What We Don’t Know Because of Polarization

Partisanship makes most media useless for understanding "grooming gangs", and new research shows psychologists self-censor evidence for "taboo" conclusions. Issue #134

One of the terrible things about the culture war is that it makes it impossible to get clear-eyed information about controversial topics. When the outrage machine starts churning and neither side is interested in what the other knows, everyone suffers. The last few weeks gave us two stark examples: the travesty of “grooming gangs” discourse, and new research showing that psychological researchers self-censor taboo evidence.

Good luck getting up to speed on “grooming gangs”

A few weeks ago the issue of UK “grooming gangs” exploded into public discussion after a series of posts by Elon Musk. As this issue flooded my feeds, I found that hyper-polarization made it nearly impossible to make sense of the story. Our media systems – publishers and algorithms alike – have failed to serve us an honest and accurate picture.

As the New York Times reported, “The phrase refers to a decade-old scandal involving a series of child sex abuse cases in which girls were assaulted and raped by gangs of men in several towns and cities. Most of the perpetrators were of British Pakistani heritage.” There were thousands of victims over decades.

This is a very sad and complicated story. The race (and religion) of the perpetrators also makes it an extremely controversial topic. As a result, few sources seem interested in telling the story in a thoughtful, comprehensive way.

On one side, we have Musk and others who see this as an attempt by the government and media to cover up race-based violence against young white girls by South Asian men. This is being used to argue that immigration is harmful, that wokeness is destroying civilization, or that “multiculturalism has failed.”

On the other side, people on the left have argued race is irrelevant, that this is an instance of toxic right-wing rhetoric run amok, that this story is only garnering attention now because it serves a racist narrative, and that anyway most pedophiles are white.

Neither of these positions effectively informs society. The right is suggesting a cover-up, yet we now have two decades of media reports, government inquiries, hundreds of arrests and dozens of convictions. Yet the left is mostly ignoring multiple credible reports from police officers that they were discouraged from investigating these crimes because of racial sensitivities.

In other words, we are caught in a dysfunctional dynamic where the left justifies suppressing evidence because the right weaponizes it. One early incident perfectly captures this pattern:

In 2004, [filmmaker Anna] Hall made Edge of the City, an investigation into Bradford grooming gangs, for Channel 4. The broadcast was delayed for three months after the [far right] BNP tried to exploit it as party propaganda

We’ve seen versions of this play out in other contexts. After a long sorry history of racist crime coverage, American media is now less likely to report the race of perpetrators if they are non-white. But this corrective is counterproductive if race is legitimately part of the story, and today the British public does not trust the BBC or the government to provide accurate information on the grooming gangs issue (though they trust Musk even less).

There isn’t even agreement on whether race is, numerically, a factor here. In theory the relevant statistics will tell us, but we seem to be caught in the middle of a math fight. People on the left point out that the vast majority of convicted child sex offenders in the UK are white (92% in one tally). People on the right point out that the majority of perpetrators of group abuse are South Asian (84% by one count). There’s a lot of complexity here, with the statistics collected over different time periods, from different areas, using different definitions – and many police reports didn’t record ethnicity at all. Asking if “crime” is up or down has always had these sorts of definitional problems; in this case it’s even more complicated.

Here too, honest brokers are scarce exactly where we need them most. I have yet to encounter an analysis of the statistical evidence that wasn’t also trying to push a particular racial politics narrative. (Yes the perpetrators are mostly South Asian; no they aren’t.)

More fundamentally, it’s worth asking why we’re so fixated on racial statistics. Even if everyone agreed that the perpetrators were disproportionately Pakistani men, what then should we make of it?

This is where our polarized media systems have failed us most. I haven’t anywhere seen the reasonable take, which might go something like this: A 2020 UK government inquiry acknowledged that “community and cultural factors are clearly relevant to understanding and tackling offending.” One Pakistani community member claimed in 2014 that local leaders were aware of the problem and “talked in mosques but not to police.” Culturally, Pakistan is currently ranked 164 out of 193 in the UN’s Gender Development Index and child brides are still common. We need to be careful here, because all of this plays into racist stereotypes of backwards “folk devils” and rhymes a little too closely with old-fashioned lynching to protect the virtue of white women. Some people will conclude that South Asian men are dangerous, but that commits the base rate fallacy: “most perpetrators are South Asian” doesn’t imply “most South Asians are perpetrators.” Thus there are potential harms from discussing race candidly, but that doesn’t mean that color-blindness is the right response. Targeted and culturally-specific outreach, education and enforcement may be appropriate.

Our information ecosystem — publishers, bloggers, and algorithms alike — has failed because it has not produced an honest and compassionate discussion about racial and cultural factors. Instead we have seen either blanket denials that race is relevant, or calls to punish an entire group for the crimes of a few—including punishing potential immigrants who had nothing to do with any of this. So far, we’re simply not having a grown-up conversation about this issue. (Instead, we get Musk calling for overthrowing the British government.) And it’s not the culture warriors who will suffer. As one long time victims’ advocate put it:

Those “focused exclusively on the ethnic origin and religious affiliation of a particular set of abusers”, she added, are “not really interested in the girls at all”.

Self-censorship in psychology

Meanwhile, new research looks at how American psychology professors handle research on controversial topics, and finds two notable things: there’s wide disagreement on the scientific evidence for ten “taboo conclusions,” and almost everyone is worried about social sanctions that could come from publishing results that support them. The result is that scientists who have found evidence on one side of these debates have censored themselves.

In this study, Cory Clark and her colleagues presented a group of academics with PhDs in psychology or related fields with ten taboo conclusions. These are a remarkable map of how the broader culture war intersects with scientific inquiry:

1. The tendency to engage in sexually coercive behavior likely evolved because it conferred some evolutionary advantages on men who engaged in such behavior.

2. Gender biases are not the most important drivers of the under-representation of women in STEM fields.

3. Academia discriminates against Black people (e.g., in hiring, promotion, grants, invitations to participate in colloquia/symposia).

4. Biological sex is binary for the vast majority of people.

5. The social sciences (in the United States) discriminate against conservatives (e.g., in hiring, promotion, grants, invitations to participate in colloquia/symposia).

6. Racial biases are not the most important drivers of higher crime rates among Black Americans relative to White Americans.

7. Men and women have different psychological characteristics because of evolution.

8. Genetic differences explain non-trivial (10% or more) variance in race differences in intelligence test scores.

9. Transgender identity is sometimes the product of social influence.

10. Demographic diversity (race, gender) in the workplace often leads to worse performance.

There was strong disagreement among the professors as to the truth or falsity of the ten statements. But, Clark pointed out, “Professors more confident in the truth of the taboo conclusions reported more self-censorship, a pattern that may bias perceived scientific consensus toward rejecting controversial conclusions.”

Most professors, tenured or not, said they feared how people would react if they spoke about their beliefs. Professors who were younger, left-leaning, and female tended to be more inclined to punish peers who publish controversial research, though most everyone expressed contempt for petitioning to retract papers on moral grounds.

Overall, Clark writes, “there was a great deal of disagreement regarding the accuracy of various controversial conclusions, but also general agreement that scholars should not be discouraged from pursuing even the most taboo conclusions.” And yet, her research also finds that most people do hold their tongues at least a little for fear of how their peers will react.

We are forced to conclude that, on average, we should believe each of the above conclusions somewhat more than published evidence would support.

Quote of the Week

me: oh boy i bet the right will make so much hay out of the fact that we capped insurance premiums for uninsurable homes that are now burning to the ground. and they'd be totally correct about that...

the right: WOKE FIREFIGHTERS

— @andersonbcdefg