Patterico's Pontifications

6/6/2021

The Constitutional Vanguard: Are Emily and Greg Really More Employable Than Lakisha and Jamal?

Filed under: General — Patterico @ 12:26 pm



Big Media has drilled into our heads that employers are racist. One way we “know” this to be true is by reviewing studies that send out resumes with distinctive black names, together with identical resumes with white-sounding names. The resumes with distinctive black names get fewer callbacks.

Totally proves racism, right?

Wrong. That is what today’s newsletter is about.

There is good reason to question whether race is really the dominant factor at issue in these studies. The more likely culprit is socio-economic status.

. . . .

The implicit assumption here is that potential employees who have recognizably black names are likely to possess the same characteristics as those who have names thought to be “white-sounding.” The only difference between the two groups, so the argument goes, is race. Therefore, if you, as a potential employer, are more queasy about hiring a Lakisha than an Emily, the reason is obvious: you are a racist.

The problem here is that the underlying assumption is false. In fact, Roland Fryer and Steven Levitt did a study in 2004, which I bet you never heard about, which found that distinctively black names “provide a strong signal of socioeconomic status” and tend to be characteristic of the subset of black folks “living in racially isolated neighborhoods.”

The newsletter gives you all the links and data you need to refute this common leftist talking point.

Today’s newsletter is for paying subscribers only. The number is growing all the time. You have no excuse not to join up, really. I feel like I should provide the subscription link in the same format you see from those annoying pop-ups:

Yes, I want to subscribe and learn things Big Media won’t tell me!

No thanks, I hate having the data needed to win arguments with dopey leftists

67 Responses to “The Constitutional Vanguard: Are Emily and Greg Really More Employable Than Lakisha and Jamal?”

  1. I really do think that today’s entry is pretty enlightening. I hope y’all enjoy it.

    Patterico (e349ce)

  2. I never thought that distinctively black names was the same thing as race. I don’t even see how anybody could think so. Certain distinctive black names do tend to signal low formal education of the mother – which shouldn’t be held against someone. But when each name is sui generis, there’s going to be a statistical difference in whether or not a person is eliminated.

    Sammy Finkelman (51cd0c)

  3. Things is, they are.

    Especally Big Media.

    At CBS, the unwritten ‘rule’ was, only hire ‘hotties,’ not ‘uglies’ for secretarial spots [preferably blondes] no blacks at reception desks; token quota hires kept to low profile ‘bck room’ departments. ‘Talent’ [that is, the on camera faces, reporters and such] had token hires; literally ‘window dressing;’ but department management was male and white- w/some in-roads by women, chiefly as producers, which is hard-azz-all-hours-work.

    DCSCA (f4c5e5)

  4. The real reason is the terrible schools that the education system has tolerated in the inner cities. In part this is parenting, but it is also the dogged insistence on modes of operation that have been repeatedly proven not to work. And rather than bring things that do work, such as separate tracks for those that want to learn, or separate tracks for the gifted, they are instead removing tracks from every school possible and forcing everyone to learn at the slowest pace, which may well be “not at all.”

    Kevin M (ab1c11)

  5. You know that this won’t keep better-read (but not necessarily less dopey*) leftists from telling you that it only helps confirm the Marx/Bakunin class struggle theory in that there are privileged classes and less privileged classes and they can exist regardless of race. And that they have no more compunction in combining class struggle with racial struggle than Stalin had in combining communism with Russian nationalism.

    *They might not actually believe it themselves. Like all the multi-millionaire socialists we know, with or without a (D) after their name.

    nk (1d9030)

  6. Although what’s a few million these days, right? They might sincerely see themselves as the underprivileged struggling against the billionaires.

    nk (1d9030)

  7. Trump gets it, too. We cannot take that away from him*.

    His Cletuses are the white people Robert Byrd knew a lot of in his time, and they view the minorities of color and foreign birth as their competitors for privilege. And they combine it with meretricious nationalism.

    *But I suspect that he had a head start — largely translating his penis envy into class envy.

    nk (1d9030)

  8. I can’t address the article itself, because I’m not subscribed, but I can tell you what the Fryer and Levitt study says, underneath the noisy data (it is a very messy study).

    Despite their initial implications of a broader group, their group was black women in California, born to unwed mothers in California in 1973 or 1974, who then gave birth between 1989 and 2000. When addressing economics, they did not look at individual wages, jobs, or family wages or jobs, only at the median income in the zip code and whether or not she had private insurance when she had the baby. Generally what they found out was that among black women born to young unmarried mothers who then also had children quite young, black women who had very black sounding first names were only slightly worse off than black women who had more white sounding first names.

    For some perspective, a person born in 1973 or 1974 would’ve graduated HS in roughly 1992. A significant portion of their data is skewing the results, because a person’s first name would not be determining their economic success on their own behalf until they entered the labor force as an adult and the data was taken using the birth records of the child, so at minimum any data collected before the girls would’ve graduated HS in about 1992 is causing significant problems with their results, at maximum for some girls if they were in college at the time of the birth of their child, it could be up to data collected before about 1997 (average time to complete at a UC or CSU was roughly 5 yrs at the time IIRC).

    I understand partly why they used the method they did (information available in birth records became wider over time and the time period spread with the data they wanted was very tight at the time they originally did the study.), but if you really wanted to look at the data they looked at and get closer to real results, you would have to redo the study and look only at women who were old enough to have graduated from HS at the time of the birth of their child.

    Anyway, their study isn’t a great indicator of how much a first name does or does not effect anything.

    Nic (896fdf)

  9. “No thanks, I hate having the data needed to win arguments with dopey leftists”

    Lefties don’t “argue”, they repeat the day’s talking points and denounce. Data is immaterial and usually counterproductive.

