Accession Number : ADA307513
Title : Predicting Naval Aviator Attrition Using Economic Data.
Descriptive Note : Master's thesis,
Corporate Author : NAVAL POSTGRADUATE SCHOOL MONTEREY CA
Personal Author(s) : Bookheimer, William R.
PDF Url : ADA307513
Report Date : MAR 1996
Pagination or Media Count : 71
Abstract : Understanding and accurately predicting attrition is vital to correctly managing the retention of naval aviators. This thesis investigates the ability of models incorporating a number of economic measures to predict naval aviator attrition rates. Using data from 1978 to 1990, this study examined a wide range of potential economic explanatory variables and their effects on naval aviator attrition rates. The naval aviator data set was grouped into six populations, separated by aviation community (helicopter, jet and propeller) and by years of service (5-8 and 9-12). Three separate linear regression models for each of the aviator groups were developed, and their predictive ability evaluated. The study found that: no single model was best at predicting attrition rates for all groups; simple models using one or two variables performed better than complex, multivariate models; the most useful predictor variable was the national unemployment rate; attrition rates with the highest levels and variability were in the jet and propeller pilot groups with five to eight years of service, and the most significant models, able to outperform a naive prediction, were found for these groups.
Descriptors : *NAVAL PERSONNEL, *ATTRITION, *PERSONNEL RETENTION, DATA BASES, MATHEMATICAL MODELS, NAVAL AVIATION, ECONOMIC IMPACT, COST ANALYSIS, MULTIVARIATE ANALYSIS, PILOTS, ACCURACY, THESES, CAREERS, CORRELATION, MANPOWER, MATHEMATICAL PREDICTION, SALARIES, PERSONNEL SELECTION, OFFICER PERSONNEL, LINEAR REGRESSION ANALYSIS, JOB SATISFACTION, PROMOTION(ADVANCEMENT), GRADE STRUCTURE(PERSONNEL MANAGEMENT), RETIREMENT(PERSONNEL), UNEMPLOYMENT, ECONOMIC MODELS, MILITARY DOWNSIZING.
Subject Categories : Personnel Management and Labor Relations
Economics and Cost Analysis
Distribution Statement : APPROVED FOR PUBLIC RELEASE