Anyone Who Plays Squash

I · The Airport Test

In 2006 a doctoral student in sociology took an unpaid job in the recruiting department of an elite American firm and stayed for nine months. Lauren Rivera had negotiated an unusual arrangement: she would do the administrative work of hiring season—the scheduling, the paperwork, the shepherding of candidates between rooms—in exchange for permission to watch. She was not allowed into the interviews themselves. She was allowed everywhere else, which turned out to be where the decisions were made. Alongside the observation she conducted a hundred and twenty interviews with the people who do the choosing at investment banks, law firms, and management consultancies: forty in each industry, hiring partners and managing directors and the human-resources managers who assemble the lists.

What she found was not that elite firms hire snobbishly. That would be a smaller and duller finding. What she found was that the people doing the hiring, asked to rank their own criteria, put cultural fit at the top—more than half of them named it the single most important consideration at the interview stage, ahead of analytical ability and ahead of communication skill. Once a candidate had cleared the first screen, which is usually the prestige of the university, fit was typically given more weight than grades, than coursework, than work experience. Concerns about shared culture, as Rivera put it in the American Sociological Review in 2012, often outweighed concerns about absolute productivity.

The interviews are better than the summary. One banker described his method as the stranded-in-the-airport test: “Would I want to be stuck in an airport in Minneapolis in a snowstorm with them?” Another was blunter about what he was buying. “A lot of this job is attitude, not aptitude,” he told her. “You will see more of your co-workers than your wife, your kids, your friends… So you can be the smartest guy ever, but I don’t care. I need to be comfortable working everyday with you, then getting stuck in an airport with you, and then going for a beer after. You need chemistry.” And a third, scanning a résumé in front of her, resolved the question in six words: “She plays squash. Anyone who plays squash I love.” He ranked the candidate first.

It is tempting to read this as a story about a sport, and it is not. It is a story about a signal, and the most important thing about the signal is that it does not mean the same thing in every room. A lawyer at a scrappier firm read an almost identical line—lacrosse, squash, crew—and laughed. “I’m sort of giving him a personality type here, and I don’t think he’s going to fit in well here… we’re more rough and tumble. I’m going to let him go.” The same three words, the same three sports, the opposite verdict.

This essay is about what follows from that, and it travels in an order that gets steadily less comfortable. That elite institutions sort on class-coded signals is now about as well established as anything in the sociology of work, and the audit evidence for it is strong enough to be worth setting out precisely. That the signals are therefore a code one can learn—the proposition on which a whole shelf of advice literature and a certain amount of hope depends—is much weaker than it looks, and fails in a specific and interesting way. And that mastering the code closes the gap is the part that does not survive at all: the people who have demonstrably succeeded at it are still paid measurably less, decades later, than the colleagues they learned to sound like.

II · Cabot and Clark

Interviews tell you what people say they do. To find out what they do, you have to send them something. In 2016 Rivera and the sociologist András Tilcsik published, in the same journal, an audit study of a kind familiar from the research on racial and gender discrimination but rarely turned on social class. They built four fictitious law-school candidates, sent their applications to three hundred and sixteen law-firm offices across fourteen American cities—a hundred and forty-seven firms, some sixty per cent of the entire universe of such offices—and counted the invitations.

The candidates were, on paper, indistinguishable in ability. All four sat in the top one per cent of their class at a selective but second-tier law school, the kind ranked somewhere between fiftieth and hundredth of two hundred. What varied was a constellation of small class markers, none of them a claim about competence. One candidate was surnamed Cabot, another Clark. One had won a “university athletic award”; the other, an “award for outstanding athletes on financial aid.” One had mentored first-year students, the other first-generation students. One sailed; the other ran track. One listed interests as sailing, polo, and classical music; the other as track and field, pick-up soccer, and country music. The signals are almost comically legible once assembled, which is the point—they are legible to the reader because they were legible to the reader in the firm.

The overall callback rate was 6.96 per cent, and it was not distributed as anyone’s theory would have predicted.