    “His Cletuses are the white people Robert Byrd knew a lot of in his time, and they view the minorities of color and foreign birth as their competitors for privilege. And they combine it with meretricious nationalism.”

    ‘MERETRICIOUS NATIONALISM’, eh? nk believes there should be literally no privileges for anyone who’s an American citizen and will gladly slander anyone who believes otherwise as ‘redneck Cletuses’. Certainly no foreign country will ever take advantage of or encourage this attitude!

    This view, needless to say, was not shared by the Founding Fathers, America’s greatest heroes, or anyone who did anything worthwhile for posterity, as most realized basic facts like “in order to get common people to fight, take risks, and work hard in the now, it’s necessary to respect their desires for a country with a purpose beyond ‘built by the lowest bidder’.”

    But it’s quite common among immigration lawyers, Chamber of Commerce greedheads, and anyone in ‘urban planning’ whose paycheck depends on stuffing as many warm bodies in as many tax farms as possible. I’d much rather live well with the descendants of Washington, King, and Jackson than some dude with strong opinions about how ‘national loyalty’ really means that you just live in whatever ‘optimized’ fashion the corporations tell you. But they always find people with any pre-existing loyalties troublesome.

    If you traitors hadn’t stolen the election, Trump would be pushing right now legislation to cancel all USTs in Chinese hands, and Fauci would already be under arrest. But thanks to the Principled People, you can all run your requests for justice by the Psaki Circle Crew. Maybe they’ll listen to your data!

    Psaki Circle (a3731c)

  10. I can’t address the article itself, because I’m not subscribed, but I can tell you what the Fryer and Levitt study says, underneath the noisy data (it is a very messy study).

    I’m not a sociologist, Nic, and I don’t have access to the study itself — one of the frustrating things about writing about such topics is that everything is kept on JSTOR or an equivalent. The best access I have to their actual methodology is this piece, which is linked in the newsletter. I’d like to have a conversation with you about this, though, since you display some familiarity with the study and I think it might benefit my understanding to have an interchange with you about it.

    Despite their initial implications of a broader group, their group was black women in California, born to unwed mothers in California in 1973 or 1974, who then gave birth between 1989 and 2000. When addressing economics, they did not look at individual wages, jobs, or family wages or jobs, only at the median income in the zip code and whether or not she had private insurance when she had the baby. Generally what they found out was that among black women born to young unmarried mothers who then also had children quite young, black women who had very black sounding first names were only slightly worse off than black women who had more white sounding first names.

    Here’s what the article linked in the newsletter says about their methodology:

    Do Fryer and Levitt’s data indicate that having an ethnic name leads to worse adulthood outcomes after controlling for background characteristics? The authors used data on women who were born in California in 1973 or 1974 and later gave birth there by 2000. The authors compared information on the woman’s own birth certificate, which provides information about her socioeconomic conditions at birth, with information about her that is available on her child’s birth certificate, which provides information on the woman’s adulthood economic outcomes.

    You say:

    For some perspective, a person born in 1973 or 1974 would’ve graduated HS in roughly 1992. A significant portion of their data is skewing the results, because a person’s first name would not be determining their economic success on their own behalf until they entered the labor force as an adult and the data was taken using the birth records of the child, so at minimum any data collected before the girls would’ve graduated HS in about 1992 is causing significant problems with their results, at maximum for some girls if they were in college at the time of the birth of their child, it could be up to data collected before about 1997 (average time to complete at a UC or CSU was roughly 5 yrs at the time IIRC).

    So: the study was done in 2004. Can you explain why you think a “significant portion of their data” would have been “collected before the girls would’ve graduated HS in about 1992” — some 12 years before the study was published?? This is the part of your comment that puzzles me the most.

    I understand partly why they used the method they did (information available in birth records became wider over time and the time period spread with the data they wanted was very tight at the time they originally did the study.), but if you really wanted to look at the data they looked at and get closer to real results, you would have to redo the study and look only at women who were old enough to have graduated from HS at the time of the birth of their child.

    Anyway, their study isn’t a great indicator of how much a first name does or does not effect anything.

    The newsletter has other studies published since. Do you think the famous 2004 study showed much? Most of my thesis is that it did not.

    Patterico (e349ce)

  11. If you traitors hadn’t stolen the election

    You sound like Steppe Nomad and in any event you sound like an idiot. Goodbye.

    Patterico (e349ce)

  12. People like Psaki Circle fascinate me. They come across as so aggrieved, and seem incapable of a calm, give-and-take conversation. I am reminded of my late father, who just could not sit down and discuss a problem in a civil manner. He would let things build up, and then explode. As soon as he was done ranting and raving, he would storm out of the room. There was no attempt to understand the other side.

    norcal (6fde5d)

  13. @Patterico@13 If you have a burning desire to read it, Harvard has a PDF of the study posted here. It’s a bit of a slogging mess and they needed a better editor.

    Can you explain why you think a “significant portion of their data” would have been “collected before the girls would’ve graduated HS in about 1992

    Yes, sorry, I assumed you were looking at the study itself.

    Their data set is women born in 1973/74 who gave birth from 1989 to 2000, so it covers the period during the time the women would have been in HS and could have been in college and one of the metrics they used for economics was that private insurance was listed in the birth records of their child. There’s no indicator of whose private insurance she was using and if the girl/woman was in high school and/or college possibly up to age 23 (1996/1997) when she gave birth, then there would be a high likelihood that the insurance belongs to a parent, rather than the girl/woman herself, so it wouldn’t indicate her economic outcomes, but rather her parent’s.

