Predicting US Army First-Term Attrition After Initial Entry Training
Abstract
The goal of this research is to identify demographic and administrative factors of active component, first-term enlisted soldiers who have completed their Initial Entry Training to construct models to predict the probability of failure in completing their initial contractual obligation. We construct a binary logistic regression model, classification tree and a random forest classification model to predict a soldiers probability of first-term attrition based on the individuals unique service record. We design web based graphic user interface that allows leaders, soldiers or recruiters to input demographic variables of current soldiers or recruits and get the predicted probability of attrition as an output. We find that a soldiers deployment history and the duration of the initial contract are significant predictors of whether a soldier will complete his or her first term. Knowledge of the key factors, and other influencing variables, assists the Army Resiliency Directorate in creation of models and tools to better advise U.S. Army leadership and development of intervention strategies and preventative measures to preclude the loss of first-term soldiers.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 30, 2018
- Accession Number
- AD1063055
Entities
People
- Andrew Anglemyer
- Anthony D. Smith
- Jarrod Shingleton
- Jon Alt
- Karey Speten