Revision of the United States Naval Academy Selection Composite

Abstract

This research investigated alternative methods of weighting predictors in the Candidate Multiple, the U.S. Naval Academy's selection composite, to determine whether validities could be improved against criteria of academic and military performance and attrition at the Academy. Using student and applicant data for the classes of 1983, 1984, and 1987, adjustments were made to correct for the effects of selection on predictor-criterion correlations. Corrected validities were then used to construct alternative composites that differed from each other in terms of which predictor weights were predetermined by the Academy and which were free to be empirically optimized. Composites were cross-validated on the class of 1988 and compared with earlier operational versions of the Candidate Multiple. Results indicated that all four experimental composites improve the prediction of Academy criteria as compared to previous operational composites. Keywords: Naval training; Performance (Human).

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 1988
Accession Number
ADA200237

Entities

People

  • Edward F. Alf
  • Idell Neumann
  • Joyce D. Mattson

Organizations

  • Bureau of Naval Personnel

Tags

Communities of Interest

  • Human Systems
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Data Science
  • Education
  • Engineering
  • Information Science
  • Mathematics
  • Psychological Tests
  • Ratings
  • Regression Analysis
  • Schools
  • Standards
  • Statistical Analysis
  • Statistics
  • Students
  • Training
  • United States
  • United States Naval Academy
  • Validation

Fields of Study

  • Education

Readers

  • Psychometric Testing or Psychological Assessment.