    Since the data set covers only 11 years and up to eight of those years almost certainly includes data of girls/women who were still on their parents insurance, their data regarding economic success was compromised.

    This is the part of your comment that puzzles me the most.

    They were looking for specific sets of information that weren’t available during the entire 1961-2000 period and they also needed to match sets of birth records. Because of that it limited the data that they could use. A lot of the data they needed to match for the women’s birth wasn’t available before 1973 and the matching data on their child’s birth records wasn’t recorded until 1989. so the oldest women they could use were women born in 1973/1974 for children born in, at earliest, 1989.

    The 1989-2000 years included too many years when the women would have still been dependents of their parent/s, so it wouldn’t indicate their own economic factors, but only that of their parent/s. Theoretically you could use those same women, but use data from children they had after 1997 (to eliminate college or other births on parental insurance) and it would reduce or remove data skewed by being dependent on their parents and outside the work force due to young age or the education process.

    Do you think the famous 2004 study showed much?

    The study did a reasonable job of showing that in California there had been an increase in the use of black cultural names over time from 1961 to 2000, possibly influenced by a variety of cultural factors. 😛

    Nic (896fdf)

  14. Just guessing, but I’m guessing that for resume-level jobs, networking is probably the most important factor. Which gets us to talkin about “systemic” and I apologize for that. But elephants in rooms and all that. What connections do Emily and Gregg have as opposed to Lakisha and Jamal?

    nk (1d9030)

  15. nk (1d9030) — 6/7/2021 @ 5:25 am

    What connections do Emily and Gregg have as opposed to Lakisha and Jamal?

    The resumes are said to be identical, just submitted to different companies or in response to different ads. I think this test dealt with preliminary screening, and they couldn’t go much further anyway, since none of these people actually existed,

    Sammy Finkelman (b434ee)

  16. Why is this allowed?

    DRJ (03cb91)

  17. @19: I assume you are speaking of the comment and not the commenter. The comment was pretty much what I expected, given the source.

    Kevin M (ab1c11)

  18. I am less exasperated by the persiflage than I am about the attitude of some comrades, whether it’s about getting a job or getting the death penalty, that they’ll be the ones dishing it out instead of the ones suffering it, DRJ.

    nk (1d9030)

  19. Nic,

    Thanks for the link. I will look at it later.

    I should have been clearer: I was not referring to Fryer and Levitt as the famous 2004 study, but rather the one everyone cites for the purpose of “proving” employers are racist (Bertrand and Mullainathan). I’m willing to listen to critiques of Fryer and Levitt but my main thesis, backed up by other studies and common sense, is that Bertrand and Mullainathan did not convincingly prove what Big Media claims they proved. Do you have a view on that?

    Patterico (59e140)

  20. DRJ,

    I don’t really understand the Kevin M/DCSCA interchange and I have read it twice.

    People, be polite please.

    Patterico (59e140)

  21. What connections do Emily and Gregg have as opposed to Lakisha and Jamal?

    In the real world there may be some difference. In the study cited to “prove” racism nobody has any connections because the resumes are all made up.

    Patterico (59e140)

  22. I’m going to bet that if Lakisha and Emily both have degrees from Harvard, Lakisha gets the job and it’s still racism.

    Kevin M (ab1c11)

  23. His meaning is clear.

    DRJ (03cb91)

  24. DCSCA is why I rarely bother to read comments here anymore.

    DRJ (03cb91)

  25. “I don’t really understand the Kevin M/DCSCA interchange and I have read it twice.”

    Kevin: DCSCA is too obnoxious to be liked by attractive women so gay men are his best bet
    DCSCA: To have that understanding, Kevin must be well practiced at having sex with gay men

    I’m confident that Sammy will tease out the truth

    AJ_Liberty (ec7f74)

  26. I’m not sure I agree with your dismissal:

    Therefore, if you, as a potential employer, are more queasy about hiring a Lakisha than an Emily, the reason is obvious: you are a racist.

    The problem here is that the underlying assumption is false. In fact, Roland Fryer and Steven Levitt did a study in 2004, which I bet you never heard about, which found that distinctively black names “provide a strong signal of socioeconomic status” and tend to be characteristic of the subset of black folks “living in racially isolated neighborhoods.”

    In the scenario you describe, the potential employer has much more than just a “distinctively black name” by which to judge the applicant. They have an entire resume presumably documenting job qualifications which were considered sufficient for a person with a different name.

    I can’t see how it’s not racism to disqualify someone from a job on the basis of their name sounding “distinctively black”, regardless of whether it’s correlated with socio-economic status. It seems unconscionable to me, in fact.

    Dave (1542be)

  27. DCSCA is why I rarely bother to read comments here anymore.

    DRJ (03cb91) — 6/7/2021 @ 9:08 am

    He will occasionally have something insightful to say, but the signal to noise is terrible.

    Time123 (306531)

  28. …or address, in many ways that might do more harm to Lakeisha and Jamal. Perception of available transportation, likely elementary/secondary ed (though I’m told nk’s own Lane Tech was the only HS worthy of mention one’s fulltime career resume or vita)

    urbanleftbehind (1c01cf)

  29. Speaking as a white guy from a white suburb who has a (frequently mangled) Hispanic sounding name that usually signifies the bearer is African-American, I suspect there have been a few times my resume got trashed as a result. But I will always treasure my invitation to speak at the African-American accounting association here in ATL.