Higher-class men were invited to interview at 16.25 per cent—thirteen of eighty. Everyone else, taken together, was invited at 3.83 per cent: more than a fourfold difference, significant beyond the one-in-a-thousand level. So far, so much as expected. The unexpected part is inside the remainder. Higher-class women were invited at 3.80 per cent, three of seventy-nine—which is not merely worse than higher-class men but worse than the lower-class women, who were invited at 6.33 per cent. For women, the polish that helped the men was, if anything, a liability.

The authors went looking for the mechanism and found it in the language evaluators used. In a survey experiment with two hundred and ten American lawyers, and in twenty follow-up interviews, higher-class signals raised perceived fit for candidates of both sexes. But for women only, those same signals lowered perceived commitment—the affluent-sounding woman was read as someone who might not need the job, or stay in it. Respondents, in the authors’ summary, praised both higher-class men and women for their fit while questioning the higher-class women’s commitment; conversely they praised the work ethic of the lower-class applicants while doubting their fit, and some suggested those candidates might be happier in government or the public sector.

There is one further detail in Rivera’s interview study that deserves to be dragged into the light, because it demolishes the most common defence of all this. If cultural fit were really a proxy for the interpersonal demands of the work—if firms screened for chemistry because the job runs on chemistry—then fit should matter most where people work most closely together. It does the opposite. Fit was ranked first by seventy-five per cent of the law-firm evaluators and by only forty per cent of the consultants, and consulting is by some distance the most client-facing and team-dependent of the three. As Rivera notes, the use of cultural fit is not an artifact of a job’s social demands. One consultant said the quiet part outright: “I don’t think [fit] should be [a consideration] at all… But it’s what [firms] want, so it’s what you do.”

III · The Fish and the Water

The theory that anticipated all this is more than fifty years old and is routinely oversold, including by people who have read it. Pierre Bourdieu, the French sociologist who spent the nineteen-sixties studying Kabyle peasants in Algeria and the rest of his life studying the French, gave the discipline three words it has not been able to do without. Habitus is the set of durable dispositions—taste, posture, expectation, the sense of what is done—laid down by early experience and operating mostly below deliberation. Bodily hexis is its physical residue: in his formulation from the Outline of a Theory of Practice, political mythology em-bodied, turned into a permanent disposition, a durable manner of standing, speaking, and thereby of feeling and thinking. And cultural capital is what those dispositions are worth when exchanged.

The image everyone remembers is the one he gave with Loïc Wacquant in 1992: when a habitus encounters the social world of which it is the product, it is like a fish in water—it does not feel the weight of the water, and it takes the world about itself for granted. That is a better description of the squash conversation than anything a management theorist has produced since. The banker was not applying a criterion. He was recognising a fellow-swimmer.

What gets lost in transmission is Bourdieu’s carefulness. Two qualifications in particular are dropped almost every time the theory is invoked, and both matter for the argument that follows. The first is that habitus is durable and transposable, not fixed; he resisted the word “unconscious” as too psychoanalytic and preferred to speak of what operates before reflection. A theory of disposition becomes a theory of destiny only when that qualification is quietly removed, which is how a great deal of popular writing on the subject manages to be simultaneously fatalistic and inspiring.

The second is subtler and cuts against a claim one now hears constantly—that cultural capital is worth whatever the local market says it is worth, and nothing more. Bourdieu did hold that a capital’s efficacy is specific to the field in which it operates. He also held, in the same 1986 essay, that economic capital is at the root of all the other types; that a credential confers on its holder a conventional, constant, legally guaranteed value with a relative autonomy vis-à-vis its bearer, in explicit contrast to the autodidact’s capital, which may be called into question at any time; and, writing on the state in 1994, that the state holds a kind of meta-capital granting power over the rates of conversion between the others. It is the exchange rate that is local. The value is not. A New York reader will recognise the distinction immediately: a diploma travels, a manner does not, and the person who has only the manner is holding the more fragile asset.

