The New Art Science of Selecting Employees - part 3

Three Criteria to Establish
a Legal Selection Technique

The basis for non-discriminatory hiring rests on three principles.

First, the selection process should not have a discriminatory impact, unless, second, there is a valid justification of business necessity. Finally, if there is discriminatory impact, even with a valid business justification, and there is another equally valid alternative selection method that would have no discriminatory impact, then the alternative should be used:

Non-Discriminatory Impact

Discriminatory impact, on the surface, is straightforward. Any selection process, including recruiting, background checks, interviewing, or testing, etc., that results in "protected classes" being hired or promoted significantly less often than non-protected classes is discriminatory.

Significant impact, as a rule of thumb, means that the percent of protected classes who are hired, promoted, or "pass" a selection criterion is less than 80% of the pass-rate of non-protected individuals who are hired or promoted.

For example, if, as the result of a screening interview, 50% of all white male applicants pass but only 20% of females or "minorities" pass, then the screening interview can be presumed to have discriminatory impact because the pass-rate of minorities is only 40% of white male candidates. If, however, 50% of white males pass and 44% of minorities pass, then the pass-rate of minorities would be 88% of white males and may be presumed to be non-discriminatory, since the 44% pass-rate for minorities is not less than 80% of the 50% pass-rate for white males.

This "Four-Fifths" rule depends on the sample being of sufficient size to be realistic. If three of four white males are hired and only one out of two females are hired from a total of six candidates, then the numbers are too small to make any valid conclusions.

For small companies that have too few employees in a position to validate, any testing procedure that has no discriminatory impact may still be used. For example, Chally's selection test scales are non-discriminatory in and of themselves.

A Business Necessity - Justification For A Discriminatory Selection Technique

In general, a selection technique must be shown to be a valid predictor of success on the job to demonstrate that it is consistent with business necessity. This, in turn depends on two factors:

1. Job Relevancy: The factor of position criterion being tested or evaluated is an essential (for disabled) or meaningful (for other protected class legislation) aspect of the process.
   
   
2. Validity: The measure or test does predict who will perform that job-relevant factor (skill) successfully and who will not.

Lack Of Non-Discriminatory Alternative With Equal "Business Necessity" Value

When a selection practice has a discriminatory impact but can also be proven to be job relevant, i.e., valid and consistent with business necessity, then a complainant must demonstrate that a non-discriminatory alternative selection technique is equally valid and consistent with business necessity.

For example, if a background check is job relevant but discriminatory, and a test is available that correlates equally well with success on the job but does not have discriminatory impact, then the non-discriminatory test is the best legal solution to the selection need.

The Bottom-Line:

It is clear that testing has proven to be the most effective component of a well thought-out selection process, not only because it works, but also because it is legally defensible.

FOOTNOTES:

¹ Chris Lee, "Testing Makes a Comeback," Training, Vol. 25, Dec. 1988, p. 49.

² Ibid, p. 50.

³ Peter Brimelow and Leslie Spencer, "When Quotas Replace Merit, Everybody Suffers," Forbes, Feb. 1993, p. 80.

⁴ E. James Randall and others, "A Successful Application of The Assessment Center Concept To The Salesperson Selection Process," Journal of Personal Selling & Sales Management, May 1985, p 53.

⁵ Griggs vs. Duke Power Co. -- US Sup Ct (1971), 3 FEP Cases 175.

⁶ American Society of Personnel Administrators, Resource, 1988.

⁷ "Testing Report," Human Resource Executive, May 1992.

⁸ Brimelow and Spencer, p. 82.

⁹ Human Resource Executive, p. 46.

¹⁰ John & Ronda Hunter, "Validity and Utility of Alternative Predictors of Job Performance," Psychological Bulletin, July 1984, p. 90.

Back to "Classic Articles" page