Department

University of Tennessee at Chattanooga. Dept. of Psychology

Publisher

University of Tennessee at Chattanooga

Place of Publication

Chattanooga (Tenn.)

Abstract

Retaining qualified active duty members in the military is an essential mission for DoD. This research presents findings on the relationship between an active duty member’s plans to stay on active duty (as indicated on a survey) and the member's actual retention behavior in the U.S. military two and four years later. Retention plans, as measured on the DoD’s Status of Forces Surveys, have often been interpreted as an indicator of subsequent retention behavior, but the relationship between survey responses and actual retention behavior has not been verified using actual retention data. This study seeks to examine this relationship. Further, in cases in which a service member indicated high retention plans, but did not remain in the Military, it may be valuable to explore factors that have led to this change. Recent transactional data of U.S. military separation codes was restructured among time dimensions and military rank criteria and compared against prior survey data measuring retention plans. Employing weighted logistic regression analyses, retention plans were used to predict a dichotomous outcome of actual retention behavior (i.e., “retained” or “left”). Moreover, predicted probabilities were used to show how actual retention behavior varies as members indicate higher and lower levels of plans to remain on active duty. Analyses showed that an active duty member’s plans to stay on active duty had a positive relationship with actual retention behavior two years later and a stronger relationship four years later. Chief among the findings were that Marine Corps members and junior enlisted member were less likely than members in other Services and paygrades to remain in the Military after two and four years. However, the relationship between retention plans and actual retention behavior was stronger among these subpopulations (Marines, junior enlisted) than members in other Service and paygrades. This research is part of a larger body of on-going research that seeks to implement data wrangling techniques to merge survey data with transactional administrative data, augmenting the use of survey items in predicting retention outcomes. Future research will focus on identifying factors related to a Service member indicating high retention plans but separating from the Military (i.e., those who, at one point, intended to stay, but separated 2-4 years later). Findings may support the Department of Defense in tailoring its policies and programs (i.e., promotional opportunities, family policy, etc.) to better support retention.

Date

October 2018

Subject

Industrial and organizational psychology

Document Type

presentations

Language

English

Rights

Under copyright.

License

http://creativecommons.org/licenses/by/3.0/

Share

COinS
 
Oct 27th, 1:30 PM Oct 27th, 2:30 PM

Using Past Surveys of Attitudes to Predict Current U.S. Military Retention

Retaining qualified active duty members in the military is an essential mission for DoD. This research presents findings on the relationship between an active duty member’s plans to stay on active duty (as indicated on a survey) and the member's actual retention behavior in the U.S. military two and four years later. Retention plans, as measured on the DoD’s Status of Forces Surveys, have often been interpreted as an indicator of subsequent retention behavior, but the relationship between survey responses and actual retention behavior has not been verified using actual retention data. This study seeks to examine this relationship. Further, in cases in which a service member indicated high retention plans, but did not remain in the Military, it may be valuable to explore factors that have led to this change. Recent transactional data of U.S. military separation codes was restructured among time dimensions and military rank criteria and compared against prior survey data measuring retention plans. Employing weighted logistic regression analyses, retention plans were used to predict a dichotomous outcome of actual retention behavior (i.e., “retained” or “left”). Moreover, predicted probabilities were used to show how actual retention behavior varies as members indicate higher and lower levels of plans to remain on active duty. Analyses showed that an active duty member’s plans to stay on active duty had a positive relationship with actual retention behavior two years later and a stronger relationship four years later. Chief among the findings were that Marine Corps members and junior enlisted member were less likely than members in other Services and paygrades to remain in the Military after two and four years. However, the relationship between retention plans and actual retention behavior was stronger among these subpopulations (Marines, junior enlisted) than members in other Service and paygrades. This research is part of a larger body of on-going research that seeks to implement data wrangling techniques to merge survey data with transactional administrative data, augmenting the use of survey items in predicting retention outcomes. Future research will focus on identifying factors related to a Service member indicating high retention plans but separating from the Military (i.e., those who, at one point, intended to stay, but separated 2-4 years later). Findings may support the Department of Defense in tailoring its policies and programs (i.e., promotional opportunities, family policy, etc.) to better support retention.