The answer is simple: hire HR directors and specialists that understand proper hiring practices to not run afoul of the law. An effective hiring manager realizes that race is neither effective nor legal to use in screening applicants. Once again, the law is already clear on this. If you have an employer with a selection process that unfairly discriminates against blacks (intentionally or not), they are subject to punishment under the law.
Firstly, it's important to remember that this study is not evaluating success rates vs. whites. It is intended to study the success rate of various "whitening techinques" among non-white applicants. I think it is conducted poorly when compared to the research they cite (link below).
To the study: I have a few issues with the methods:
1. The sample size is extremely small considering the scope of employers and geographical area covered. A similar study evaluated 2445 resumes in just two major metro areas. This study evaluates 1600 nationwide.
2. Varied industries and various job types were used for the same generic resume. The study doesn't explain it's randomization or the selection rates across the job types.
3. No employer got more than one resume.
4. There was no control group.
They submitted four names:
Lamar J. Smith vs. L. James Smith (Black vs. "Whitened")
Lei Zhang vs. Luke Zhang (Asian vs. "Whitened")
The OVERALL rejection rate (no callbacks) - 83.3% (which is surprisingly low based on my own experience with resumes). The total number of resumes generating some form of response was ony 267. They applied to 1600 different jobs at 1600 different employers across 6 job types.
From the study:
"Whitening the name only (versus not whitening at all) didnot make a statistically significant difference for black applicants (13% versus 10%) but led to amarginally significant increase in callbacks for Asians (18% versus 11.5%, z = 1.83, p < .10)"
I believe this means they chose too ambiguous a name in both cases. Luke Zhang and Lei Zhang don't suggest any difference in race to the common observer. I have a similar thought to the name Lamar, but I'll grant them that one.
Quoting again:
"If hewhitened both his first name and the experiences, he would receive 2.5 times as many callbacksas he would with the original, unwhitened résumé" - The strongest argument of the study. And a difference of
35 resumes across 1600 (once again, across industries and across employers) is the total difference... I would much rather see the variance between employers, but the university IRB rejected that possibility.
The study shows what it shows, but I don't think it's very useful as it was designed, as there no way to assign causality to any one factor.
http://www.nber.org/papers/w9873.pdf - One of the cited studies,
"Are Emily and Greg More Employable than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination" -
Bertrand and Mullainathan (2004) contains a better design, is much more well-explained, and contains a detailed breakdown of the confounds (page 20-it's worth the read), although the N is still low (2445).