Changing My Mind About Race and IQ

Changing one’s mind is something everyone likes in theory, though much fewer actually do, when on the internet a pundit tells you about how he changed his mind about something, it will likely be a tale of a long ago conversion to his (currently held) idea. Back when he held the wrong-headed idea he had the foresight not to write about it. In religions both literal and ideological, the conversion narrative is often central, Evangelical Christians have their born-again experiences, racial realists often discuss their “racial awakening.” These narratives are welcoming to the man who changes his mind only in one direction. They are problematic because they turn ideas into an identity, be one of the awakened ones, White man! Don’t be one of the sheep! The virtue-signallers, of course, are kings at this, their whole identity is de-rooted from any race, ethnicity, culture, or religion, an identity based on holding the virtuous views. For the Evangelical Christian, the confluence of ideals and identity is made plain, for the ideologue, it must usually be denied.

But if you want to have the opinions which are most likely to be objectively true, it is best to try to separate out your identity from your opinions. In the words of CGPGrey, opinions “are in a box you carry with you and should be easily replaceable if they turn out to be no good.”

For the last few years I’ve believed in, and written under my pseudonym about, hard racial differences in intelligence. I was quite sure that most or all of the gap between the measured IQ of Black and White Americans was genetic in origin, everything I observed seemed consistent with that, and I assumed it would eventually be proven scientifically.

Yet, after reading, and re-reading, the entire race-IQ series by Chanda Chisala, I am now agnostic as to the real cause of the intelligence gap. I am not saying I agree with everything in the series: for instance, he makes much of the fact that African and Caribbean immigrants outperform African Americans. Chisala cites the fact that they have taken many of the Ivy League seats supposedly set aside for the descendants of American slaves, they shouldn’t, Chisala says, because African Americans’ 20% White admixture should give them an advantage. But the White admixture in African Americans should only give them an advantage of a few IQ points. Immigrant selection, perhaps on the slave ships where the stupider were captured, perhaps on modern planes where the smarter made it out of Africa, can easily do away with such a small gap. In addition, it should be noted that motivation matters a lot in terms of getting into the Ivy League, it’s not simply a matter of innate ability. Harold Ekah, cited by Chisala as an example of African immigrant achievement, co-authored an e-book about how he got accepted into all 8 Ivy Leagues. He says about it:

The biggest tip: Start early. Literally. Both men would begin their days by 5 a.m. during their senior year in order to get more work done.”

Who wants to do that?

But what convinced me were two things, African performance in the UK and African performance in scrabble.

African Performance in the UK

Chisala cites this study of CAT scores at age 11 for British pupils of different backgrounds. It has a respectable sample size(Black African N=2,197) and shows Black African scores of 92, 94.1, and and 94.1 for its verbal reasoning, quantitative reasoning, and non-quantitative reasoning, respectively, on its IQ like scale. You can massage the data downward a bit, to 90.5, 93.5, and  92.9 if you set the White British score at 100 and the standard deviation of the White British score at 15. When I first saw the data years ago, when someone other than Chisala cited it, I thought it was entirely believable from a hereditarian perspective given the immigrant selection of African migrants to Britain, who are demonstrably much more educated than their peers who remained in Africa. The difference in scores between where it “should be” and where it is is no larger than that which is regularly seen when highly-educated Whites segregate themselves into rich college towns.

Next, Chisala cites the performance of various ethnic groups on the GCSE tests. These showed Black African pupils nearly equaling Whites. This was not at all what I would have predicted, given my hereditarian priors, but still possible under a hereditarian theory, given motivational factors. Those can be seen in the gender gaps which are quite large. There is no logical reason to dismiss a difference in IQ(Black-White) because groups do the same while accepting an equality in IQ(between boys and girls) when one group scores higher. The Africans would not be the only group which managed to “punch above their weight” in terms of IQ and education, the East Asians in America have long done so. Asians’ IQs of 5 points above Whites are not enough to explain such an extreme over-representation in the Ivy League, tiger mothering provides the rest of the puzzle. That case, however, is not so counter-intuitive, you expect Asians to perform better than Whites and they do.

