Read on Twitter
Re: scientific racism, this conversation between @rasmansa and one of the authors of the religion, violence, and IQ study is instructive https://twitter.com/EPoe187/status/1271271074320322561
Bo is careful to dodge the following implication: If IQ is a valid measure, and if the data are good, then by normal clinical standards an astonishingly large number of people living in these countries would be classified as intellectually disabled. https://twitter.com/rasmansa/status/1271263227016556544
If we reject this implication, because of strong testimony that contradicts it, then we also must reject at least one of the argument's premises.
First, we could reject normal clinical standards. Bo doesn't do this.
Second, we could reject the quality of the data. Bo sticks to his guns, claiming these are the best data available. https://twitter.com/EPoe187/status/1271265429852499970
(He does not consider the possibility that, although these may be the "best" data available, they are still quite bad, as detailed in this informative thread.) https://twitter.com/RebeccaSear/status/1271547090221572096
Finally, we could question the validity of IQ on the grounds that it is biased or confounded. For obvious reasons, Bo doesn't do this.
At about this point, normal academic debate would stop. All three premises are open to attack, and I personally find the criticisms of the data quality damning. This should be enough for us to reject the paper and its implications.
But there are deeper questions at hand: Why do research using data with such obvious shortcomings? And why publish a report that makes clear, astonishing implications while publicly rejecting those implications?
This is where we see some key red flags I want to make explicit.
This is where we see some key red flags I want to make explicit.
Here, it's tempting to impute intentions, to guess the true beliefs of the author. I want to avoid that, and instead talk only about how the form of their writing supports some real world uses and not others, because I think this is important for identifying scientific racism.
First, vagueness. Vagueness affords innuendo. "Well, the measure is valid, the data are good, the clinical standards are normal. What could be going on here? Something's up. Are most Africans mentally disabled? I don't know. Who ever will solve this mystery." Trump does this.
Vagueness also affords plausible deniability. If someone accuses Bo of claiming an astonishingly large number of Africans are mentally disabled, he can say "That's not what I said." He merely laid out premises from which the conclusion follows. https://twitter.com/EPoe187/status/1271269178201210880
This implicative strategy can, perversely, shift responsibility for the implication to the reader. "I'm not racist, you're racist for making this horrifying implication!"
Another feature is treating strong conclusions as if they were weak.
The report comes to the strong conclusion that religion causes violence in low-IQ areas (which happen to be mostly black). They strongly emphasize their certainty in their statistical methods. But they also soft pedal, saying "Our results should be interpreted with caution." Why?
Understatement is a close cousin to vagueness because it also affords plausible deniability. Importantly, understatement confers plausible deniability on those who disseminate the reserach. "They didn't say we should re-colonize Africa, they merely said more research was needed!"
And just one more for now: Selective reverence for previously published work. In this case, the bad dataset. Here, Bo defends use of the dataset by noting that its author "has responded to criticisms." This is obviously insufficient. https://twitter.com/EPoe187/status/1271248162833997824
Watch for reverent treatment of work that supports authors' claims, and intensely skeptical treatment of work or other evidence that appears to contradict them.
An important result of publishing using disputed data is that it may affect the outcome of the dispute! The more published research appears to use a dataset, the more reasonable such uses appear to be.
To review: Implications that are discussed only vaguely, sweeping claims that are strangely understated, and reverence for previously published work even if it is under dispute. These strategies take advantage of what I take to be widespread norms in peer review.
The norms: Do not impute intentions. Take authors completely literally—if they don't say it, they don't mean it, and if they do say it, that's all they mean. Trust that previously published findings have merit. Do not preemptively hold authors responsible for misuse of their work
In collegiate, collaborative settings, these norms make some sense. But they fall apart in adversarial settings.
I suspect editors are looking for the wrong kinds of adversaries. Typically, editors are on the lookout for normal scientists overclaiming to attract media attention and to bolster their own reputations. But scientific racists strategically understate their most sweeping claims!
Editors may also be on the lookout for explicit or implicit signs of prejudice or hatred, or language that seems designed to inflame hatred in the reader. But this is not the goal of scientific racism.
The risk is not publishing work that discriminates by making astonishingly strong claims or inflaming hatred. The risk is publishing apparently weak or minimal claims that, if interpreted as straightforward facts, could be used by institutional policymakers to justify prejudice.
Importantly, I think editors and reviewers should be on the lookout for these issues even if there is no sign that the authors intend to justify prejudice!
I don't think this requires a paranoid attitude and I don't think it's that tricky. The thing to do is simply ask oneself: Could this work be used to justify prejudice? If the answer is yes, then careful evaluation beyond the usual standard of peer review is required.
Maybe the answer is, "Yes, but only if it's misunderstood." Time to clarify. It might be, "Yes, but the benefits may outweigh the costs." Time for deliberation. What's shocking about this paper is that the answer is "Yes, it appears to justify a return to full-blown colonialism."
Even though the claims are carefully hedged, even though there's no language that seems designed to inflame hatred—the racism here is not subtle. But you have to know what you're looking for.
