RACE VIII: (MORE) CRITICISM OF THAT EMPLOYMENT STUDY: I gave a long, rambly description of a recent study attempting to measure anti-black discrimination in employment, here. Reader John Brewer thinks I was way too easy on the study; he makes a lot of good points. Everything that follows is by him.

It is difficult to take this study seriously, even by the standards of what passes for publishable social-science research. (There is also an interesting-at-least-to-me lurking question about the morality of systematically lying to people who do not know that they are being used as unpaid subjects for a social-science experiment. What did they say to the employers who invited someone in for an interview? “Sorry, I’m not real, I’m just an invention of some academics who were hoping to catch you being racist”?)

Using people with exotic-sounding names as a proxy for the entire universe of black potential job applicants is both bizarre and, I would think, potentially somewhat insulting to the non-exotically-named majority of the black population. What would you think of someone who ran a study using live “tester” applicants with purportedly identical credentials, 10 blacks and 10 whites, but then made sure that all of the black applicants had dreadlocks (or wore dashikis, or something equivalently “ethnic,” to their interviews) but that all of the white applicants had

“normal” hairstyles and/or clothing? Would you think it plausible to attribute any statistical difference in the results of the interviews to “racism” in the first instance? (Alternatively, what if they classified any applicant wearing a suit as “white” even if he was, in fact, black, and somehow structured the interviews so that the interviewer could see the clothes but not the skin? Their study treats, e.g., the late Jam Master Jay as having a stereotypically “white” name, which seems implausible. With Jam Master Brendan, maybe they’d have a point.)

[He later added (sorry for caps, I don’t have time to retype this):] THE QUESTION IS WHETHER THEIR “WHITE” NAMES ARE ACTUALLY “PERCEIVED” AS BEING WHITE, OR WHETHER IT’S AN INTERESTING STATISTICAL FACT THAT DOES NOT ACTUALLY AFFECT THE PERCEPTIONS OF THE RESUME-REVIEWER. NOTE THAT THEY STRUCK CERTAIN NAMES THAT WERE STATISTICALLY LIKELY TO BE BLACK (JEROME AND MAURICE, E.G.) BECAUSE THEY THOUGHT THEY WOULDN’T BE PERCEIVED AS SUCH BY THOSE EVALUATING THE RESUMES. THE SAME PROBLEM SHOULD HAVE, I THINK, LED THEM TO STRIKE VIRTUALLY THEIR ENTIRE LIST OF “WHITE” NAMES. WITH THE EXCEPTION OF BRENDAN AND MAYBE GEOFFREY (SPELLED THAT WAY VS. JEFFREY) NONE OF THEIR WHITE NAMES STRUCK ME AS “PERCEPTUALLY” WHITE, IN THE SENSE THAT IT WOULD SEEM ODD, NOTEWORTHY OR COMICAL TO COME ACROSS A BLACK PERSON BEARING THEM, WHILE IT WOULD BE ODD ETC. TO RUN INTO A WHITE (BUT NON-ARAB) AISHA. I THINK THE ONLY WAY YOU COULD EFFECTIVELY COME UP WITH “PERCEPTUALLY WHITE” NAMES WOULD BE TO USE OVERTLY ETHNIC NAMES (LUDMILLA, SVEN, ETC.). I THINK BRENDAN (AS OPPOSED TO THE YUPPIE FAVORITE BRANDON) IS STILL LARGELY AN ETHNIC NAME (I.E. I CANNOT OFF THE TOP OF MY HEAD THINK OF A NON-IRISH BRENDAN WHEREAS I CAN IMMEDIATELY THINK OF A BLACK TODD).

THUS, I APPRECIATE WHAT THEY WERE TRYING TO DO IN TERMS OF TRYING TO GENERATE NAMES THAT WOULD CONNOTE RACE, I JUST DON’T THINK THEY DID IT VERY WELL AND IT MAY BE IMPOSSIBLE TO DO (UNLESS YOU WANT TO COMPARE “ETHNIC” WHITES TO BLACKS RATHER THAN “REGULAR” WHITES TO BLACKS). THAT IS, I WOULD SAY, A FEATURE RATHER THAN A BUG. ON THE WHOLE (AND WITH NO DISRESPECT TO THE LAKISHAS AMONG US) THE FACT THAT MOST AMERICAN BLACKS (AND AN EVEN GREATER MAJORITY OF THOSE BORN PRIOR TO THE LAST FEW