IV · The Chameleon’s Wages

If the codes are real and legible, the obvious inference is that one should learn them. This is the promise of cultural code-switching, and it is made with great confidence in a large advisory literature: read the room, adjust the register, and the door opens. As a description of something people do, it is unimpeachable—anyone who has moved between classes or countries has done it, often before breakfast. As a claim about outcomes it is remarkably poorly evidenced, and the poverty is worth documenting, because the confident version is everywhere and the evidence is nowhere near it.

Start with the strongest adjacent measure, which is the personality trait psychologists call self-monitoring: the disposition to observe and adjust one’s self-presentation to the situation. The founding career study is a good one. Martin Kilduff and David Day followed a single M.B.A. cohort for five years after graduation, measuring self-monitoring before anyone had a job and taking outcomes from alumni-office records rather than self-report. High self-monitors changed employers more—a correlation of .32, the strongest thing in the paper—and won more cross-company promotions. But promotion within a company was not significant and was signed the wrong way, at −.08. The authors put the trait’s explanatory reach at between four and eleven per cent of the variance in career mobility. Chameleons, on this evidence, move. They are not shown to rise.

The meta-analytic picture is smaller still. Pooling a hundred and thirty-six samples and some twenty-three thousand people, David Day and colleagues found self-monitoring correlated with job performance at .10—which is precisely Cohen’s threshold for a small effect, about one per cent of the variance—and with leadership at .21. The decisive detail is a moderator. Against subjective performance ratings the correlation was .15; against objective measures of output it was .03. Whatever the trait buys, it is bought from human raters.

And then a meta-analysis published in 2025 asked which humans. Pooling fifty-five samples and just over nine thousand people, Lele Lei and colleagues broke the self-monitoring–leadership relationship down by who was doing the rating, and the result is the most quietly devastating number in this essay.

Self-rated leadership effectiveness correlated with self-monitoring at .223, comfortably significant. Supervisor-rated effectiveness, at .071, was not significant. Subordinate-rated effectiveness, at .034, was not significant either. The people who work for the chameleon do not see what the chameleon sees. The authors’ own advice to practitioners is that they should be cautious not to be easily impressed by individuals skilled at cultivating favourable social images, which is a sentence one does not often find at the end of a meta-analysis.

It should be said that the counter-literature is no better armed. The best-known statement of code-switching’s psychological cost is an article in the Harvard Business Review, which is a practitioner magazine and not a peer-reviewed venue, and whose three cost claims are borrowed from adjacent research rather than measured in its authors’ own studies. Their peer-reviewed paper, published two years later, found that both Black and white participants rated a code-switching professional as more professional—the opposite of what the popular title suggests. The mechanism usually invoked for the cost, ego depletion, failed a preregistered replication across twenty-three laboratories and two thousand one hundred and forty-one participants, returning an effect of 0.04 with a confidence interval straddling zero. Both the promise and the warning, in other words, are running well ahead of their evidence.

V · What Cannot Be Faked

The deeper problem with learning the code is visible in Rivera’s own data, in a place most summaries skip: a footnote. Similarity, she observes there, is risky to fake. People react badly to inauthenticity in self-presentation, and—this is the operative clause—evaluators reported spot-checking candidates’ stated experiences to see whether participation had been genuine and extensive. The squash line works because it can be pulled. Someone will ask where you play.

Her conclusion generalises the point and quietly dismantles the whole coaching proposition. Mere familiarity with a cultural signal, she writes, was insufficient. Successful candidates needed enough cultural breadth to find common ground with whichever professional they were paired with, and enough depth in white, upper-middle-class cultural signals to relate to interviewers who were overwhelmingly white, upper-middle-class, and Ivy-educated. Breadth is teachable in an afternoon. Depth is a childhood. The mechanism does not reward the ability to switch; it rewards having been immersed, and it is specifically instrumented to detect the difference.

