NotesFAQContact Us
Search Tips
ERIC Number: ED356272
Record Type: Non-Journal
Publication Date: 1992
Pages: 7
Abstractor: N/A
Reference Count: N/A
Estimating Future Adverse Impact Using Selection Ratios and Group Differences in Test Score Means.
Aamodt, Michael G.; And Others
Estimating the validity of a test is only one concern for the human resources professional developing a personnel selection battery. An equally important concern is whether the test will result in adverse impact against a member of a protected class. It would be useful if the probability of adverse impact could be estimated prior to spending time and money administering the test to an actual applicant pool. This paper presents a table based on the normal curve that uses selection ratios and effect sizes (d scores) obtained from information in test manuals to determine the ratio of the minority selection rate to the majority selection rate. The table is used by finding the point in the table where the selection ratio and the effect size intersect. The number found at this point in the table represents the percentage of minorities that will be selected as a percentage of the non-minority selection ratio. Any number less than the "magical" 0.80 or four-fifths indicates that the selection device will probably result in adverse impact. Table 1 shows minority selection ratio as a proportion of the non-minority selection ratio for effect sizes from 0.01 to 3.00. Table 2 gives effect sizes for preliminary use with the adverse impact table (Table 1) for biodata, interviews, personality tests, honesty tests, ability tests, interest inventories, and values tests. (SLD)
Publication Type: Reports - Evaluative
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A