    Appalled (1a17de)

  30. ^Makes me think of this dude, who played pro football under a “blackish” name and achieved fame in TV and politics under a whiter name.

    urbanleftbehind (1c01cf)

  31. That just sounds better then saying blacks are inferior so racists can pretend they are not racist. The strong socio-economic signal is that they are lazy inferior blacks. Their is always a market for making racists feel better about themselves. Larry elder is the perfect example of this.

    asset (9dd89c)

  32. Kevin: DCSCA is too obnoxious to be liked by attractive women so gay men are his best bet
    DCSCA: To have that understanding, Kevin must be well practiced at having sex with gay men

    I’m not sure this is the correct translation but it might be. The whole thing is stupid. Maybe I’ll just delete all of those comments.

    Patterico (e349ce)

  33. In the scenario you describe, the potential employer has much more than just a “distinctively black name” by which to judge the applicant. They have an entire resume presumably documenting job qualifications which were considered sufficient for a person with a different name.

    I can’t see how it’s not racism to disqualify someone from a job on the basis of their name sounding “distinctively black”, regardless of whether it’s correlated with socio-economic status. It seems unconscionable to me, in fact.

    You have a habit of reaching opinions about articles you haven’t read. This stuff is addressed in the piece as a whole. If you don’t want to subscribe and read it, that’s fine, but it’s kind of a waste of time to level critiques that are addressed by the broader piece — and I’m not interested in regurgitating it all here for you, because I already spent a bunch of time writing it.

    Patterico (e349ce)

  34. I’m going to bet that if Lakisha and Emily both have degrees from Harvard, Lakisha gets the job and it’s still racism.

    How so? If everything else is equal except skin color, would picking Lakisha be considered racist? Unless you’re one of those who believes you can’t hire a minority until all Caucasians are fully employed.

    Rip Murdock (2975ef)

  35. Nic,

    I’ve skimmed through some parts of the study. You seem to fixate on one variable — insurance — but the study itself appears to me to cover a wide range of variables, including years of education of the woman or the father of her child, mother’s age at first birth, the baby’s birth weight, the number of total children born to date, per capita income in the ZIP code, and so forth.

    They concede that their “outcome measures are coarse” as “[w]e do not observe individual wages or family economic circumstances, but rather, the median income of their zip code, years of education, etc, which are highly correlated with the relevant outcomes.” Again: this study is only one building block in my argument — albeit the one I highlighted in my teaser — but there appears on the whole to be a lot of data suggesting that the famous study highlighted by Big Media confounds race and socio-economic status, and there are studies that attempt to disentangle these factors that come to different results.

    Patterico (e349ce)

  36. Patterico, When Substack came out I told myself I wasn’t going to start paying. This subject is interesting and I’m so convinced by the previous data I’ve seen I want to understand where the disconnect it.

    Time123 (306531)

  37. Cross posting because of names and guesswork with regard to suspects’ ethnic/racial background:

    https://heavy.com/news/wynne-lee-marcus-eriz/

    urbanleftbehind (23a4bd)

  38. You have a habit of reaching opinions about articles you haven’t read. This stuff is addressed in the piece as a whole. If you don’t want to subscribe and read it, that’s fine, but it’s kind of a waste of time to level critiques that are addressed by the broader piece — and I’m not interested in regurgitating it all here for you, because I already spent a bunch of time writing it.

    Well, fine, but you posted what I quoted and responded to here, on this site…

    Dave (1542be)

  39. I’m not sure this is the correct translation but it might be. The whole thing is stupid. Maybe I’ll just delete all of those comments.

    Not quite, at least on my part. But I see no reason to raise this up any further. I have no problem with what DCSCA wrote.

    Kevin M (ab1c11)

  40. The topic is very interesting but I get sidetracked by the insults. I don’t object to insults per se but these are
    gratuitously mean and gradeschool caliber. At least be entertaining and clever about it.

    DRJ (03cb91)

  41. The people of Boone County, W. VA would be crazy to hire these White people:
    “The Wild and Wonderful Whites of West Virginia”.

    I once ran into some folks from the area where my mom was raised and in coversation we quickly narrowed it down them being from the same 250 person town area as my Mom. They asked my moms last name so I told them…. pregnant pause….. “oh….” so I jumped in with my Grandma’s maiden name. “Oh yes, Wonderful people, your grandmother married a…?
    So I said something like: “well, back then they might of just been the last two single people standing at the church picnic” and they said “well, it couldn’t have been at church”.
    Rest assured that if I sent them a resume with my grandpa’s name on it, I wasn’t getting the job.
    My opinion is we all bring biases to choosing who we want to work for us, its more of a case of being honest with ourselves about it. My wife has a bias against people from the state of Guerrero MX. Most Mexican people I know have opinions about the Mixtecos and seeing a Mixteco name on a job app that isn’t for manual labor would raise some eyebrows.
    I’ve heard of Nigerian-Americans not hiring US Africa Americans, US African Americans not hiring Haitians or Jamaicans.
    The white people I am around look at credentials and experience, but who knows if they were put off by a Mom’s misplaced accent in a first name, or an overly grandiose name?
    If you are hiring a data entry person, would you possibly be subliminally put off by a first name that looks misspelled and has an accent where none should be?
    I might be.

    steveg (ebe7c1)

  42. Well … it’s a complicated world.

    The Presiding Judge of the Chancery Division in Cook County also signs name changes and, in the previous century, I would see some of them while waiting for my trial courtroom assignment. He would ask the petitioners why they wanted the name change. To make sure it was not for fraudulent reasons.

    Many were black people and they wanted to correct the spelling of their names on their birth certificates. All they had to say was “Mississippi”. He got it. Birth registrars in Mississippi would deliberately misspell the names of black babies on the birth records. So when they tried to register to vote 21/18 years later, the name they had been known by all their life could not be found.