But, when reading Chisala’s series, I initially overlooked a much more powerful point he made in his favor: what happens when the African group is broken down into subgroups? This shows that many African groups perform much better than White Britons, indeed, some even out-perform the Chinese. Accepting that Tiger-mothering could be so much more effective among these groups as to be able to out-perform a group that is notorious for also practicing it and beats them in IQ by ~13 points(if Black African immigrants to the UK are assumed to be selected to a point where their children have IQs of about 92, while the Chinese are assumed to be 105) is much harder. I’m uncomfortable with “super-selection” theories which have the IQs of the children of African immigrants being higher than 95.(They have to regress to a mean, remember?) African immigrants are undoubtedly well-educated, but it is also undoubtable that Africa is not a meritocratic continent. Some were merely lucky to be the right person’s cousin, or the pretty girl the dictator decided to take into exile after the coup.

A lot of people get confused about regression to the mean, some believe it’s a multi-generational curse that should pull the descendants of smart blacks all the way TO the mean,(rather than merely in its direction, for one generation only) others than it should not apply if the migrating individuals are part of some vaguely defined meritocratic class. The meritocratic class theory could be theoretically possible. Suppose there’s a policy of only allowing White South Africans to immigrate and taking a representative sample among them. It would be wrong to say “these South African immigrants have IQs of 100, they should regress toward the mean of their country of 75.”(75 is made up, not sure of the actual mean.) But if, rather than taking White South Africans, you took any Black South African with an IQ over 85, you’d expect some regression to the mean. Why?

What causes regression to the mean? Suppose you have two identical twins, one with an IQ of 131, the other 121. If you make a cut off point and take a group of people with only IQs above 130, you’d take only the smart twin. His offspring should be expected to regress to below the cut-off point, as his genetic potential is not any higher than his brother’s, he just got lucky, either in random developments in his brain or in his luck on the day of the test.

The White South Africans in the first example were not selected by IQ. They were not on average any smarter than their siblings. But if you took any Black South African with an IQ over 85, you’d have many people smarter than their siblings, who just got lucky, and who’d be expected to regress to the mean in the second generation.

Here it is important to note that of selection of a class of people, essentially, men and their siblings, and selection by IQ, men who leave their duller siblings behind, both would negatively impact the IQ of the second generation. The more brothers and nephews of smart people you have coming, the more you aren’t selecting for IQ, whereas the more you select for IQ, the more you expect regression to the mean to lower the next generation’s IQ. If you had a group like White South Africans, a group of Africans where the group average is 15 points higher, who haven’t intermarried with duller Africans for generations, well, does that sound like a likely situation in any African country? You’d notice it.

In the study of IQ-like CAT scores, Africans have a standard deviation that is lower than Whites on all the subtests, which raises a question about the existence of high-scoring subgroups, and lends credence to the notion that motivation or tiger-mothering is the reason for the gap. However, it could also indicate that IQ-like tests may have a real bias, perhaps linguistic, which the Africans are able to make up in between the time they took the CAT(at age 11) and then took their GCSEs. Note that in the CAT tests Africans did worse in verbal than mathematical reasoning, a reversal of the pattern seen among Blacks in America. Either way, the tests alone should not lead one to conclude outright that the high-scoring sub-groups of Africans are truly less intelligent than Whites in Britain: real world accomplishment should always be a better measure of intelligence than IQ when they conflict. If one group has greater genotypic potential for height but was held down to equality with another due to insufficient nutrition, it would be silly to say they were “really” taller. Of course, if it turns out that Africans flounder after they leave school, or if the cause is tiger-mothering and the Africans stop and their achievements decline, it will be another matter. Note that the African over-performance on the GCSE’s is a recent phenomenon, so the outcomes of the current class of parents, many of them poor immigrants, should not reflect on their children’s ability.

Africans in Scrabble

Chisala has made a persuasive case about African over-performance in the game of Scrabble. The basic point is simple, competing in world scrabble competitions requires a high degree of ability and with IQs averaging 70, there simply should not be enough Africans able to do it.

