Eyferth the Cat

by Dr Swaggins

Earlier today a friend of mine, the venerable Eagleshigh, posted an article about the black/White IQ gap which aggravated me to the point of actually writing a blog post to do away with that unkillable cat we politely call the Eyferth Study once and for all.

To give a little context for the uninitiated, the Eyferth study compared the average IQs of German children fathered by Allied soldiers during WWII. The results- trumpeted around the world, of course- were that the half black children scored an average of 96.5, a statistically insignificant difference from the White children’s score of 97.2. About half a century later, nobody has replicated this study’s results, and hereditarians have refuted its conclusion so thoroughly that Eyferth’s credibility should appear to have been run through a meat grinder. Yet the liberal news rags repeatedly celebrate the study’s apparent conclusion the world over, as if the scientific method simply doesn’t require replication of an experiment’s results anymore as long as the results are adequately pleasant. Put simply, Eyferth is the cat the keeps coming back.

Fortunately, with cats who keep coming back, there are multiple ways to skin them.

I’ll do away with the first of its nine lives by repeating Jensen’s observation in his 1998 book The g Factor: The Science of Mental Ability that the black GIs involved had all been selected for intelligence beforehand; those blacks with the lowest IQs, the bottom 30% of their population in this respect, were rejected from the military.

A hereditarian can use this datum alone to do some thorough skinning. Since the overwhelming majority of American GIs at the time were draftees, they should make a pretty good approximation of the black population as a whole in terms of IQ; I’d be willing to wager that the blacks called up for the draft back then averaged an IQ of 85 since draftees were chosen for their age and sex alone (neither of which affect IQ enough to skew our mean significantly). If you send the bottom 30% back home, you’ve truncated the left flank of your normal curve clean off! I used the trapezoidal method of integration to approximate the average IQ of those blacks remaining after the bottom 30% were sent home and came up with a result of 92.

Hilariously, if these men were to beget children by women with an average IQ of 100, the expected average IQ of the offspring would be around 96, or perhaps a hair higher due to maternal IQ being a slightly better predictor of a child’s IQ than paternal IQ. 96.5, maybe? Taking paternal IQ estimates into account, the Eyferth results were as consistent with the hereditarian model as with the egalitarian model- perhaps even more so!

I wish I could say that the study was only inaccurate to that extent. Unfortunately, I have more work to do. As Rushton and Jensen pointed out in their 2005 review article Thirty Years of Research on Race Differences in Cognitive Ability, 20-25% of the fathers of the Eyferth study were not even black Americans, but rather French North Africans. Since Algeria constitutes most of that territory (and is statistically representative of it as I’ll outline later) I’ll discuss the possibility of Algerians making up 20% or so of the Eyferth fathers. Richard Lynn found the Algerian average IQ at 83, but some of that relative deficit is due to Algeria’s high consanguinity rate. Still more of it may be attributed to Algeria’s mixed record as it relates to nutrition; UNICEF reports that in 1995, 92% of Algerian children had adequate dietary iodine, but by 2005 that figure had dropped to 61%. The 83 average IQ figure was taken in 2002, temporally closer to the 61% iodine adequacy figure than the 92% figure. Also, I’ll justify using Algerians as representative of French North Africa as a whole by pointing out that the other two countries have roughly the same consanguinity rates, that Lynn found Tunisia and Morocco’s average IQs to be 83 and 84 respectively, and that the UNICEF report I linked above found Morocco’s iodine deficiency rate to be within two percentage points of Algeria’s. Furthermore, all three countries are nearly entirely composed of Arabized Berbers. For these reasons, all three countries most likely have about the same genetic IQs.

We have little clue as to what effect the presence of these Algerians, Tunisians, and/or Moroccans would have specifically, although we certainly know that tossing in Berber caucasoids with the “black” sample constitutes an instance of flagrant intellectual dishonesty. I’ll posit that since first cousin marriages drop the offspring’s IQ by up to 10 points on average, we can estimate that those 22-25% inbred Algerians are around 10 points duller than their peers. Assuming that the inbred and outbred Algerians together have an average IQ consistent with Lynn’s figure of 83, algebra would have us believe that the inbred Algerians would score about 75 on average and their peers about 85; I’ll round up to 90 under the assumption that North African nutrition prior to the 1940s wasn’t stellar, because that would affect their IQ but not the IQ of their children.