DECADES) HAVE NAMES WHICH COULD BE BORNE BY “REGULAR” AMERICAN WHITES WITHOUT IT SEEMING WEIRD IS A HELPFUL REMINDER THAT WE AND THEY HAVE A COMMON CULTURE AND A COMMON HISTORY WHICH CANNOT BE REDUCED SIMPLY TO AN IMPOSITION BY US ON THEM (WHICH WOULD DENY WHAT WE HAVE RECEIVED IN RETURN). WHITES AS WELL AS BLACKS CAN GIVE THEIR KIDS THE CHRISTIAN NAMES BORNE BY BENJAMIN BANNEKER, HARRIET TUBMAN, ETC. ETC. IF YOU WOULD PREFER HOLLYWOOD CELEBRITIES TO HISTORICAL ACHIEVERS, YOU COULD E.G. NAME YOUR SON AFTER WESLEY SNIPES (WHO COMPLETES THE BLACK SHOWBIZ PROTESTANT BOYS’ NAME TRIFECTA WITH CALVIN BROADUS AND LUTHER VANDROSS).

THE QUESTION IS WHETHER THE SIGNIFICANT NUMBER OF BLACKS WITH SUCH [distinctively black] NAMES ARE A RANDOM SAMPLE / FAIR CROSS-SECTION OF THE BLACK COMMUNITY OR WHETHER, ON AVERAGE WITH LOTS OF INDIVIDUAL COUNTEREXAMPLES, THEY HAVE A DIFFERENT PROFILE (EITHER IN REALITY OR IN PERCEPTION). IF THE LATTER, THE REST OF THE POPULATION WOULD BE JUSTIFIED IN BEING ANNOYED THAT STATEMENTS ABOUT THAT SUBGROUP AND HOW SOCIETY DEALT WITH THAT SUBGROUP WERE BEING ATTRIBUTED TO THE REST OF THEM. I WOULD BE ANNOYED IF SOMEONE DID A SURVEY OF FANS AT A NASCAR RACE AND REPRESENTED IT AS INDICATIVE OF THE BELIEFS OR EXPERIENCES OF “WHITE MALE REPUBLICAN CHRISTIANS” EVEN THOUGH (AS A NON-NASCAR-RACE-ATTENDING WMRC) I DON’T THINK IT’S “BAD” TO BE A NASCAR FAN. THE MOST INTERESTING QUESTION, I SUPPOSE, IS WHETHER THE BLACK PARENTS WHO GIVE THEIR KIDS THESE KIND OF NAMES SYSTEMATICALLY DIFFER FROM BLACK PARENTS WHO GIVE THEIR KIDS “REGULAR” NAMES IN WAYS THAT ARE NOT AS EASY TO MEASURE AS INCOME, EDUCATION LEVEL, AND MARITAL STATUS — IN PARTICULAR, IN ATTITUDES TOWARD ASSIMILATION/INTEGRATION VERSUS SEPARATISM/NATIONALISM. FOLLOWUP QUESTION IS WHETHER ANY SUCH DIFFERENT ATTITUDES LEAD TO MEASURABLE DIFFERENCES IN HOW THE KIDS ARE LIVING THEIR LIVES TWO OR THREE DECADES DOWN THE ROAD.

To make an even more basic criticism, it is not clear to me that the results show any sort of discrimination (even what I might call normal-versus-weird discrimination rather than white-versus-black discrimination). Table 3 categorizes 87.37% of the employers as showing “equal treatment” with 8.87% “favoring” whites and 3.76% “favoring” blacks. One would need to know more about statistics than I do to know how likely the gap between the latter two numbers is meaningful rather than noise. One can also report the same results in different fashions: An alternative headline could be: “Study shows that 57% of all employers who extended callback invitations to any applicants included at least one black applicant in the group of applicants extended such invitations.”

