Notes on: Reilly, W. (2021). Testing the Tests for Racism. Acad. Quest. 34 (3)DOI 10.51845/34.3.4

Dave Harris

Audit studies have emerged to test the claims of decreased American racism. Social scientists have pointed out that racism seems to be declining [references include Sowell and Reilly himself]. Other scholars specifically doing audit tests say that these still show 'modern era bias against blacks and other racial minorities', including Pager.

The evidence of racism declining includes an increase in approval of black/white interracial marriage, rising from 4% among whites in 1958 to 84% in 2013. A Pew survey in 2017 asks questions about whether interracial marriage was a bad thing, and fewer than 10% of whites agreed. Actual multi-race marriages among newlyweds rose from 3% in the mid-1960s to 17% today. Most other anonymous poles find similar results, including those asking for voting for various kinds of candidates for president — 8% of whites would not vote for black qualified same party candidates, 7% Catholics, 9% for Jews, 19% for Mormons in 2015.

Performance gaps in terms of income or sat scores often attributed to racism or genetic factors but when we make adjustments for nonracial characteristics, gaps in income between blacks and whites 'vanish almost entirely'(18) if we allow for 'years of education, median age, region of residence and any aptitude test score'. So one study showed that '"overall, black men and 82.9% of the white wage. Adjusting for black-white differences in geographic region, schooling, and age raises the ratio to 87.7%; adding differences in (standardised) test scores raises the ratio to 95.5%, and adding differences in years of work experience raises the ratio to 99.1%"'.

Similar adjustment has the same effect with rates of black encounters with police although this is more controversial. If we adjust for a black crime rate which affects police encounters with black people, this is often more than twice the white rate and so completely eliminates the disparity between 13% representation of blacks within the US population and the 25% representation of identified Blacks among those shot and killed by the police. At a higher level of analysis, 'including multiple victim and officer characteristics', 'whites are 27.4% more likely to be involved in police shootings than are similarly situated black Americans' (19).

However, audit studies 'conducted primarily by sociologists' finds almost the opposite, 'rather intense discrimination'. Pager is a famous case where attractive male college students were used to test labour market discrimination in Milwaukee in 2001. She found black applicants 'were a bit less than half as likely as white applicants to be called back for jobs — 14% versus more than 30%' and black applicants without a criminal record were still less likely to be higher than white applicants were admitted to having one, '14% versus 17%'. A similar but larger study used hardcopy resumes [CVs for Brits] and looked at the effect of '"black sounding" names. The results were 'depressingly similar… Black sounding names received only about two thirds as many callbacks'. Dozens of similar studies have taken place since, most of them since 2010, and most have found discrimination effects 'at least at the 7 to 9% level'.

So there is a contrast between 'multiple well designed regression analyses' which adjust for race neutral factors', and 'multiple significant audit studies' showing 'intense bias in the job market and for that matter the housing market'. The 'simple and greatly understudied reality of plain human adaptation to moderate levels of discrimination must play some role ("just apply for two more jobs online, Jamaal")' (20), but another answer rests with 'weaknesses and omissions which do exist within the audit literature'.

Sometimes there are 'personal hobby horses' and other questions [one is what would have happened if the application process had actually taken place in earnest — if there is a 14% final callback rate, this would imply that a black person needs to apply to only about seven more jobs get hired, so actual hiring rates 'for dedicated applicants' might be different]. Another problem is the same sector of the job market is being examined — the 'zone of entry-level private sector jobs, primarily with white owned employers'. The two original papers have been often cited and so the same methodology 'remains prevalent' with the same jobs databases tapping the same occupational categories — none involved college achievement beyond high school degrees, for example. The sector involved 'could well be the only large sector of the US economy where a qualified middle-class minority applicant would face significant discrimination in 2021'. Many other employers 'engage in very large scale programs of pro-minority affirmative action', which do convey 'a significant advantage over equally qualified whites'

There are other problems. It is often forgotten that 'the United States is now just 60.1% non-Hispanic white and minorities seem no less prone to the universal human sin of bias than whites'. Levels of in group preferences among various US populations' show that 'while most Americans "like"  one another well enough, black Americans preferred their racial group over alternatives by 15.58 points as versus 13.94 points for Asian-Americans, 12.83 points for Hispanics, and just 11.62 points for white conservatives' (21 – 22). Given this it surprising that no studies to date have looked at 'the 18.3% of businesses owned by members of racial minority groups', or the '38.2% of businesses owned by women and minorities combined'(22). It would be interesting to compare patterns.

At least one recent study examined 'co-ethnic hiring among a heavily – POC subset of new American business ventures' and saw 'a preference for hiring members of the same ethnic group as "ubiquitous" among immigrants' (22). Same group hiring, often amongst fairly small groups, averaged 22.5% and was sometimes as high as 40%. It will be interesting to see if similar patterns obtain among the owners of the 'thousands of successful minority businesses listed on [various] directories'.

