One of the most widely reported and talked about studies in recent years has been that of an MIT/University of Chicago experiment that sought to find racism in employment. The name of the study is titled: “Are Emily and Greg More Employable than Lakisha and Jamal?” The provocative title is very important, as we will discover later, and for goodness sake MIT did it, so it must be bulletproof. The study in their own words does the following:
“We study race in the labor market by sending ﬁctitious resumes to help-wanted ads in Boston and Chicago newspapers. To manipulate perceived race, resumes are randomly assigned African American or White sounding names. White names receive 50 percent more callbacks for interviews… Deferential treatment by race still appears to still be prominent in the U.S. labor market…”
This study was published by Marianne Bertrand and Sendhil Mullainathan in 2003 and then again in 2004 (not exactly sure why, they didn’t bother to explain). There are a few statistical differences between the two but the same conclusions are made in both. Normally publishing a study for the second time with slightly different information would warrant a comment or explanation as to why there was a revision, the latest study offers no such thing. It is not clear why, what is clear is that this is the least of their problems.
The study proclaims that white sounding names receive 50% more callbacks than black ones and later explains that on average a person with a black sounding name will have to seek employment for 15 weeks rather than just 10 for whites.
This study has been cited hundreds and hundreds of times by “scholarly” publications and articles. The New York Times reported the results of this study along with CBS news and it was part of the Freakonomics documentary. It is a favorite of professional racialists like Tim Wise who has been citing this study since 2003 in nearly all of his speeches including TV appearances, debates, lectures, multiple articles, and his books. In fact, Wise uses this study as proof that President Obama who surely knows the “truth” according to Wise was lying and pandering to the white majority in 2004 when he said at the Democratic National Convention “We’re just not a Black America and a white America and a Latino and Asian America. It’s just the United States of America.” To Wise, it is impossible for an honest person with knowledge of this employment discrimination study to think America is anything but a deeply racist nation.
It is essentially the go-to proof of widespread discrimination for people like Wise: it is simple and easy to understand and it can be shocking to some that something so common like applying for a job is so entrenched with racism. But in this 42 page study, the authors Bertrand and Mullainathan reveal on page 39 how utterly useless this study is in measuring racial discrimination. In Table 8 it becomes clear that the study essentially picked 9 names for each gender and race and then documented the call back percentage for each black/white sounding name, without the understanding that they will inevitably have some names just randomly doing better than others. In fact the only reason for the disparity was one or two randomly high or randomly low call back percentages for each group. Aisha, which is apparently a black sounding name, got 2.2% call back which killed the overall call back percentage for black women. I can’t imagine employers racially discriminating against someone named Aisha (2.2%) but not someone named Ebony (9.6%, the name literally means black); it’s more likely that Aisha just wasn’t very lucky, but the authors don’t address this at all. The authors of the study found real resumes online that were apparently comparable with each other, but maybe Aisha’s resume lacked something that Ebony’s didn’t. This is the obvious explanation of the difference between Ebony and Aisha, but the authors won’t admit this because if they do, their entire study is undermined due to the obvious fact that you can’t have a comparison of two people with multiple differences but attribute their discrepancy to whichever one variable is convenient.
The authors also need to explain the disparity of “White sounding names” like Emily (7.9%) and Kristen (13.1%) and what kind of discrimination this seemingly illustrates, or the apparently insidious deep seeded Brad (15.9%) supremacy over the Todd (5.9%) race. Perhaps the most embarrassing fact of this study is its title, “Are Emily [7.9%.] and Greg [7.8%] More Employable than Lakisha [5.5%] and Jamal [6.6%]?” This is supposed to be obvious that yes of course Emily and Greg are more employable that Lakisha and Jamal, duh, we live in a racist country. But if the title simply used different names within the same study then the title would convey the exact opposite, for example “Are Emily (7.9%) and Todd (5.9%) More Employable than Ebony (9.6%) and Jermaine (9.6%)?” Or in other words this study is tainted and obviously not scientific but clearly agenda driven. The specific names/resumes made the difference, not race. If some of these names received lower call back percentages due to anti-Emily/Aisha sentiments or these names just randomly did not get called back as often as others, it doesn’t prove racism because the result would be entirely different if different names/resumes had been used. Ashley (not used in the study) could have gotten .5% or 20% and changed the results of the study completely, making this study less science and more editorial.
The study boasts the use of nearly 5,000 resumes sent to conduct the experiment, yet for an unexplained reason only about a quarter of the resumes (1,124) were split between white and black males. Not surprisingly males had a much lower call back difference between whites and blacks, effectively rigging the results. If half of the resumes tested would have been male, using their statistics, then the study would have a far different result.
Of the 1,124 male resumes 575 were white and 549 were black, both of which had 9 names used each for the study. Assuming that these resumes were evenly split between each name, this means 63 white resumes were sent for each white name and 61 for each black name. So taking the names used in the title of the study, Greg (7.8) and Jamal (6.6), we see a grand total difference in callbacks equalling (brace yourself)… 1. White name Greg received 5 callbacks while black name Jamal got 4. This is representative of all the male names, the difference is so small that making them into percentages and comparing them is extremely deceptive. So Greg got about 25% (1) more callbacks than Jamal. From that percentage figure we can just make up a situation that sounds scary, and that’s exactly what they did. For example we can say if it takes Greg 10 weeks to find a job, it will take Jamal an additional two and a half weeks of looking. It is a completely fictitious figure based on the percentage, which in reality was only the difference of one more callback.
This study was published by the National Bureau of Economic Research along side another study that focused on the role of names in the labor market. The other study titled “The Causes and Consequences of Distinctively Black Names” lead by Roland G. Fryer is much less known, in fact according to “Google Scholar” which keeps track of how many times a particular study is cited in other studies and books, Fryer’s study has been cited about 160 times compared to the over 800 times Bertrand and Mullainathan’s study has been cited.
Fryer’s study analyzed 16 million births using census data in California between 1960 and 2000, and found no significant effect on how someone’s life turns out based on one’s name. This study is objectively more extensive, concrete, and scientific than Bertrand and Mullainathans and comes to a completely different conclusion. Yet Fryer’s study is largely ignored despite being published along side Bertrand and Mullainathan’s. Of course the reason for this is that the result of Fryer’s study may lead people to believe we are not all racists, which is boring and does not reinforce the racialists creed. Instead of being relieved by Fryer’s conclusion, racialists ignore it and choose to promote the study that will shock people and gain media exposure.
This is yet another example that racialists will simply believe whatever they want, and repeat it enough times until it is dogma.