There is also a well-documented way the switching backfires when it becomes visible. The Dutch psychologist Roos Vonk called it the slime effect: warmth directed upward at superiors, paired with coolness directed downward, produces strong observer dislike and attributions of manipulation. Code-switching is punished precisely at the moment it becomes legible as switching—which is an awkward property for a strategy that requires an audience on both sides of the switch.

The most instructive finding in this neighbourhood, though, is the one that appears to say the opposite. Silvia Bellezza, Francesca Gino, and Anat Keinan reported in 2014 what they called the red-sneakers effect: under some conditions, visible nonconformity raises perceived status and competence, because observers infer that the deviant could have conformed and chose not to.† The conditions are the interesting part. Shop assistants in Milan’s luxury boutiques rated a nonconforming shopper as higher in status; demographically matched women recruited at Milano Centrale rated the very same person markedly lower. Familiarity with the environment was the only moderator that produced an outright reversal, and it was large. Intentionality was a hard requirement: label the deviation accidental and the effect vanished entirely. And a later independent study by Janneke Oostrom and colleagues found that low-qualified job applicants were punished for nonconforming dress while high-qualified ones were granted leeway.

Read together, these are not an escape route but a restatement of the trap in flattering language. Breaking the rules pays only for someone the room already credits, in front of an audience that knows the rule being broken, and only when the breach is legible as a choice. That is a fairly exact description of a person who did not need the advice.

VI · The Consolations

Every unequal arrangement generates a consoling literature, and this one is no exception. If the excluded cannot have the returns, perhaps they have the virtues. For about a decade the most-cited version of that consolation was empirical: a series of studies out of Berkeley reporting that lower-class individuals were more generous, more compassionate, and better at reading other people’s emotions than the wealthy, who were correspondingly more inclined to cut off other drivers and to help themselves to the sweets meant for children. It was a wonderful set of findings. Two of them appeared, briefly, to be about cars in traffic, which made them irresistible to newspapers.

They have not held up, and the manner of their collapse is instructive. In 2023 a team including one of the original authors on its review panel ran preregistered replications of the two field studies at two and a half times the original sample sizes. The first original had found an effect of .12 with two hundred and seventy-four vehicles, at a p-value of .047; the replication, with six hundred and eighty-three vehicles and eighty-eight per cent power, returned .013. The second original had found .17 with a hundred and fifty-two observations; the replication, with three hundred and ninety-four and ninety-three per cent power, returned −.028. Two preregistered replications of the earlier laboratory work produced an overall effect of .01.

Then, in 2025, the meta-analysis. Jiaqi Wu and colleagues pooled one thousand one hundred and six effect sizes from four hundred and seventy-one studies—two million three hundred and forty thousand participants across sixty societies—and found the association running the other way: higher-class individuals slightly more prosocial, at .065. A field experiment in the Netherlands supplies the likeliest explanation and rescues everyone’s dignity. Misdelivered envelopes containing cash were returned by eighty-one per cent of high-status households and thirty-eight per cent of low-status ones; envelopes containing a bank-transfer card of no cash value showed no significant difference at all. The gap was never about character. It was about what the money was worth to the person holding it.

One further piece of the consolation deserves correcting, because it is stated confidently and is precisely backwards. It is often said that the poor are perceived as less agreeable while the rich enjoy an easy likability. The stereotype-content research, among the most replicated frameworks in social psychology, finds the wealthy occupying the low-warmth, high-competence quadrant—the position that elicits envy rather than affection. Susan Fiske’s own formulation is that if rich people seem cold but competent, then working-class people seem warm but incompetent. The comparative paper on the subject is titled, with a directness the field rarely permits itself, Poor but Warm, Rich but Cold (and Competent). There is no warmth premium at the top to be explained. The premium is competence, and it is paid for in liking.

The reason to spend a part of this essay on findings that failed is not schadenfreude. It is that the flattering result travelled around the world while the correction stayed in the journals, and the same asymmetry is operating on the code-switching literature right now, in the same direction, for the same reason. A field in which the consoling story outran the evidence for a decade is not a field in which the next consoling story should be believed on sight.