    “Sorry. We have a Lakisha Jonestown, but no Letitia Johnston.” Like that.

    nk (1d9030)

  43. Most Mexican people I know have opinions about the Mixtecos and seeing a Mixteco name on a job app that isn’t for manual labor would raise some eyebrows.

    Anecdotally, I have been made aware that hiring decisions of Indian (dot, not feather) managers favor or privilege other Indians of their own locale and caste, then US whites, then Indians of other regions and castes, then other ethnicities of US citizens. Arguably the sample of Indians in position to make the decisions described is probably small. But it would be interesting to review that data and see if US white managers are more, or less, biased than our newly arrived managers-of-color.

    Actually, that seems to me to be a fair question all the way around. Not “are whites in the U.S. biased?” but “How much more, or less, biased are whites in the U.S. in positions of power than people of other ethnicities in comparably positions of power?” Do whites exploit their power to sustain an existing system based on “whiteness”? Or do non-whites bias their power in the white system to advantage one, their own, non-white group at the expense of other non-white groups? Or, as CRT seems to advocate, do non-whites in the white system affirmatively address existing inequities to re-balance an excess of whiteness with a new draw and mix of many non-whites?

    pouncer (6c33cf)

  44. @Patterico@35 When they were looking at economic outcomes, they used 2 factors, that of listing private insurance and that of average median income in the zipcode. The zipcode has similar issues to private insurance because if they are still in school then the zipcode they are in is probably their parent’s zipcode, rather than one they have chosen to live in themselves (using zipcodes also brings up other issues regarding mobility, like if you live in a tight community or one where your family has roots, you may not move significantly far from it, even if your economic factors improve, so you may just be in a better neighborhood in the same zipcode.)

    On the Bertrand study, which you can find here. Their sample size of 5000 resumes sent out is good. Their methodology of randomization for which names go on which resumes is good (controls for the researcher’s unconscious biases) . The ability to generalize their results may be limited. With Boston especially there is a niche culture that may not be generalizable to the whole country. Chicago results could be more generalizable, but there really can be a lot of regional differences in the US, despite us all sharing the same over-culture, that I would be reluctant to do so. They are primarily looking at very specific and easily observable piece of data, which is also good methodology since there isn’t much ambiguity or valuation bias in whether a company called in response or not, it either did or it didn’t. 😛

    They also looked at the relationship between socio-economics and ethnic naming trends to see if there might be some class based intersection in results and I thought it was interesting that what they found back east was that there wasn’t necessarily a similar pattern in naming practices to the one that Fryer and Levitt found in California.

    They controlled very well for race, less well for gender (sometimes they deliberately over-selected one gender over the other for certain jobs, so you can’t compare call-back rates for men and women because they are skewed)

    They are thorough in discussing other factors that could have contributed to call-back rates other than race and why those may or may not apply.

    Interestingly they sited Fryer and Levitt in their study. 😛

    So, in conclusion, the sample set was good, the data was good, the methodology was good for everything except gender (which was incidental to what they were studying). They controlled well for economic perceptions by randomizing zipcodes. The study was also well written. I don’t always agree with their incidental conclusions (in some cases I feel like they are overemphasizing their look into correlations in certain areas) but their primary conclusion appears, in general, to be sound based on their data. Generally speaking it’s a good study. Something racial appeared to be effecting call backs for people with African American names in Boston and Chicago. It is not entirely clear what it was, but there are strong indications that there is something racial there.

    Nic (896fdf)

  45. RIP Clarence Williams III (81).

    Rip Murdock (d2a2a8)

  46. Nic,

    I am appreciating the discussion, and it’s clear that you have expertise in this area — I’m guessing much more formal expertise than I have (though I don’t know what you do for a living). It’s a little frustrating to me that you have not read my whole piece; if you want to email me I could send it to you as I think the discussion would be fruitful for me. Simonsohn’s observations about the Bertrand study in particular seem on target to me.

    Also, thanks for the links to the studies themselves. For whatever reason, I had trouble finding the entire studies online — meaning I had to rely on third-party discussions of methodology at times — so I find these links helpful.

    All of that said:

    I find the Bertrand study’s dismissal of the class confound utterly unconvincing. They concede that it is “plausible” that “employers do not care at all about race but are discriminating only against the social background conveyed by the names we have chosen” but then go on to argue against this interpretation with arguments that, for me, land with a thud.

    Their first argument (at p. 1007) seems to assume that employers are paying particular attention to addresses. They say that “If the African-American names we have chosen mainly signal negative social background, one might have expected the estimated name gap to be lower for better addresses.” To me, the signal sent by a name distinctively signalling low socio-economic status is not remotely made up for by “better addresses” — which I suspect employers pay relatively little attention to.

    Similarly — and I am skipping around a bit because I am short on time right now — we are told “African-American babies named Kenya or Jamal are affiliated with much higher mothers’ education than African-American babies named Latonya or Leroy. Conversely, White babies named Carrie or Neil have lower social background than those named Emily or Geoffrey.” OK . . . I’m not sure that’s intuitively obvious to most people. If we’re going to slice things that finely, we have to be consistent about how we react to differential findings based on such non-intuitive differences — and when we are consistent, we can be surprised by the results, depending on what we choose to emphasize. For example, when we go to the table to check all this, one thing that jumped out at me is that “Ebony” has a higher callback rate than “Emily.” Huh? Latonya and Kenya are higher than Jill or Anne. Yes, the “black” names have an overall lower callback rate but also have an overall lower education for the mother.

    I think Simonsohn’s general observations that this study does not address these confounds well has merit. Maybe I can discuss more at a time when I have more time available.

