Accession Number:

ADA248520

Title:

Predicting Coast Guard Enlisted Attrition

Descriptive Note:

Master's thesis

Corporate Author:

NAVAL POSTGRADUATE SCHOOL MONTEREY CA

Personal Author(s):

Report Date:

1992-03-01

Pagination or Media Count:

59.0

Abstract:

This study examines enlisted attrition behavior for the U.S. Coast Guard and develops a model that projects attrition figures. Survival analysis techniques are used to analyze the empirical attrition behaviors associated with an individuals sex, race, marital status, and military occupational skill MOS. In this study males tend to have higher survival probabilities than females, non-caucasians higher than caucasians, and married persons higher than those not married. Aviation MOSs have the highest survival probabilities and technical MOSs have the lowest. Modelled survivor functions are developed for two paygrades because each contain small personnel inventories. These modelled survivor functions do not fit the data as well as desired but are nonetheless used pending the development of sharper alternatives. Finally, a counting model based on the Binomial Distribution is developed that projects monthly enlisted attrition figures. Attrition, survival analysis, Weibull regression model, Binomial counting model, survivor function, accelerated failure time model, Coast Guard Attrition.

Subject Categories:

  • Personnel Management and Labor Relations
  • Statistics and Probability

Distribution Statement:

APPROVED FOR PUBLIC RELEASE