A Methodology for Long-Term Forecasts of Air Force Pilot Retention Rates: A Management Perspective
Abstract
Personnel planners in various Air Force agencies use models, among other things, to aid them in forecasting pilot retention rates. This Theses attempted to forecast retention rates three years ahead with the use of multiple regression analysis techniques. Such models can be of use of Air Force leaders to develop proactive policies and programs to combat poor retention forecasts. Economically quantifiable variables were primarily used in the modeling effort. However, some year groups could not be adequately explained with the use of economic variables alone. The models for year groups eight, twelve, and thirteen used the retention rates of 'peer groups' to assist in explaining their own retention rates. All models were subjected to common internal tests associated with linear regression. External validity was verified by the use of a withheld data set. Forecasts were made for Fiscal Years 90, 91, and 92, using independent variable data from 1987, 1988, and 1989, respectively. All tests and forecasts were thoroughly documented. The practical and policy implications of these forecasts were discussed, and some thoughts about possible policies and programs to increase retention were advanced. Improvements to further the utility of these models were suggested.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1990
- Accession Number
- ADA229541
Entities
People
- Bruce A. Guzowski
Organizations
- Air Force Institute of Technology