How to be an anti-racist
Diversity training doesn't work — here's some stuff that does
Stanford political scientist Hakeem Jefferson remarked the other day that “what's so wild to me about all of these arguments for ‘popularism’ is that popularists never seem to entertain the idea that the real problem of our politics is not wokeness but white racism.”
This is, I think, pretty unfair and mostly an example of how searching for “the real problem” can be somewhat fruitless. On climate policy, for example, it’s pretty obvious that “the real problem” is that many people do not care as much as they should about harms that occur in the future or harms that accrue to people living in foreign countries. To say “I would simply get everyone to care a lot about problems afflicting residents of island nations and low-lying river deltas in the 2080s” is not really a constructive engagement with the issue. I think all left-of-center factions now basically agree that the way forward on climate needs to deal with the electorate’s aversion to explicit carbon pricing rather than wishing it away.
Where racial prejudice is different, it seems to me, is that it’s a social problem over and above its political manifestations, so it’s not crazy to ask people to try and address racism in addition to trying to cope with it politically.
I don’t have an incredibly hot take on how to do that beyond the observation that I think this is somewhat empirically understudied. American higher education is full of internal diversity initiatives and also lots of well-qualified social scientists who could be trying to evaluate these programs. Whole graduate schools of education exist that could, in theory, be attempting to empirically evaluate the efficacy of various anti-racism classroom initiatives. This is mostly not being done, and the evidence available on corporate diversity trainings is pretty negative. But there does seem to be some stuff that works.
Unfortunately for people who don’t like the advice that politicians should pander more, as best I can tell, none of the literature seems to support the idea that in-your-face calling-out tactics are effective. What seems to work best are fairly gentle suasion tactics plus efforts to get more people into casual integrated interactions.
“Deep canvassing” might work
Progressives like the study by David Broockman and Joshua Kalla in which they did a field experiment showing that when trained canvassers went door-to-door and talked to people, they were able to reduce transphobic sentiments.
This involved non-confrontational 10-minute conversations “encouraging actively taking the perspective of others,” and they found that “these conversations substantially reduced transphobia” and “also increased support for a nondiscrimination law, even after exposing voters to counterarguments.”
It’s a great paper.
In terms of its relevance to contemporary issues, I would note that it does seem challenging to bring this to mass scale. More important, I think, is that in this context, “reducing transphobia” means things like saying you disagree that “it is morally wrong for a man to present himself as a woman in public” and agreeing to support a pretty basic housing and employment non-discrimination law.
When we talk about the impact of racial views on electoral politics, we’re not living in 1968 and generally not talking about the votes of people who want to repeal the Civil Rights Act. For the record, though, the Civil Rights Act was clearly above water in 1964 (though with a large undecided bloc) and polls as overwhelmingly popular today.
There is also deep canvassing literature on undocumented immigrants, and in general, this seems like a promising model for investment, albeit a costly one. I just want to flag for caution that I don’t think the existing studies really test for what contemporary progressives are interested in with regard to race.
“Diversity training” is pretty bad
The question of whether corporate diversity training initiatives work has been pretty extensively studied.
Business executives believe that doing these programs has genuine value to the bottom line in terms of protecting them in the face of lawsuits, so they are fairly widespread. Critically, however, the lawsuit-protecting attributes of training do not require the trainings to be effective, and they generally are not. Indeed, as this summary from Frank Dobbin and Alexandra Kalev in Harvard Business Review hints, the main question in this literature is whether the trainings backfire by annoying people:
Do people who undergo training usually shed their biases? Researchers have been examining that question since before World War II, in nearly a thousand studies. It turns out that while people are easily taught to respond correctly to a questionnaire about bias, they soon forget the right answers. The positive effects of diversity training rarely last beyond a day or two, and a number of studies suggest that it can activate bias or spark a backlash. Nonetheless, nearly half of midsize companies use it, as do nearly all the Fortune 500.
Some of the backlashes can be very bad. Leigh Wilton, Evan Apfelbaum, and Jessica Good find that emphasizing themes of multiculturalism can increase subjects’ belief in race essentialism (consider Tema Okun’s work in this light) while Madeline E. Heilman and Brian Welle find that when teams are assembled with an explicit diversity goal in mind, women and Black group members are perceived as less competent, and “this effect occurred regardless of the proportional representation of women or the degree of the groups’s heterogeneity.”
I don’t think many on the left are actually super enthusiastic about these diversity trainings, but the general sense is also that only a bitter crank would actually complain about them. But there is real evidence that they are at least sometimes making things worse, which strikes me as a big deal. For example, Michelle Duguid and Melissa Thomas-Hunt find that when you tell people that stereotyping is widespread, they stereotype more.
This suggests to me that a very underrated step toward progress would be to eliminate the judicial and legal standards that suggest diversity training has litigation-protective effects. That could have three benefits:
Denying companies’ legal protections on the basis of having diversity training programs could help plaintiffs in discrimination cases obtain justice.
Doing fewer diversity trainings would, based on the evidence, likely somewhat reduce racism.
