ERIC Number: ED222503
Record Type: RIE
Publication Date: 1981-Oct
Reference Count: 0
Reducing Adverse Impact Via a Measure of Applicant Disadvantagedness.
Mann, Walter G.
Recent attempts to reduce the adverse impact of examinations have focused on alternatives to written tests. The present report, however, demonstrates how the adverse impact of written tests can be reduced by correcting for the degree to which a job applicant had been educationally and/or economically disadvantaged or deprived. A measure of disadvantagedness containing nine items was internally and externally validated with 11,931 applicants for a nationwide examination. Internal validation was demonstrated by a factor analysis that yielded two factors. External validation was based on the relationships of scores on the two factors with other variables: minority status, test performance, and educational level; each assumed relationship was confirmed. A composite measure of the two factor scores, called D, had a point biserial correlation of -.31 with passing the test and .54 with minority status. The adverse impact of the test was substantially reduced by partialling D out of test performance. It was also demonstrated that partialling D out of test performance would not necessarily reduce the validity of the test and could actually improve it. (Author/PN)
Publication Type: Reports - Research
Education Level: N/A
Authoring Institution: N/A
Identifiers: Adverse Impact; Internal Validity