed in a quarter against new contract data obtained for fiscal year 1986. The results of the forecasting comparisons are much worse than expected. Possible causes for the large error percentages in the comparisons are mentioned in this study but not examined in detail. (Theses).
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1986
- Accession Number
- ADA176021
Entities
People
- Paul T. Serfass Jr
Organizations
- Naval Postgraduate School