What’s In A Black Name? Ctd

Andrew Sullivan —  Jan 22 2015 @ 7:32am

A reader writes:

An experiment I’ve always wanted to see: mail out resumes with African (e.g. Nigerian) names and see what the response would be. I’d be also curious to see if resumes were sent with photos attached of the same person with the same resume with only the names changed, say Michael Smith vs. Jamar Smith, if there would be any difference. Basically what I’d like to know: is this an actual of pure racial bias, or this more of a cultural bias against people with more stereotypically black names?

Another reader:

One plausible explanation for the discrimination identified by the Gaddis study, other than naked racial animus/aversion, is that employers “discount” the educational achievement of black applicants based on the assumption that affirmative action gave them a boost along the way.

As your previous coverage of racial preferences in college admissions reflects, this is not an entirely unreasonable assumption, though many people do overestimate the effects of such racial preferences. As a quick proof-of-concept, I did a rough comparison of SAT numbers for one of the university pairs Gaddis highlights – Harvard vs. UMass Amherst.  This article from the Harvard Crimson indicates that the average SAT score for all incoming Harvard freshman last year was 2237 (on the 2400 point scale), but only 2107 for incoming African-American freshman.  For incoming White freshman, by comparison, it was 2233 (i.e., roughly the average for all freshman).

The UMass website, meanwhile, indicates an average SAT score for all incoming freshman of 1218 (on the 1600 point scale), which, assuming comparable performance on the essay section, translates to 1827 on the 2400 point scale.  I couldn’t find a racial breakdown of UMass scores, but let’s assume that incoming white students scored roughly the overall average, just as they did at Harvard.

So, when comparing Harvard students to UMass Amherst students in a race-blind hiring process, the average SAT gap would be 410, but when comparing a black Harvard student to a white UMass Amherst student, the average SAT gap is only 280.  That’s roughly a 30% reduction of the achievement premium supposedly signaled by the Harvard name.

Given that the potential employers in Gaddis’ study seemed to discount the achievement premium for African-American Harvard students all the way to zero, there’s obviously more going on here, but given that other studies have shown people tend to significantly overestimate the effects of racial preferences in admissions, I wouldn’t be surprised if somehow controlling for that factor eliminated a meaningful portion of the apparent hiring discrimination.