Detecting and Measuring Improvements in Validity
Abstract
The Navy Personnel Research and Development Center (NPRDC) has developed new aptitude tests for possible addition to the Armed Services Vocational Aptitude Battery (ASVAB). Validity studies are currently underway to determine whether the new tests produce an increment in validity using either a training or a job performance criterion. Because most samples are small, it is necessary to pool information across sites to obtain a sufficiently powerful test for incremental validity. Methods for obtaining the sampling distributions of incremental validities (the differences between multiple correlations) from the same and from independent samples were developed and are presented. These results were applied to yield methods for pooling incremental validities, testing the statistical significance of pooled validities, constructing confidence intervals for the pooled incremental validity, and conducting a power analysis of the pooled test for incremental validity. It was concluded that the test for the statistical significance of the pooled estimate should have adequate power to detect increments in validity of .02 given pooled sample sizes of N greater than or equal to 4,000.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1992
- Accession Number
- ADA257446
Entities
People
- Betsy J. Becker
- John H. Wolfe
- Larry V. Hedges
Organizations
- Battelle Memorial Institute