A comparison of adverse impact levels based on top-down, multisource, and assessment center data: Promoting diversity and reducing legal challenges
We compared levels of adverse impact (AI) against minorities and women based on three promotional decision methods for 428 associate store managers of a Fortune 500 retailer: top‐down appraisal (TDA), multisource appraisal (MSA), and an assessment center (AC). We found significant effects for race a...
Saved in:
Published in | Human resource management Vol. 51; no. 3; pp. 313 - 341 |
---|---|
Main Authors | , , |
Format | Journal Article |
Language | English |
Published |
Hoboken
Wiley Subscription Services, Inc., A Wiley Company
01.05.2012
Wiley Periodicals Inc |
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | We compared levels of adverse impact (AI) against minorities and women based on three promotional decision methods for 428 associate store managers of a Fortune 500 retailer: top‐down appraisal (TDA), multisource appraisal (MSA), and an assessment center (AC). We found significant effects for race and minority status (favoring whites) but no significant effects for gender across all three methods. Comparisons were analyzed using two definitions of AI (four‐fifths rule and the Fisher Exact Test) with a selection ratio (SR) of .25. No evidence of AI was found against blacks for any of the methods. For Hispanics, AI was found under the four‐fifths rule for the AC. For women, the four‐fifths rule was violated using TDA. Results provide some partial support for the argument that diversity goals are more likely to be met using assessment centers compared to TDA when the preponderance of raters are white men. Our data also suggest that carefully crafted, job‐related, top‐down appraisal, combined with rater training, may be effective in facilitating the promotion of more women and minorities and reducing the probability of legal problems. © 2012 Wiley Periodicals, Inc. |
---|---|
Bibliography: | ArticleID:HRM21472 istex:7099435760DBBF90A4E04624EBE880D2D5D779CB ark:/67375/WNG-33V3RT7P-H |
ISSN: | 0090-4848 1099-050X |
DOI: | 10.1002/hrm.21472 |