    Patterico (e349ce)

  47. @Patterico@46 I’ve had a year of undergrad research design and a year of graduate research design, and I did a small sociological study in undergrad (on average women use more words when describing a picture than men do) and a full publication level research based thesis in grad school. My job involves a lot of organizing and/or providing interventions to support at risk students, so I also spend a fair amount of time reading research on drop out rates, graduation rates, whether or not state testing scores really mean anything, public/private partnerships in schools, and other education related stuff mostly relating to letter salad programs meant to provide interventions and/or “best first teaching”. Don’t get rid of your career and technical ed programs, people.

    I should start off by saying that I have a strong disinclination to believe other people’s assessment of studies and very much prefer to read the studies themselves and reach my own conclusions. (mostly because other people’s assessments are clearly wrong. OK, not really, mostly due to pop-psych hot-takes that simplify results into what they want them to say and the tendency of the internet to sensationalize things.)

    My concern about leaning too much on “it’s just class” in the Bertrand study is that if employers are using that as a factor, it should also be seen within the groups, not just between the groups. Employers don’t seem to be doing that within the group of the “white” names on the list (Carrie gets way more call-backs than Emily, even though Emily “should” be perceived as a higher status name) or within the group of the “black” names on the list (Aisha gets far fewer call-backs than Latoya, even though Aisha “should” be considered a higher status name). It’s only seen between the groups. If the perception is simply that people with black ethnic names are lower status than people without black ethnic names, that is a racial perception even if it isn’t necessarily racism. A socio-economic comparison regarding perception of names could, however, definitely be better done. If you are interested, someone has done some work in that area, though take it with a grain of salt since it’s not officially published and I’m not sure about their data collection method (It could very well be fine, I just don’t know enough about the website they used). The table you would want is at the very bottom. I just thought it was a fun read and, since you sent me down that rabbithole to see if anyone had actually done that work, I thought I’d return the favor and bring the rabbithole to you as well.

    Simonsohn’s concerns are interesting, though I have similar concerns as to his data collection method as I do with the above name study I linked (they used the same website). The study he was referencing is here (listen, I realize that all the study linking is very geeky.) The methodology is very different than Bertrand, because Bertrand sent all 4 different resumes to each job, so you were theoretically getting the same set of biases for each resume. The study Simonsohn was looking at sent different ethnicities and genders to different employers so, like, they send 4 white male resumes to job 1 and 4 black female resumes to job 2, etc. They were attempting to control for racial bias, rather than look at it. (they were successful) but because the methodology and focus were different it would need to be further researched to be able to reach Simohn’s conclusion that it was because of name choice. My other concern about Simonsohn is that he doesn’t want the researchers to use frequency data to pick the names they used, but to instead hand-pick the names, and that could bias the study because then the researcher could be using his own personal perceptions which could skew name choice. You’d have to be really careful on how you did it.

    FWIW, my guess is that the perception of which names indicate what kind of socio-economic status is probably pretty regional. I did not grow up in CA and my association of black ethnic names (at the time) was that their parents were probably pretty involved with black cultural and/or political movements back in the day. Also my mom’s family had some whole thing about which Irish names were OK, the only part of which I remember was that you didn’t name your daughter Moira because she, er, would say yes too much.

    The Darolia study (which I can’t, very annoyingly, find in full text, only the intro) intro here appears interesting. I very much like that it is more nationally based. My concern is that they did not use black ethnic first names due to concerns about socio-economic bias (they used Chloe and Ryan as first names), but rather last names (Washington and Jefferson) common among black people, which they agree is a concern

    A tradeoff is that the surnames in our experiment may not indicate racial background to employers as strongly as the distinctly African-American sounding names in their study.

    I would not personally see the name “Chloe Washington” and assume she was black. I suspect the data on latino applicants would be more solid, but I can’t see it. Bah.

    Anyway, those are my thoughts.

    Nic (896fdf)

  48. Nic, and Patterico, This is a great discussion. Thank you for having it publicly.

    One thing that the Bertrand Study appears to have done is provided other indicators of SES.

    Billy Van Frey might sounds lower class then William Frey, but if he went to an elite school on an polo scholarship other conclusions might be in order.

    Time123 (cd2ff4)

  49. “Van _____” might mean old money in New York State

    urbanleftbehind (812b8e)

  50. If the perception is simply that people with black ethnic names are lower status than people without black ethnic names, that is a racial perception even if it isn’t necessarily racism.

    This is the confounding part — pun entirely intended — and I think it’s entirely due to the Bertrand study’s (in my view) questionable methodology.

    Again: imagine a study where recognizably redneck names like Cletus and Billy Jack represent the white applicants, and are coupled with (to take two names of black colleagues of mine) Todd and John to represent the black. If employers formed a perception that people with names that tend to be associated with low class whites are indeed lower status than people without such names, is that a “racial perception”? I say it’s more of a class perception.

    There is a basket of signals that can send messages about class. In addition to names, there is profanity, quickness to anger, tendency to speak loudly, certain accents or styles of speech, and so forth. No one thing is reliable. I have worked with people with distinctively black names who are very well off and educated. The loudest motherfucker I know is white and his childhood home is around the corner in my affluent neighborhood. No one signal is reliable but taken together they can send a signal, and an employer might favor a quiet well-spoken non-profane even tempered person with a non-distinctive name of either race (white or black, specifically, here) over a loud profane angry-sounding person with a distinctive accent and name associated with lower classes of either race.

    In the end my thesis is not “it’s totally class and race is shown to have nothing to do with it” but rather “it’s a lot more complicated than you have been told, and class may well predominate over race in the observed effects but Big Media will never say so.”