Companies that don’t want to get sued would have an incentive to invest money in designing and identifying effective programs because a program that actually reduced bias incidents would still be valuable.
One important finding that runs through this literature is that mandatory trainings have a higher risk of backlash. I think that’s a general pattern here, which is that more forceful anti-racism measures tend to be less effective.
Beware counterproductive intersectionality
Unfortunately, Erin Cooley, Jazmin Brown-Ianuzzi, Ryan Lei, and William Cipolli have found that if you take a group of white people who have progressive views on social issues and give them something to read about white privilege, they become less sympathetic to the plight of poor white people. Now to be clear, liberals start out more sympathetic to the poor than conservatives, and the decline in sympathy for the white poor doesn’t reverse that relationship. But relative to baseline, reading about white privilege causes liberals to adopt “greater punishment/blame and fewer external attributions for a poor white person’s plight.”
This is bad on the merits; it’s also incredibly politically counterproductive.
The problem isn’t — or at least isn’t only — that talking a lot about white privilege might alienate moderate white people and cost you their votes; talking about white privilege makes liberal white people less supportive of efforts to help low-income white people. And as Slow Boring has frequently argued, race-neutral economic redistribution is the most substantively and politically effective tool for narrowing racial gaps in outcomes of interest.
The white poverty rate is 7.3%, much lower than the Black (18.8%) or Hispanic (15.7%) rates. But there are twice as many non-Hispanic white people as Black or Hispanic people combined, so even at the lower rates, around half the poverty population is non-Hispanic white. And it’s true that while there are plenty of policy ideas for helping the poor and near-poor (expanded Child Tax Credit, Medicaid expansion, make housing assistance a fully-funded entitlement, etc.) nobody right now actually has a program that would somehow target assistance away from low-income white families and only benefit non-white ones. What’s more, Paul Frymer and Jacob Grumbach find that “gaining union membership between 2010 and 2016 reduced racial resentment among white workers” so in the case of labor union race-neutral progressive economic policy would not only have racially egalitarian outcomes it could actually shift racial attitudes.
Posting is praxis
Lena Song, a Ph.D. candidate at NYU, has a job market paper out on a subject near and dear to my heart — the value of tweeting.
She did some online experiments to try to test the impact of exposing people with different pre-standing beliefs to different kinds of racial justice content. She finds that “racial moderates are persuaded and become more progressive after the exposure to moderately progressive content,” but that this content has no impact on conservatives or progressives. But Song also finds that “extremely progressive content generates a backlash for racial conservatives” and that “racial conservatives and moderates rate more progressive content as less informative, less reliable and more objectionable.”
I find that heavy social media users have a somewhat ambivalent relationship to the idea that what they post matters.
Most normal people don’t like the idea of being subjected to the kind of message discipline that a candidate for office would need to use. At the same time, I assume that anyone who is posting things on an unlocked Twitter account is hoping that somebody reads them. And if you’re seeking an audience, you should probably care about the impact that your content has on your audience.
The idea that center-left content can persuade moderates while far-left content provokes backlash seems pretty intuitive and is probably applicable beyond the racial justice sphere (tweeting about “climate emergency” rather than “climate change” makes you seem less credible, for example). But the empirical evidence we have is focused on racial justice, and it suggests that if you are very fired up about this subject, you should absolutely continue to post about it — but also that to post effectively, you probably need to trim your sails somewhat.
It’s important, I think, to be mindful of the difference between expressing your feelings and influencing others in the desired way. Remember the study that said telling people stereotyping is widespread leads to more stereotyping? There’s also a study from Sohad Murrar, Mitchell Campbell, and Markus Brauer asking what happens when we tell people “that their peers hold pro-diversity attitudes and engage in inclusive behaviors.” It turns out that this makes white students have more positive attitudes toward outgroups, while members of marginalized groups report “an increased sense of belonging” as well as higher grades and better treatment from their peers.
Across almost all issues, I feel like there is conventional wisdom among progressives that acknowledging progress is inherently right-wing. If you say that living standards have risen dramatically over the past 30, 50, or 100 years, you’re somehow an apologist for every feature of the status quo. I don’t think that’s true in any area, and specifically on race, spreading good news seems more productive than telling people bigotry is lurking behind every corner.
Integration is good
The government in Singapore doesn’t put a lot of value on individual autonomy, so they force every neighborhood and every high-rise apartment building to roughly match the country’s overall ethnic balance.
From the standpoint of the government, this solved a problem that existed early in Singapore’s history where neighborhoods dominated by the large Malay minority in the country would become hotbeds of anti-government agitation. Now there are no Malay neighborhoods and no Indian neighborhoods, either. Every neighborhood is mostly Chinese, but with Malays and Indians integrated — no minority clusters or ghettos.
The United States is not like that, but freshman dorms at American colleges sort of are — central planners dictate where people will live and force racial integration. And that lets us ask, “what if you are assigned a Black roommate?” The answer is that students assigned to a Black roommate end up with more Black friends, even if you exclude the roommate from the count. A somewhat similar survey looked at peer group assignments in the Air Force Academy, with broadly similar results.