VII · The Ceiling

Suppose all of it works anyway. Suppose the codes are learned, the interviews are passed, the offer is signed, and the person from the wrong background is now inside the firm, indistinguishable at the airport. This is the scenario in which cultural matching has done everything asked of it. It is also where the most robust finding in the whole area lives, and it is not encouraging.

Working from the British Labour Force Survey and a hundred and seventy-five interviews across four elite occupations, the sociologists Sam Friedman and Daniel Laurison established what they called the class ceiling: people from working-class origins who have reached elite occupations earn about sixteen per cent less than colleagues of privileged origin doing the same kind of work—on the order of six thousand four hundred pounds a year. In 2024 Laurison and Friedman ran the American version using the Panel Study of Income Dynamics, arriving at an analytic sample of some twelve hundred and ninety people across nearly three thousand person-years. The raw gap in top professional and managerial occupations was close to twenty thousand dollars a year. After controlling for age, gender, race and ethnicity, education, and occupation, it was still about eleven thousand seven hundred.

These are not people who failed to get in. They are people who got in. Whatever the codes are worth—and Part II establishes that they are worth a great deal at the door—they are not worth enough to be collected in full afterwards, and no amount of fluency appears to close the residual. Working-class-origin women in the American data fare worse still, at about nineteen thousand five hundred dollars below their privileged-origin colleagues, which rhymes uncomfortably with the commitment penalty that kept the higher-class women out of the interview room in the first place.

Honesty requires the same treatment of the romantic counter-story, which holds that the class migrant pays for the ascent in psychic coin. Bourdieu had a term for it, habitus clivé, the cleft habitus, and applied it to himself in Esquisse pour une auto-analyse. When Friedman went looking for it empirically, across thirty-nine life-course interviews with upwardly mobile Britons, he found it in ten of them—about a quarter, all of whom had travelled the longest distances. Fourteen, more than a third, had trajectories he described as largely psychologically smooth. He is explicit that this is not to say his subjects were unhappy people or suffered from psychological disorder; most, he writes, seemed to be battling their multitude of emotions valiantly, even perhaps successfully. Several reported a gain: a capacity for reflexivity and self-analysis that their settled colleagues lacked. The wound is real for some and it is not the general case, and the wider quantitative literature on mobility and well-being, once origin and destination are properly separated, finds upward mobility neutral to beneficial and concentrates the harm in mobility that goes the other way.

Which leaves an unglamorous conclusion standing on its own. The codes are learnable, partially and at cost. The advantage is not.

VIII · Anyone Who Plays Squash

A word on what this literature does not cover, since a New York magazine ought to say so plainly. Almost all of it is American and British, and nearly all of it is drawn from three industries—law, investment banking, and management consulting—which are not representative of professional life even in New York. The studies document how an interviewer can mistake social familiarity for competence; they cannot establish which particular pastime will carry that signal in every office. The comparative work suggests that the sorting mechanism travels more reliably than the content of its codes. A word on a résumé may open a door on Wall Street and mean little in a laboratory or a public school. The city contains all three, and the hiring decision still has to be defended on the work.

What does travel, and is the most solidly evidenced finding anywhere in this essay, is the voice. A meta-analysis of accent bias in hiring, pooling a hundred and thirty-nine effect sizes across four and a half thousand participants, found candidates with standard accents judged more employable at an effect size of 0.47. The bias was not moderated by how easy the speaker was to understand, which disposes of the polite explanation; its magnitude tracked the perceived social status of the accent, which does not. It was stronger for jobs involving a lot of talking, and stronger for women. Whatever else is uncertain here, the first three seconds are not.

It is worth being clear about who comes out of this badly, and it is not the banker. He was not being malicious over the résumé; he was doing what the research says people do, which is to reach for a proxy under time pressure and to experience the resulting recognition as insight. The failure is institutional and it is measurable, and it will not be fixed by teaching candidates to say squash. It might be reduced by structuring the interview so that the proxy has less room to operate, which is the one intervention the audit literature consistently supports and the one firms consistently find least appealing, for reasons that Rivera’s consultant identified in a single sentence.

