Identifying Optimal Keys to Enhance Personality Scale Validity
Abstract
The U.S. Army Cadet Command uses personality scales that have been validated against the Cadet Order of Merit List metric to help award U.S. Army Reserve Officer Training Corps scholarships. These personality scales use a conventional algorithm to compute scale scores (i.e., the scale score is set equal to the mean item rating with some item ratings reversed for direction). To improve the utility of these measures, the U.S. Army Research Institute for the Behavioral and Social Sciences (ARI) is evaluating alternate procedures that can be used to key and score these personality scales. Using a cross-validation design, analyses were conducted using training samples to identify the optimal, integer-based scoring key for five personality scales for predicting the Cadet Order of Merit List metric (i.e., each possible key in the universe of scoring keys was evaluated for each scale). Subsequent analyses using validation samples and profile similarity metrics demonstrated that the validity for four of the five personality scales was improved by using optimal keys identified from the training samples. For a battery of five personality scales, regression analyses using the validation samples demonstrated that the composite validity of the personality battery significantly and substantially increased over the use of conventional scores based on the standard keys (R = .48 vs R = .32).
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 2021
- Accession Number
- AD1147892
Entities
People
- Benjamin S. Kerner
- Oren R. Shewach
- Peter J. Legree
Organizations
- George Mason University
- Human Resources Research Organization
- U.S. Army Research Institute for the Behavioral and Social Sciences