This general phenomenon is the Contact Hypothesis — that actual interaction with members of diverse groups will lead to less prejudice. And while Contact Hypothesis doesn’t hold up in all cases, meta-analyses tend to strongly support it overall. A really interesting study by Xuechunzi Bai, Miguel R. Ramos, and Susan T. Fiske finds that “at national, state, and individual levels,” places with more diversity feature less stereotyping. Detailed research from Census records suggests that white kids who grew up living next door to a Black family are more likely to grow up to be Democrats. There is a similar outcome based on the rise and fall of integration-promoting busing in North Carolina.
We’ve obviously not going to go Full Singapore, but we should absolutely reform zoning laws to promote more integrated neighborhoods and (as D.C. does) draw school boundaries to err on the side of integration rather than doing the reverse the way most localities do.
Limited political impacts
This is all good stuff and suggests we could make a fair amount of progress on racial issues by dropping dumb defensive trainings, getting racial justice advocates to exert message discipline, emphasizing the “good news” about America’s standing as a relatively high-diversity/high-tolerance country, and most of all by emphasizing the large economic benefits to everyone of housing reform.
What I don’t know is whether these changes would do the kind of work that people annoyed about popularism want it to do.
At the end of the day, the electorate is roughly 70% white, and white people are overrepresented above and beyond that in Congress. And no matter what you do to diminish bias, stereotyping, and other racist attitudes, I think most voters are going to be mostly selfish most of the time. This means that a politics that really centers the idea of benefits for a smallish minority of the population is always going to be electorally counterproductive. A recent Pew poll found that most Hispanic people think that we as a society pay too much attention to anti-Black racism and not enough to anti-Hispanic racism. That’s just how people are; they want to hear about how politicians are going to help them.
For example, Paul Frymer and Jacob Grumbach find that union membership reduces racial resentment among white unions members, which is great. But obviously white people aren’t going to join a union because they heard it will make them more sympathetic to Black people. You would need to convince them that joining a union is beneficial to them or convince voters and policymakers that pro-union legislative changes are good for most people.
There’s a reason Martin Luther King, Jr. launched a Poor People’s Campaign, and it’s not that he was averse to talking about racism or fanatically committed to a doctrine of strict colorblindness. He and Bayard Rustin and William Julius Wilson and other now slightly unfashionable thinkers had this fundamentally right, up to and including the fact that even dramatic reductions in racism wouldn’t make the basic issue go away. But separate from those big picture material concerns, it really is bad to have people walking around with biased behavior and doing lots of malign stereotyping, and we really can and should do things to reduce that.
Thanks for the kind words about my study with Josh Kalla. I agree with what you wrote regarding the fact that a) we haven't studied racial prejudice yet and b) this kind of canvassing is far too costly to scale nationwide -- and even if you did, although the effects are large by social science standards, they are not large enough to transform public opinion in any sense. I think the practical upshots of our studies are: a) some insights everyday people can apply in interpersonal conversations (e.g., don't lecture, have a two-way conversation), b) some ideas that might be able to be applied in more scalable mass communication (e.g., tell stories of individual outgroup members!), and c) a tactic that could make sense in small elections (e.g., there were some ballot initiatives on immigration and trans issues in small to mid sized towns where it actually is plausible you could try to canvass the whole electorate).
One theme in the literature on contact is that it really depends on whether the contact is positive-sum or neutral or negative-sum. Even neutral, shallow contact between groups seems to increase prejudice quite often (see Ryan Enos' work). Matt Lowe has a great experiment on the difference between positive-sum and negative-sum contact and as you might expect, positive-sum contact between groups really reduces prejudice but negative-sum contact increases it (https://www.aeaweb.org/articles?id=10.1257/aer.20191780). I think the corollary of these kinds of findings is not so much that integration is good per se, but that we have to pay attention to the kind of integration. An example of a win that is nicely consistent with your and my policy preferences: there's a doctor shortage in the US, so we should let doctors from other countries immigrate here and work in rural areas where there is a lot of anti-immigrant sentiment and also a big doctor shortage! That is going to do wonders for people's views towards immigrants -- and in fact, there is some research consistent with that intuition: https://www.pnas.org/content/118/14/e2022634118.short.
I agree with this post and with the SB crowd on most days. Maybe it’s just me, but it feels like the tone these days has drifted into outright hostility toward any and all forms of antiracist effort. Matt tagged Hakeem in his tweet of this post today, and Hakeem retweeted it. If Hakeem or his supporters read these comments, I’m pretty sure he’d come away feeling even more self-assured that Matt is just feeding the white racist trolls. Hakeem’s initial tweet was “why don’t the popularists condemn white racism?” It feels like a strawman to me as most people do in fact condemn what they consider to be racism, they just might not agree with Hakeem on what racism *is.* But as this group gets more and more comfortable throwing around inflammatory racial statistics without any caveats, and shows less and less interest in solving real problems, I dunno, it’s starting to feel icky. Is it just me? Can we keep our eyes on the prize which is finding ways to reduce racial animosity and severely adverse racial disparities at the same time?