The candidate, meanwhile, was ranked first for a reason that had nothing to do with her. Somewhere in her file there was presumably an account of what she had actually done, and it is possible that it was excellent. We will never know, because the conversation had already moved on to the airport in Minneapolis, where it was snowing.

Notes on Sources

Rivera, L. A. (2012). Hiring as cultural matching: The case of elite professional service firms. American Sociological Review, 77(6), 999–1022. Source of the nine-month participant observation, the hundred and twenty interviews (forty each in banking, law, and consulting), the ranking of fit above grades and experience, the industry breakdown (seventy-five per cent in law against forty in consulting), the observation that fit does not track a job’s social demands, the spot-checking footnote, the breadth-and-depth conclusion, and all quotations from evaluators. See also Rivera, L. A. (2015). Pedigree: How Elite Students Get Elite Jobs. Princeton University Press.

Rivera, L. A., & Tilcsik, A. (2016). Class advantage, commitment penalty: The gendered effect of social class signals in an elite labor market. American Sociological Review, 81(6), 1097–1131. Source of the audit design (316 applications, 316 offices, 147 firms, 14 cities), the callback rates in Fig. 1, the class-signal constellation, the survey experiment with 210 lawyers, and the commitment mechanism.

Bourdieu, P. (1977). Outline of a Theory of Practice, for bodily hexis; (1986). The forms of capital, in J. Richardson (ed.), Handbook of Theory and Research for the Sociology of Education, 241–258, for the root status of economic capital and the “conventional, constant, legally guaranteed value” of the credential; (1994). Rethinking the state: Genesis and structure of the bureaucratic field, Sociological Theory, 12(1), 1–18, for meta-capital and rates of conversion; and Bourdieu, P., & Wacquant, L. (1992). An Invitation to Reflexive Sociology, for the fish in water.

Kilduff, M., & Day, D. V. (1994). Do chameleons get ahead? The effects of self-monitoring on managerial careers. Academy of Management Journal, 37(4), 1047–1060. Day, D. V., Schleicher, D. J., Unckless, A. L., & Hiller, N. J. (2002). Self-monitoring personality at work: A meta-analytic investigation of construct validity. Journal of Applied Psychology, 87(2), 390–401. Lei, L., Wang, C., & Pinto, J. (2025). Do chameleons lead better? A meta-analysis of the self-monitoring and leadership relationship. Personality and Social Psychology Bulletin, 51(7), 1139–1158, for Fig. 2 and the quoted conclusion.

Vonk, R. (1998). The slime effect: Suspicion and dislike of likeable behavior toward superiors. Journal of Personality and Social Psychology, 74(4), 849–864.

Bellezza, S., Gino, F., & Keinan, A. (2014). The red sneakers effect: Inferring status and competence from signals of nonconformity. Journal of Consumer Research, 41(1), 35–54.† Oostrom, J. K., Ronay, R., & van Kleef, G. A. (2021). The signalling effects of nonconforming dress style in personnel selection contexts. European Journal of Work and Organizational Psychology, 30(1), 70–82.

McCluney, C. L., Durkee, M. I., Smith, R. E., Robotham, K. J., & Lee, S. S.-L. (2021). To be, or not to be…Black: The effects of racial codeswitching on perceived professionalism in the workplace. Journal of Experimental Social Psychology, 97, 104199; and the same authors’ “The Costs of Code-Switching,” Harvard Business Review, 15 November 2019, cited here as a practitioner article and not as evidence. Hagger, M. S., Chatzisarantis, N. L. D., et al. (2016). A multilab preregistered replication of the ego-depletion effect. Perspectives on Psychological Science, 11(4), 546–573.