A second problem is that audit studies find difficulties when adjusting for social class and other non-race variables. The most common indicator signalling blackness has usually been the use of the name 'which presumably sounds black'. However, 'mean levels of income and of socio-economic status vary widely among racial groups' so 'very black names (an example might be "Jamarrian")' are not associated simply with minority race but also with lower incomes and lower levels of parental education. These factors are difficult to adjust for and 'in fact almost never is adjusted for' (23). Where studies have avoided 'the risk of class signalling' by using 'ethnically distinct last names ("Washington" "Jefferson," "Garcia," "Hernandez")…. with relatively neutral first names… They find almost no effect of race on employer hiring preference. Despite criticism
[allowing for the ambiguity of first names], the authors of this paper points out that 'roughly 60% of all potential employers would have had to mistakenly believe their applicants were white in order for their primary results to be invalid' . Reilly himself comments that 'it is difficult to imagine a sizeable number of canny hiring managers mistakenly believing a "Carlos Hernandez" to be of Norwegian descent'.

Another study in 2017 found that high status black names 'are less likely to be perceived as black at all: those with higher education levels who call their kids '"Nia, Malcolm, or Malia"' are less likely to have them perceived as black as opposed to those who call them '"DaShawn, DaQuan or Lakisha"', and there is disparity with names associated with middle-class black people including '"Bria, Sade, Kylah,Lyric and Jasmine"'. These are common names and so they should lead 'at the very least in decreased discrimination', since they were perceived as black 'by under two thirds of study participants'.

The study showed that the class association of names exists for whites as well, and if anything, respondents can more correctly identify education levels of mothers. Sometimes, names associated with poor whites, especially '"Cheyanne"' were wrongly attributed as black names.

It now seems critical for audit studies intending to measure race effects to adjust for class and to use only names known to be associated with black holders and also verified by respondents as being associated with middle-class status [that is not working class or low levels of maternal education]. It might be interesting to compare resumes with stereotypical lower income white first names [he cites "Earline" and "Chevy"] and whether they are perceived as white — they might face discrimination on both class and race grounds. Another experiment might involve submitting resumes with a minority-associated last name and simply a first initial.

Another complexity is adjusting for employer or renter perceptions of probable qualifications. It might at first glance be simple racism if an employer is 1/4 less likely to hire a black rather than a white graduate from the same college or to prefer a white college graduate rather than a black Yale graduate. But there's an obvious alternative — employers 'are aware of affirmative action' and 'that black Ivy League and a white Nittany Lion [?] might well have virtually identical SAT or GRE qualifications' (25). 'This is not at all an exaggeration — affirmative action effects have lasted for decades and 'are objectively very large'.

Thus the medical school acceptance rates for doctors with an MCAT of 23 to 26 was '6% for Asian-Americans, 8% for whites, 31% for Hispanics, and 56% for blacks. At the higher level of 30 – 32 (paired with a GPA of 3.6 – 3.79 in the dataset) the same rates were respectively 58%, 63%, 83% and 94%' (25). The same sort of thing seems to apply in law schools. 2010 to 15 acceptance rates for students in the top decile of academic classes were: '10% for Asians, under 20% for whites, around 35% for Hispanics and close to 60% for blacks'.

In 2017 the mean combined math/verbal SAT scores were 941 for blacks, 963 Native Americans, 987 Hispanics, 1118 whites, 1181 Asians, a total black white gap of 177 points, given that there are only 1600 points possible on the test. If we were to add this quantitative metric representing aptitude [he thinks it is 'solid enough'] to resumes it would be interesting to see the effect it would have on employer behaviour and auditors should do so during future studies — employers are certainly 'aware such gaps exist'.

So there are problems with whether this is a correct measurement, but also the issue is whether this prejudice corresponds to that which is experienced by members of other populations. The 'well-designed totally anonymous' Gallup poll found an 8% rate of voter bias against blacks, but higher rates against Hispanics and Jews and Mormons. Few audit studies have ventured beyond traditional white/black comparisons, but those that have seem to reveal similar patterns of discrimination. So those with Asian last names 28% are less likely to receive employer callbacks than white applicants, those that stated Islamic faith 32% fewer employer emails and 48% callbacks, and 'women with "feminine sounding" names are less likely to be hired as attorneys'(26)

These are 'esoteric varieties of bigotry' and are of course not morally acceptable, but 'that's the point'. Although they encounter prejudice, Americans of Asian and Middle Eastern descent 'are arguably the most successful people in the country and female lawyers do well. If there is a similar baseline level of prejudice against the 89% of the country composed of racial minorities and women, it might mean that 'the simple existence of bias does not mean that said bias explains all observed discrepancies in performance between groups'. Auditors should test this '"many biases" theory, say by including photos with resumes where appropriate so they can 'measure the effect of physical fitness and perceived attractiveness, by looking more at prejudice against gay applicants… And by conducting social class focused studies like the test described above'.

Overall audit studies are a sub genre of research that obviously do indicate that bias remains a reality within significant sectors of employment and housing. They rarely examine rates of pro-white or pro-POC bias in higher education, the public sector and the minority business community. They fairy frequently do not adjust for social class or perceived competence. They have not extensively compared the bias faced by members of other potentially disadvantaged groups. As a result they do not seem to 'counter the basic observation that citizens of different races with the same background characteristics often perform similarly in life'. They should take into account these criticisms and go forward.