The United States Employment Service has implemented a recently developed testing program. The Validity Generalization (VG) Testing Program, adopted its name from the meta-analytic technique which cumulates the findings of test validation studies. For this testing program, predictors were developed for five job families based on a validity generalization study of 515 validation studies. The Employment Service claims that these predictors are valid and virtually all jobs are covered in the five job families.
This study is a direct test of the validity of one of the five predictors, Job Family IV Validity Generalization percentile scores. (The Employment Service estimates its true validity is .53.) Secondly, two potential moderators of that predictor's validity were investigated: jobs and locations. Three "computing and account recording" clerical jobs and two locations were examined. Finally, evidence of whether general abilities were better predictors of performance than specific abilities was examined, since the testing program's predictors are comprised of composite, general ability scores.
A concurrent validation study was conducted with 219 clerical bank employees. Two predictors, the General Aptitude Test Battery (GATB) and the American Bankers Association's test battery, were administered. Two criteria measures were obtained, supervisory ratings on the Descriptive Rating Scale for all subjects, and, objective measures - strokes per hour - for proof operators.
The observed validity for Job Family IV's predictor with the global DRS criterion was .16, Observed validity with a composite of DRS dimensions was .19. Corrected for attenuation, those coefficients were .18 and .20 respectively. However, general cognitive ability measures appeared to be slightly better than the percentile scores at predicting performance.
While there was no statistical evidence of moderators, the relatively small effect size resulted in low power for the tests and may account for the results. Nevertheless, the evidence raised questions about the possible existence of situational moderators.
Finally, measures of general ability did not appear to predict performance better than measures of specific abilities. / Ph. D.
| Identifer | oai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/82623 |
| Date | January 1987 |
| Creators | Hoover, David J. |
| Contributors | Management, Madigan, Robert M., Hills, Frederick S., Murrmann, Kent, Scott, K. Dow, Stone, Dianna L. |
| Publisher | Virginia Polytechnic Institute and State University |
| Source Sets | Virginia Tech Theses and Dissertation |
| Language | en_US |
| Detected Language | English |
| Type | Dissertation, Text |
| Format | x, 96 leaves, application/pdf, application/pdf |
| Rights | In Copyright, http://rightsstatements.org/vocab/InC/1.0/ |
| Relation | OCLC# 17336836 |
Page generated in 0.0022 seconds