Understanding IQ; The Radioactive Data of the Minnesota Transracial Adoption Study

I review an interesting study

This is the second article in a series diving into the nature of IQ and what they may mean. You can read the first here.

---

If you are a Very Serious researcher who happens on a field of study that makes these ridiculous claims: A) IQ scores predict lifelong outcomes for populations more than any other metric in the social sciences, and b) There are lasting differences between the scores of racial groups on these tests, your first stop might be to falsify it. And one way to falsify that claim is to figure out whether these racial differences in IQ are due to genetic factors or environmental factors.

Of course, it is entirely possible that the differences we note in these tests are due to environmental factors that merely correlate with racial background. After all, correlation isn’t causation. For example, if most Black people are poor, and low IQ is really just evidence of poverty, then Black people having low IQs would be irrelevant to their Blackness. Importantly, Whites of similar backgrounds would have the same IQ score. Therefore, Blacks would have low IQ scores because they are poor, not because they are black.

The best way to prove this would be to reveal low IQ to have a strong relationship with socio-economic background. One way to do that would be to look at the IQ of young Black kids adopted by White parents and compare their IQ to other members of the same racial background. If IQ scores were completely untethered from genetics, we should see these Black adopted kids eventually get IQ scores on par with White adopted kids.

It is an interesting research design, and is one many friends have told me to present if I truly believed the claim of IQ studies. Well, it turns out that this exact study has been done. And the results were not quite clear.

I first read this study about two years ago, and the results blew my mind. The most shocking thing, I’d say, was the fact that I was only just reading about it. I like to think myself a well-informed fellow, but this study threw a wrecking ball at that assumption. If this study were real, it should be one of the first things young Africans learned at university — even if for just the pleasure of refuting it. But I asked a friend studying psychology at a premier University in my country, and they were clueless about it. It isn’t like this knowledge was under some shady embargo — you could literally just google it and find a link on Google Research. Yet.

The Study

The Minnesota Transracial Adoption Study was designed by Sandra Scarr and Richard A. Weinberg. Scarr and Weinberg didn’t come to the study agnostic about the cause of differences in IQ. Both researchers were environmentalists who believed that most of these differences could be explained by environmental factors, not genetic ones. These were serious researchers too; Scarr had a degree from Harvard and became one of the pioneering researchers of developmental psychology and behavioral genetics, and Weinberg spent 40 years at the University of Minnesota, eventually directing the Institute of Child Development, one of the top-ranked programs for child psychology in the United States. So these were serious researchers, asking serious questions in a field that was seriously under-studied. And this was in 1975.

In any case, since Scarr and Weinberg were environmentalists, they had no incentive to “cook” the results of the study. They really wanted to falsify the assumptions and prove that environmental factors better explained IQ scores than any other factor. If you aren’t Nigerian, you may wonder why I have to make this explicit (after all, it seems obvious that this study, referenced perhaps thousands of times across several IQ studies, wouldn’t be a giant racial hoax) — however, we must note that Nigerians are uniquely drawn to conspiracy theories, and you would be surprised at the number of people who would dismiss these studies by claiming it is just another racist propaganda by Whites. Even in the comments of this Substack, you may find one such fellow. It bears re-stating that this isn’t what happened here. These researchers were desperately searching for evidence that IQ has no racial element. They had every incentive to cook the study in our favor. That is why the results are even more shocking.

The study itself was simple. The researchers first of all recruited 101 White middle-class families from Minnesota. These families had an average IQ of around 120, so these were highly educated middle-class families and provided the sort of “perfect” environment one would expect to foster higher than average IQs. The parents also enrolled their own biological kids in the study, so researchers had a baseline to compare eventual IQ scores to. There were 101 biological kids of these adoptive parents, 21 Black kids with two Black biological parents, 55 kids with one Black parent and one White parent, 16 adopted kids with White biological parents, and 12 adopted kids with Asian or Indigenous parents.

The kids were all adopted as young children — the earliest around the one-year mark, and the latest around the 4-year mark. So their life with their adoptive parents is all they know. The researchers planned to test the IQ of these kids at 7 (which was around the earliest you could administer these tests) and then compare it with their IQ at 18.

They were making two bets with this. The first is that the kids, by age 7, would have acclimatized enough to the environment of their adoptive parents that their IQ would reflect this environment, instead of the IQ of their biological parents. The second bet was that even if this didn’t happen, there would be enough difference between their IQs at age 7 and age 18 to show the genetic argument as nonsense. After all, if the IQs of these transplanted kids do not reflect the average scores of their ethnic population, we can safely argue that changing environments work, and there is nothing “rigid” about IQ scores.

Well, it appears that is where the argument runs into a pickle.

The first disappointment came with the first IQ test. The Black kids, at age seven, tested at an average of 97. This was initially encouraging news, as the Black mean is around 85. That means the environments of these kids had at least increased their IQ by 12 points on average. It isn’t a smoking gun, but it isn’t nothing. However, that encouragement dies with the tested mean of other groups; the White adoptees tested at 117 at the same age, and the mixed adoptees tested at 109. Even despite a uniform change in environment, the IQ differences persisted.