    I have many reactions to your comment, Nic — more than I can pack into a single comment or than I have time for this morning — and will likely write a follow-up to the next paid subscriber post with them. For now, I just wanted to respond to this one point, but it’s not the entirety of my thoughts. In any event your Input is very, very valuable to me and I appreciate your taking the time.

    Patterico (e349ce)

  51. Another thing: it would hurt me very much if someone I knew with a distinctively black name were to interpret my argument as calling them low class. That is not what I am saying any more than my dad Billy Van was low class because of his name. But imagine how easy it would be for Twitter (for example) to take my argument and reduce it to that. This is part of what makes honest discourse between well-meaning people very difficult these days. And it could happen any time, like a bolt of lightning. Your actual words or meaning don’t matter. Look at what happened to Ellie Kemper. They made her apologize.

    Patterico (e349ce)

  52. @51 Patterico, your concern is well founded and I appreciate your willingness to discuss this subject in public.

    imagine a study where recognizably redneck names like Cletus and Billy Jack represent the white applicants, and are coupled with (to take two names of black colleagues of mine) Todd and John to represent the black. If employers formed a perception that people with names that tend to be associated with low class whites are indeed lower status than people without such names, is that a “racial perception”? I say it’s more of a class perception.

    One thing I’ve been looking for is data around frequency of low SES white name names. Your point about white names makes sense, but I’m not sure how frequently it occurs.

    If most low SES whites have names that aren’t easily differentiated from high SES names I think that means that there’s no real way to see if the behavior in call backs is the same for white’s as for blacks.

    Time123 (53ef45)

  53. I have to admit I don’t generally hold sociological studies in high regard. In my experience they generally detract from, rather than add to, the conversation — and constantly provide an exercise in answering the question “what variable did they fail to control for in order to reach this result that Big Media and the professor’s colleagues will love”?

    Patterico (e349ce)

  54. I’ve been working with marketing and customer data for last couple years after working in more physics based areas. From that perspective I can tell you that even when there’s no political motive involved the question about what variables to use, how much money to spend to get data on them, and how to weight them can be vigorously debated.

    Time123 (53ef45)

  55. I have to admit I don’t generally hold sociological studies in high regard. In my experience they generally detract from, rather than add to, the conversation — and constantly provide an exercise in answering the question “what variable did they fail to control for in order to reach this result that Big Media and the professor’s colleagues will love”?

    Patterico (e349ce) — 6/8/2021 @ 8:38 am

    You realize that in this scenario you’re playing the role of “what variable….”. The first pass was that they sent out a bunch of resumes with white names and black names. The white names got a lot more call backs.

    If I had a black name I’m not sure how much I’d care if I wasn’t getting calls because they thought I was black, or because they thought I grew up poor.

    Time123 (53ef45)

  56. @Patterico@50

    Again: imagine a study where recognizably redneck names like Cletus and Billy Jack represent the white applicants, and are coupled with (to take two names of black colleagues of mine) Todd and John to represent the black.

    I know this is not the point, but, If I were doing that study, I’d pair Cletus and Billy Jack with a variety of cultural black names. Todd and John wouldn’t be good representations for black names because they aren’t culturally black names, so any conclusions regarding the data and race would be bad. You’d have to control for regionalism, too, because they are very distinctly southern names.

    an employer might favor a quiet well-spoken non-profane even tempered person with a non-distinctive name of either race (white or black, specifically, here) over a loud profane angry-sounding person with a distinctive accent and name associated with lower classes of either race.

    That is certainly true, it isn’t going to appear on a resume audit though. The person needs to get a call-back first. 😛

    “it’s a lot more complicated than you have been told,

    It’s ALWAYS more complicated than we’ve been told in the soft-sciences. People want answers not probabilities and/or caveats.

    I have many reactions to your comment, Nic — more than I can pack into a single comment

    No worries, lawyering is busy work.

    In any event your Input is very, very valuable to me and I appreciate your taking the time.

    I’m glad you find it to be so. I’m enjoying the discussion.

    @time@48 You are welcome. It’s an interesting discussion to have.

    They did use a variety of zipcodes that could be socio-economic indicators if one knew the area well enough.

    Nic (896fdf)

  57. #49
    My Dutch and German ancestors were Frisian and their surnames did not translate into wealth.
    Mudhutdwellerwithgiantshoesize

    steveg (ebe7c1)

  58. This is as good a time to correct an injustice as any. “Trump” has been Donald Trump’s family name since at least 1699. It was Drumpft before that. His grandfather was Trump when he came to America. He did not Americanize it from Drumpft to Trump as has been alleged and the allegation believed by me.

    nk (1d9030)

  59. I know this is not the point, but, If I were doing that study, I’d pair Cletus and Billy Jack with a variety of cultural black names. Todd and John wouldn’t be good representations for black names because they aren’t culturally black names, so any conclusions regarding the data and race would be bad.

    Now reverse the colors in that comment and you made (part of) my point for me!

    Are Greg and Emily culturally white names? I have a black colleague named Greg. I have no white colleague named Lakisha. This is why Time123 misses the point when he says “ The first pass was that they sent out a bunch of resumes with white names and black names.” No, the “white names” were not necessarily white. They were just not distinctively black, and what’s more, the distinctively black names tended to be associated with lower socio-economic classes.

    I feel like nobody really grappled with this thing that I said so I will say it again;

    Again: imagine a study where recognizably redneck names like Cletus and Billy Jack represent the white applicants, and are coupled with (to take two names of black colleagues of mine) Todd and John to represent the black. If employers formed a perception that people with names that tend to be associated with low class whites are indeed lower status than people without such names, is that a “racial perception”? I say it’s more of a class perception.