[later:] I’M NOT SURE THE 57% STAT IS MEANINGFUL. BUT I’M NOT SURE IF ANY OF THEIR STATS ARE MEANINGFUL. IF YOU HAD AN APPLICANT POOL THAT WAS 30% BLACK AND BLACKS AND WHITES WERE EQUALLY QUALIFIED, LACK OF ANY BLACKS IN THE LAST 100 HIRES IS PLAUSIBLE STATISTICAL EVIDENCE OF DISCRIMINATION. HIRING 28 OR 29 RATHER THAN EXACTLY 30 IS NOT. WHERE IN BETWEEN ZERO AND 28 THE RESULT BECOMES SUFFICIENTLY STATISTICALLY UNLIKELY AS THE RESULT OF RANDOM SELECTION TO SUPPORT A REASONABLE INFERENCE THAT SOMETHING ELSE IS GOING ON, I DON’T KNOW, BUT THERE ARE MATHEMATICAL WAYS OF APPROACHING THAT QUESTION WITH SOME DEGREE OF RIGOR. I LIKEWISE DON’T KNOW WHAT THE EQUIVALENT DEGREE OF DEVIATION FROM PERFECT EQUALITY THAT SUGGESTS SOMETHING OTHER THAN RANDOMNESS WOULD BE IN THEIR EXPERIMENT. I WOULD FEEL BETTER IF THEY HAD A FOOTNOTE WITH SOME INCOMPREHENSIBLE FORMULA IN IT THAT ADDRESSED THIS ISSUE.

I am struck by the fact that the gaps across the “races” seem to be significantly smaller than some of the gaps within the groups. E.g. Brads got more than twice as many callbacks as Neils. I presume that either sheer randomness or some explanation other than systematic anti-Neil (and/or pro-Brad) discrimination likely accounts for this. I therefore see no immediate reason to presume that racism is the most likely explanation for the smaller gaps between the “races.” Note also that, e.g., while Emily does better than Lakisha she does worse than Latoya and Latonya. The title of the study raises the first comparison but not the second and third.

Even assuming the differences they show are statistically meaningful evidence of some kind of discrimination, I think one would want to seriously consider normal-name-versus-weird-name bias as an equally plausible

hypothesis. Perhaps there is no prior research or scholarly literature on the subject, but it would seem pretty easy to construct similar experiments without a racial subtext. For example, one could compare Bretts, Todds etc. with weird-but-presumptively-white names (Percy, Thurston, Winifred, Cletus, Jethro, Merle, Adrian, Iain, & Nigel — to pick three names each that are exotic in three different ways). One could also test a different hypothesis in other ethnic contexts by testing contrasting pairs of what might be called assimilation-oriented versus non-assimilation-oriented names (I think this is a reasonably polite way to phrase the “ghetto culture” point): Jason Goldstein v. Shlomo Goldstein; Brandon Lopez v. Raul Lopez; Joseph Martino v. Giuseppe Martino. Finally, one could test a generational-novelty hypothesis by seeing how Ashleys, Brittanies & Kylas do when compared to similarly-qualified peers who have more traditional names. (The key to the hypothesis here is

that the hiring decisionmaker may be a decade or two older than the applicants and discriminate at the margin, perhaps subconsciously, against names that were not extant in his/her own cohort. Most of the “black” names they used only came into use for kids born in the 70’s — as can be confirmed both by the babyname statistics on the social security administration’s website and a look at the black kids in my high school yearbook, who were born circa ’65.).

I would not be surprised if any of the studies I have just suggested detected a small but perceptible bias — certainly a bias of the same statistical magnitude as that ascribed to “white” versus “black” names — in favor of “normal” versus “weird” names in resume evaluation. Unless such studies were conducted and affirmatively demonstrated the absence of any such effect, I would not take this study seriously as a demonstration of racism rather than a bias in favor of normality.

One can, I suppose, argue that discrimination against dreadlock and dashiki-wearers (and thus a fortiorari discrimination against the Afrocentrically-named) is just as immoral as discrimination against black skin — this is the shift from the rhetoric of colorblindness to the rhetoric of multiculturalism. Saying that a non-celebrity Calvin Broadus would be marginally more likely to get an entry-level job at a convenience store by submitting a resume with that name rather than a resume with his assumed name (Snoop Dogg) would, on this account, be an indictment of a racist society in which blacks can succeed only by denying or disguising their essential Kwanzaaness and adopting the outward forms of the oppressors’ society. Whether or not one agrees with this, the problem of discrimination against black skin per se has been bad enough in the past that it would be helpful to have non-bogus and non-loaded data on the extent to which it has or has not been surpassed. This study is

not, in that regard, helpful.