Differences in ability are famously most salient at the highest level. It is not hard to wander around and find an athletic girl to outrun a fat, out of shape boy, but at the very high level, 125 men have broken the 10-second barrier, no women are in this exclusive club. 1/6th of Black Americans are smarter than the average White, but of the American winners of Nobel Prizes for Chemistry, Physics, and Physiology or Medicine, none have been Black.(No non-American Blacks have won either.) In science, IQ test results predict that the third world countries should make barely any contribution to world science, which is exactly what is observed.

Cartogram sized by scientific papers published

Yet in scrabble Africans manage to play, and win, in both English language and Francaphone Scrabble competitions. The only way for this to be possible while preserving African’s IQ level at 70, reflecting actual intelligence, is for a huge proportion of the population who are at that level to chose to engage in scrabble. This in a continent full of poverty and violence. I conclude based on this that the estimate of a 70 IQ reflecting Africans true abilities, the position of Richard Lynn and J Philippe Rushton, which I thought was reasonable given the state of the place, is now untenable. Rather, I believe it is an environmental deficiency which reduces their scores, possibly due to a lack of familiarity with testing, but not their underlying intelligence, which is at a higher level, maybe as much as 85. This is not a position entirely alien to hereditarianism, as most hereditarians regard the Flynn Effect to have be partially or wholly non-biological.

And even then, settling on a score of 85 might explain it, but would it explain it best? Or would a true genotypic IQ of 100 better explain a country which is very poor producing world champions at a game which requires a very high IQ level? What would your estimate be for, say, the very poor Armenia? Its PISA scores are low. Its economy is so bad Armenians illegally immigrate to Turkey. Yet it produces chess champions. This certainly doesn’t prove Africa’s true IQ is higher than 85, mind you, but it does point in that direction.

Conclusion

Sibling pairs of who are the children of one or two parents of mixed heritage differ in a small amount by their proportion of ancestry. Some mixed race African Americans, in the same family, with the same environment, who believe themselves to be equally Black or White, are slightly more African and some are slightly more European. This will provide a method of essentially proving or disproving the hereditarian hypothesis by measuring the IQs and genetic ancestries of thousands of these sibling pairs, and I expect it is reasonable that someone will do this in the next fifteen years.

If someone did do this, put the papers in my hand, told me they proved the gap was either almost all genetic or almost all environmental, and asked me to bet on it, even odds, I’d still say it’s mostly genetic. But I would not go much higher than 1:2 odds. There is still a lot of evidence in support of the hereditarian position, and in the comments sections of Chisala’s articles some have simply dumped this evidence there, ignoring the points Chisala has made. But I think this is wrong-headed. If 95% of evidence supports a position and 5% opposes it, should we conclude it’s probably true? Not necessarily. Restate it and see why: 95% of the evidence is consistent with the position, 5% contradicts it. There are many things that could cause X, but if your explanation for X relies on Y never being there and it’s there, then your theory is in trouble. Similarly, if it could be proven that the IQs of the children of African immigrants in Britain are 100, and this could be proven not to be caused by selection, the hereditarian hypothesis would be falsified. That African IQs are no higher than 85 in America, Haiti, Jamaica, Brazil, and Africa itself does not matter: if the theory is that they are genetically bound below 85 and there is a place where they smash the ceiling to bits, the theory is wrong. This hasn’t been proven yet, but evidence is enough to cause me to question my priors on this issue. What would constitute proof? Chisala believes his evidence decisively disproves the hereditarian hypothesis, and has accused his critics of “moving the goalposts” on this, and so I’ll state what I would consider definitive proof. If, on IQ tests, it could be definitively proven that Black African IQs are at or close to European levels, and this could be shown for a sample of Africans which could be proven to be non-selected, by analyzing education levels or the method of immigrant selection in some other way, I would take it to be a falsification of the hereditarian hypothesis. A good potential test is provided by one of America’s stupidest immigration programs: the diversity visa lottery. Each year about 10% of the Liberian population enrolls, which greatly limits the extent to which they could be selected. A study on these immigrants, if is found that their children have mean IQs close to the White mean, would falsify the hereditarian hypothesis in my mind.