Jung, M. H., Smeets, P., Stoop, J., & Vosgerau, J. (2023). Social status and unethical behavior: Two replications of the field studies in Piff et al. (2012). Journal of Experimental Psychology: General, for Fig. 3, replicating Piff, P. K., Stancato, D. M., Côté, S., Mendoza-Denton, R., & Keltner, D. (2012), PNAS, 109(11), 4086–4091. Stamos, A., Lange, F., Huang, S. C., & Dewitte, S. (2020), Journal of Research in Personality, replicating Piff et al. (2010), Journal of Personality and Social Psychology, 99(5), 771–784.

Wu, J., Balliet, D., Yuan, M., Li, W., Chen, Y., Jin, S., Luan, S., & Van Lange, P. A. M. (2025). Social class and prosociality: A meta-analytic review. Psychological Bulletin, 151(3), 285–321. Andreoni, J., Nikiforakis, N., & Stoop, J. (2021). Higher socioeconomic status does not predict decreased prosocial behavior in a field experiment. Nature Communications, 12, 4266.

Fiske, S. T., Cuddy, A. J. C., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype content. Journal of Personality and Social Psychology, 82(6), 878–902; Fiske, S. T. (2018). Stereotype content: Warmth and competence endure. Current Directions in Psychological Science, 27(2), 67–73; Durante, F., Tablante, C. B., & Fiske, S. T. (2017). Poor but warm, rich but cold (and competent). Journal of Social Issues, 73(1), 138–157.

Friedman, S., & Laurison, D. (2019). The Class Ceiling: Why It Pays to Be Privileged. Policy Press. Laurison, D., & Friedman, S. (2024). The class ceiling in the United States: Class-origin pay penalties in higher professional and managerial occupations. Social Forces, for Fig. 4. Friedman, S. (2016). Habitus clivé and the emotional imprint of social mobility. The Sociological Review, 64(1), 129–147, for the thirty-nine interviews and the fourteen smooth trajectories; Bourdieu, P. (2004). Esquisse pour une auto-analyse, pp. 127 and 130, for the term itself. Shi, S., & Gugushvili, A. (2025). Origin, destination, or mobility? A systematic review of studies using diagonal reference models. Research in Social Stratification and Mobility, 97, 101047.

Spence, J. L., Hornsey, M. J., Stephenson, E. M., & Imuta, K. (2024). Is your accent right for the job? A meta-analysis on accent bias in hiring decisions. Personality and Social Psychology Bulletin, 50(3), 371–386. Ashley, L., Duberley, J., Sommerlad, H., & Scholarios, D. (2015). A Qualitative Evaluation of Non-Educational Barriers to the Elite Professions. Social Mobility and Child Poverty Commission, for the “poshness test” findings in British law, accountancy, and financial services.

† marks a finding this magazine reports with reduced confidence. The red-sneakers paper is not retracted, carries no expression of concern, and is not among the papers investigated in the research-misconduct findings against its second author; both co-authors have stated on the record that she collected none of its data. It also has no posted data, no independent audit, and no direct replication, and is reported here on that basis—neither impeached nor clean. Separately, Erickson, B. H. (1996), Culture, class, and connections, American Journal of Sociology, 102(1), 217–251, is the standard citation for the proposition that cultural variety outperforms highbrow taste at work; its abstract states this plainly, but the full text sits behind a paywall this magazine could not pass, so the study’s setting and its outcome measures are not asserted here and the finding is not relied on in the argument. Figures 1 to 3 plot published values; Fig. 4 renders a published decomposition and is not a physical measurement. Part VIII applies the findings to New York as an editorial argument, not a citywide estimate.

Drafted with AI (Claude, Anthropic) and fact-checked against the primary literature listed above, August 16, 2026; every effect size, sample size, and quotation was re-verified at its source, contested findings are marked as contested, and the failed replications are reported as findings rather than omitted. Set, with affection and without affiliation, in the manner of a certain Manhattan weekly.New York revision drafted with AI on October 12, 2026; editorial review pending. ◆ ◆ ◆