How many of these French North African soldiers were inbred? To what extent would outbreeding correct this deficit? I doubt we have any way of knowing. Regardless, having 20% or more of the “black” fathers actually be caucasoids with genetic IQs we can at best estimate should’ve been enough to have the study thrown out in the first place because a fifth or more of the data had literally nothing to do with the hypothesis.

I’ve killed Eyferth twice now. For my third refutation I will point out that the girls in the White sample were for whatever reason duller than the White average, at a sample average of 93. This is obviously a sampling error; it couldn’t be attributed to a high frequency of low IQ White fathers because the mentally deficient Whites had also been truncated, and furthermore because the White male children had normal IQs. It’s worth noting that if the White daughters weren’t missing half a standard deviation’s worth of IQ points, the White children would’ve been four points brighter on average, which is consistent with my first point.

My fourth point is also very indicting: the study occurred when all of these individuals were children, and heritability of IQ actually raises over time. If there were a difference in intelligence in spite of the myriad of problems I mentioned above, it likely wouldn’t even show up by that point, as the average age of the children being tested was 10. What little gap they allowed to exist within the design of the study was obfuscated by taking the data before they were ready.

My fifth point will be that the IQ of the parents was never tested. We could begin to interpret the data concerning North African IQ or the bizarre loss of 7 IQ points seen in the daughters if we had the necessary data, but we don’t.

My sixth will be that we have literally no idea how, specifically, these children were conceived. If they were conceived by prostitution or rape, then assortative mating most likely did not play a hand in the conception. In this case, the average IQ of the mothers could be 100 or even 107 if we’re to believe Lynn’s figure for German IQ. If the child’s IQ is about halfway between the mother and father’s then the fathers’ IQ would be expected to be about as far beneath the children’s IQ as the mothers’ were above them, which would be consistent with my 92 estimate; if the children were generally conceived consensually then we would see this effect less often due to assortative mating. Points 5 and 6 boil down to lacking any knowledge of a myriad of possible tertiary variables to account for. It’s just poor experimental design.

My seventh refutation will be to point out that these results have never been successfully repeated. A crucial step of the scientific process is to have somebody else use the same methods you did to produce the same result so that we know it wasn’t a fluke. Data are not typically considered valid until they’ve been corroborated, and the fact that egalitarians are latching onto a severely flawed study in spite of its uncorroborated results shows an extreme level of bias on their part.

My eighth refutation will be to point out that among African fathers, the sample specifically uses African American fathers, who genetically are roughly 20% White on average. It is possible that if the children were 50% black instead of 40% black (or less thanks to the North Africans in the sample), a slightly greater difference would be observed. Lynn gives 67 for the average of all African countries, and 80 for African Americans with no European admixture. I could certainly imagine better nutrition  and a lower parasite load tacking on 13 IQ points. And I could imagine Africans with a fifth or more White genetics having closed a quarter of the IQ gap. The closest populations I can find to the theoretical model of a purely black population living with good nutrition and so on are South African blacks (adequate dietary iodine and 77 average IQ) and British blacks (average IQ heavily bottlenecked to 92). It is possible that had the fathers been fully black, a greater difference would have been observed.

All in all we have a population of Africans likely bred to higher intelligence via White admixture, with its bottom 30% excised, with some caucasoids thrown in, bred again with White mothers to cut the remaining IQ difference in half, and then tested before heritability takes its full effect, against a nonrepresentative sample of Whites. On the whole, the fathers should have had about 36% caucasoid DNA and the mothers should have had nearly 100%, with the children being 68% genetically caucasoid. What stupidity they could have inherited from their 32% African heritage was truncated away, obfuscated by being tested before it should have appeared, and compared against a slightly dull population of Whites.

Perhaps I could drum up another refutation if I had a mind to. Jensen claims that the higher IQ could have been due to heterosis, but the mixed race IQ data sets I’ve seen show no evidence that hybrid vigor is common in such situations.

That said, I don’t see the point in trying any more than I already have. Egalitarianism is the absurd belief that over 100,000 years, in wildly varying environments, and during massive introgression events, the brain will not change at all, in spite of being an inconceivably complex, highly energy expensive, and superlatively important organ. It violates everything we know about evolution and it’s been refuted to oblivion and beyond, but after all this time you still have academics like Flynn desperately clinging to mutilated data in order to support it.

Bad ideas are often case studies in bias, because the only reason anybody believes them is that they want to. That’s why they refuse to go away- that is, until a bunch of people turn around, open their eyes, and say those magic words:

“Galileo was right all along.”