Estimating Academic Attrition from Technical Training School Data: Method and Simulation Results

Abstract

This study proposes a logistic regression-based approach for estimating academic attrition from technical training school data. The proposed approach enables Army personnel managers to evaluate tradeoffs when making decisions about where to set minimum enlistment standards. A large-scale simulation was conducted based on actual training school data from a selected MOS to evaluate the approach and to assess sampling error of estimated attrition rates under different sample sizes and operational scenarios. Major findings indicate that: (a) a simple approach based on the logistic regression using only cognitive aptitude information is adequate for evaluating impact of changes in minimum enlistment standards on academic attrition for MOS with medium validity or greater; (b) a large enough sample size allows smaller changes in minimum enlistment standards to achieve a targeted attrition rate with confidence; and (c) personnel and training decisions could be greatly improved by extending the current model to incorporate information in addition to cognitive aptitude. The report includes a ready-to-use statistical program for applying the proposed approach to actual training school data for the purpose of making operational decisions.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 2004
Accession Number
ADA426300

Entities

People

  • Mary A. Lightfoot
  • Michael Ingerick
  • Robert Fowler
  • Tirso Diaz

Organizations

  • Human Resources Research Organization

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Applied Psychology
  • Army Personnel
  • Business Administration
  • Computer Programs
  • Databases
  • Demography
  • Education
  • Employment
  • Enlisted Personnel
  • Management Personnel
  • Personnel Management
  • Personnel Selection
  • Psychology
  • Recruiting
  • Students
  • Trainees
  • Training

Fields of Study

  • Computer science

Readers

  • Computational Modeling and Simulation
  • Life Cycle Cost Analysis
  • Psychometric Testing or Psychological Assessment.