Estimating commute distances of U.S. Army Reservists by regional and unit characteristics

Abstract

This thesis develops a multiple regression model using regional and unit characteristics to estimate commuting distances of U.S. Army Reservists. The data were obtained from a 1988 file established by the Defense Manpower Data Center containing locational and biodemographic information on 238,174 enlisted reservists. A random sample of 91 reserve centers was selected for the analysis. The logistic and normal distributions were evaluated as possible candidates for fitting the commuting distance distribution. It was found that a power transformation of the fractional distance traveled fit both distributions quite well. Parameters for the two distributions are obtained through a method of maximum likelihood estimation. Finally, a multiple regression equation is used to estimate the parameters of the commute distance distribution as a function of reserve center and market characteristics.