VALIDITY AND FAIRNESS OF SOME ALTERNATIVE EMPLOYEE SELECTION PROCEDURES¹

Abstract

Despite extensive evidence that tests are valid for employee selection, Federal Guidelines have urged employers to seek alternative selection procedures that are equally valid but have less adverse impact on minorities. Research on the validity, adverse impact and fairness of eight categories of alternatives was reviewed. Feasibility of operational use of each type of alternative in an employment setting was also discussed. Only biodata and peer evaluation were supported as having validities substantially equal to those for standardized tests. Previous reviews and more recent research indicated that interviews, self-assessments, reference checks, academic achievement, expert judgment and projective techniques had levels of validity generally below those reported for tests. Data, where available, offered no clear indication that any of the alternatives met the criterion of having equal validity with less adverse impact. Results are discussed and several additional promising alternatives are described.SINCE the first validation studies reported by Munsterberg, specialists in personnel selection have relied heavily on standardized psychological tests. The usefulness of standardized tests for personnel selection is strongly supported. Ghiselli, in his 1966 book, The Validity of Occupational Aptitude Tests, and in a 1973 Personnel Psychology article, summarized the results of hundreds of criterion related validation studies including tests in five major categories: (1) intellectual abilities, (2) spatial and mechanical abilities, (3) perceptual accuracy, (4) motor abilities, and (5) personality tests. Most of ' The authors would like to thank all those members of APA Division 14 who so kindly shared their research findings with us. We would particularly like to thank Mary Tenopyr for her help and advice throughout.

A Review of the Literature on Attrition from the Military Services: Risk Factors for Attrition and Strategies to Reduce Attrition

“…As noted above, the correlation between attrition and biodata questionnaire scores is about 0.30 for military questionnaires that have been cross-validated (153,212). Similar results have been reported for other non-military questionnaires when predicting training success (53).…”

Development of Experimental Army Enlisted Personnel Selection and Classification Tests and Job Performance Criteria

“…Biodata tests are self-report multiple-choice questionnaires that attempt to measure the testtaker's prior behavior, experiences, and reactions to life events. Meta-analyses of the selection literature show that biodata effectively predict a wide variety of performance criteria (e.g., ratings of overall performance, advancement potential, commendations, sales volume, bonuses), with typical estimated validities in the .30s to .40s (Hunter & Hunter, 1984;Reilly & Chao, 1982;Schmitt, Gooding, Noe, & Kirsch, 1984). In addition to being useful as an initial selection screen, biodata instruments achieve similar validity estimates for predicting various criteria of supervisory and managerial performance (Owens, 1976;Reilly & Chao, 1982).…”

Faking biodata tests: Are option-keyed instruments more resistant?

“…Reviews by Reilly and Warech (1990) , Hunter and Hunter (1984) , Schmitt, Gooding, Noe, and Kirsch (1984) , and Reilly and Chao (1982) affirm the validity of biodata across a wide variety of jobs and criteria. As biodata instruments become more widely used, however, concerns have arisen regarding the accuracy of job applicants' self-reported data (see Fleishman, 1988 ).…”