    Nic responded to it but does not seem to understand how the response actually proves my point.

    Patterico (e349ce)

  60. I have an odd first name although not of African descent. I have noticed a big difference in callbacks based on whether I use my legal or my professional name. I often advise others to do something similar to get in the door and then introduce their preferred name later. This applies to people from other countries or from here. My friend TaKayla shortened her name to Kayla, and had better results. Now, as a hiring manager I often had to subset a group of 50 resumes to 5 for screening. (And this would be after HR did their magic.) You may only have time for 30-60 seconds per resume, so you use shortcuts after the clear rejections. I like this college because it has hard workers; I dislike this school because its graduates tend to be assholes. One of my theories for the importance of generic names is that we prefer to be able to correctly address our colleagues. If you have trouble deciphering pronunciation on the resume, will you have similar trouble in the interview, and more importantly will your clients? Could this end up being a friction point?

    HR departments are increasingly using computerized screening. This leads to people putting things on their resumes which could indicate puffery. The next people to do this study will need to take machine screening into account.

    From West Texas too (f74c85)

  61. @Patterico@59 The names they used were white cultural names (in the locales they were studying, at the time).

    The next step is to generate identities for the fictitious job applicants: names, telephone numbers, postal addresses, and (possibly) e-mail addresses. The choice of names is crucial to our experiment. To decide on which names are uniquely African-American and which are uniquely White, we use name frequency data calculated from birth certificates of all babies born in Massachusetts between 1974 and 1979. We tabulate these data by race to determine which names are distinctively White and which are distinctively African-American. Distinctive names are those that have the highest ratio of frequency in one racial group to frequency in the other racial group.As a check of distinctiveness, we conducted a survey in various public areas in Chicago. Each respondent was asked to assess features of a person with a particular name, one of which is race. For each name, 30 respondents were asked to identify the name as either“White,”“African-American,”“Other,”or“Cannot Tell.”In general, the names led respondents to readily attribute the expected race for the person but there were a few exceptions and these names were disregarded

    They used both name frequency data and interviews to determine which names were and were seen as racially specific in the areas they were studying, at the time of the study. Your argument is that the most commonly used black cultural name are perceived as low class while the commonly used white cultural name are perceived as less low class so the researchers would need to hand-pick uncommonly used white cultural names to compare. This argument, in and of itself, makes an assumption of racial bias in society.

    Nic (896fdf)

  62. 46.

    one thing that jumped out at me is that “Ebony” has a higher callback rate than “Emily.” Huh? Latonya and Kenya are higher than Jill or Anne.

    Perception of age, with Jill and Anne (and maybe even Emily?) perceived as older? Or “overqualified?”

    Sammy Finkelman (51cd0c)

  63. Nic, It would be interesting to see if there were ‘black’ names that were assumed to be higher SES and if there were white names that were assumed to be lower SES. I’d also like to see how common those sets were. “Cletus” or “Lakisha” might both code as low SES but if there are 100 Lakisha’s for ever “cletus” It makes it harder to sperate in the real world.

    Basically, if there aren’t common black+high SES or white+low SES in the real world the distinction doesn’t seem to matter.

    Additionally, poverty rates are 2.5 times higher for blacks then for whites. Disseminating against poor black people is still undesirable.

    I’ve been trying to find data to flesh this out and haven’t been very successful.

    Time123 (6e0727)

  64. 61. Nic (896fdf) — 6/9/2021 @ 9:47 am

    This argument, in and of itself, makes an assumption of racial bias in society.

    The question is, is this more independent racial bias, or derivative of something else?

    The New York Times had a front page story yesterday about statistical differences between people belonging to different racial groups in FEMA disaster aid, and stated they can’t figure out why. Actually, people closer to the ground could probably figure out why, or what are the intervening variables.

    https://www.nytimes.com/2021/06/07/climate/FEMA-race-climate.html

    A growing body of research shows that FEMA, the government agency responsible for helping Americans recover from disasters, often helps white disaster victims more than people of color, even when the amount of damage is the same. Not only do individual white Americans often receive more aid from FEMA; so do the communities in which they live, according to several recent studies based on federal data.

    Leaders at FEMA are wrestling with the complicated question of why these disparities exist — and what to do about them. The problem seems to stem from complex systemic factors, like a real estate market that often places higher values on properties in communities with many white residents, or the difficulty of navigating the federal bureaucracy, which tends to favor people and communities that have more resources from the beginning.

    There already you have part of your answer. Compensation is based on property values, not cost of replacement, and there are differences in how plugged in people, and even communities, are.

    Sammy Finkelman (51cd0c)

  65. @time@63 This is the names and SES perception I linked Patterico to earlier.

    @Sammy@64 The only difference in applicants is the name itself. They randomized the names over the resumes between jobs, so one resume that had Malik on it for one job would have Emily on it for another, etc, so there’s no other place on the resume itself that would cause a derived bias. Any derived bias would have needed to be internal from the person reading the resume and calling/not calling.

    Nic (896fdf)

  66. But the bias is entirely based on name, not anything else. That leaves open why some names were more likely or less likely to get callbacks. If you see “black” names get fewer on average that could be because such names are perceived by some people as indicating a whole variety of things.

    Why should, in this day and age, probable race of person with that name, be the thing resume screeners were most sensitive to?

    Sammy Finkelman (51cd0c)

  67. @sammy@66 Well, as I said, there seems to be something racial going on, but the study isn’t designed to really tell us more than that.

    Nic (896fdf)


Powered by WordPress.

Page loaded in: 0.1709 secs.