As companies get increasingly sophisticated about collecting and analyzing data, they can apply those skills to employee selection. But the selection criteria have to be consistent with fair employment under the law. Here are some tips for smart hiring that is also legal:

• Understand the jobs, company, and data well enough to develop a model explaining why criteria matter. A company that collects a lot of data can eventually find relationships between some variables and superior performance. Suppose you find that people from certain zip codes are less likely to quit. If you don’t dig further into what it is about people from those zip codes, you could wind up making hiring choices that show a pattern of discrimination—and don’t select for important qualities behind the numbers.

• Use analytics to test whether assumptions about job requirements really are relevant. A company that routinely selects for workers with a 3.5 grade-point average can gather data to see whether this requirement is actually associated with success. Perhaps someone with a 3.0 average and certain experiences performs even better.

• Use analytics methods to remove irrelevant criteria from the selection process. For example, having a computer do the initial screening removes the potential for unconscious bias related to factors like people’s names or photographs.

• Create robust processes. Just knowing that you want certain characteristics doesn’t mean the company will hire great people. Make sure the selection tools, including questions asked in interviews, are effective (reliable, valid, and so on). Ensure that the process is efficient and treats all candidates with respect.

• Apply the analytic methods to all applicants. If applications submitted online are screened by a computer system, then applications submitted on paper or in person should be screened using the same criteria.

Sources: Lydell C. Bridgeford, “Experts Discuss Big Data’s Effect on Hiring, Bias Claims,” HR Focus, September 2015, pp. 4–6; Kurt Naasz, “Advances in ‘Big Data’ and Analytics Can Unlock Insights and Drive HR Actions,” HR Focus, May 2015, pp. 1–4; Murad Hemmadi, “The End of Bad Hiring Decisions,” Canadian Business, January 2015, EBSCOhost, http://web.a.ebscohost.com.

Suppose a manufacturer using analytics learns that employees who graduated from one technical college are slower to get up to speed than employees who graduated from another technical college. How might applying the data to hiring decisions raise the risk of discrimination?