Perhaps the environmental change hadn’t yet kicked in enough for the Black kids, we might say. So the researchers waited with bated breath until the kids clocked 17 and they took the IQ test again. If the researchers were right, there should be little or no IQ differences between racial groups at this point. In fact, if they were right, it would be just as likely for a Black kid to attain a score above 100 as a White kid to achieve the same score. In short, we would expect the racial differences in IQ to disappear completely.

The results were not so clear cut. Instead of the average IQ score going even further up, it went down. At 17, the average score for Black kids was 89, which is more in line with the general Black mean. The scores for White kids fell by almost 15 points, and the scores for mixed-race kids fell by almost 10 points. But one thing remained constant: the Black kids scored the lowest on both rounds of IQ scores. And the White kids scored the highest. This is a very inconvenient result for all the reasons you have imagined.

Spinning the Truth

I don’t want to put any (more) spin on the results of the study. You can come to your own conclusions about what it proves. I have only written on the results. However, I will invite you to go on a journey of the study’s Wikipedia article with me to see just how careful one has to be when reading things online.

Before the article even dives into the nitty-gritty of it all, the writers input a caveat to help us judge the results of the study.

The very second paragraph of the article is this gem:

The aim of the study was to determine the contribution of environmental and genetic factors to the average underperformance of black children on IQ tests as compared to white children. The initial study was published in 1976[1] by Sandra Scarr and Richard A. Weinberg. A follow-up study was published in 1992[2] by Richard Weinberg, Sandra Scarr and Irwin D. Waldman. Another related study investigating social adjustment in a subsample of the adopted black children was published in 1996.[3] The 1992 follow-up study found that “social environment maintains a dominant role in determining the average IQ level of black and interracial children and that both social and genetic variables contribute to individual variations among them.”

I have read surely close to a hundred Wikipedia articles about scientific studies, and I rarely come across one where a study is being introduced, and refuted in the same paragraph. The fact that a verbatim sentence from the lesser known 1992 study found itself into the first two paragraphs is surely not common. But it gets worse, as far as this Wikipedia article is concerned. The introduction is rounded up by these two interesting paragraphs:

In 1994, researchers such as Levin[5] and Lynn[6] argued that these findings supported the view that genetics is a determinant of average differences in IQ test performance between races, while other researchers, including Weinberg, Scarr and Waldman, argued that the findings aligned with environmental explanations, noting that the IQ scores of the Black children were slightly higher than the national average.[7][8]

Subsequent developments in genetics research have led to a scholarly consensus that the hypothesis of Levin and Lynn is false. The idea that there are genetically determined differences in intelligence between racial groups is now considered discredited by mainstream science.

The final submission carries an air of finality; do not believe your eyes, poor pleb, smarter men than you have looked at this matter extensively, and they have found the truth. However, this isn’t true at all. In fact, the idea that there are genetically determined differences in intelligence is now more credible than ever. Here is an article that looks at multiple anonymous surveys by psychiatrists on the question, and the evidence is shocking. This idea has definitely not been discredited by mainstream science, as many researchers clearly believe in some variation of it.

Additionally, the first link to the source of this claim (that genetics plays little or no roles in intelligence differences amongst groups) is this very bad opinion article in the guardian. The thesis of the article is that there cannot be genetic racial differences in IQ since racial differences aren’t old enough. Here is how they put it;

Given that so many genes, operating in different parts of the brain, contribute in some way to intelligence, it is hardly surprising that there is scant evidence of cognitive advance, at least over the last 100,000 years.

But this isn’t true, and new research published in Nature (one of the most prestigious publications on these matters) explicitly proves it is wrong. There is significant evidence that genetic intelligence has evolved in the last 10,000 years! Not even 100,000 years. The last ten thousand.

The second authority is this somewhat balanced article from Vox that carries this strange sentence:

First, even when adoption produces substantial gains in the average IQs of adopted children, the magnitudes of the individual gains are better predicted by the IQs of the children’s biological parents than by the relative quality of the adoptive environment. This is true but irrelevant: It is merely evidence that IQ is partly heritable, which no one disputes. That effect (one more time) has no implications for understanding group differences.

I leave you to judge it. The rest of the authorities are opinion pieces referencing the same circle of moral argument; the question is unethical, if you look at the evidence this way it doesn’t prove what you think it proves, and it is very difficult to answer anyway, so let’s just leave it. I do not intend to.

It would be bad enough if this Wiki article had this position and stuck to it, but, as you know, Wikipedia editing boards are a cauldron of fire, and you never always get what you want. In the body of the article, this paragraph emerges out of nowhere:

In a 1998 article, Scarr wrote, “The test performance of the Black/Black adoptees [in the study] was not different from that of ordinary Black children reared by their own families in the same area of the country. My colleagues and I reported the data accurately and as fully as possible, and then tried to make the results palatable to environmentally committed colleagues. In retrospect, this was a mistake. The results of the transracial adoption study can be used to support either a genetic difference hypothesis or an environmental difference one (because the children have visible African ancestry). We should have been agnostic on the conclusions . . .”

So we get the same article arguing that the results of the study do not prove what they look like they prove — and that everyone who is anyone agrees — and in the same article you find the lead researcher arguing that his framing of the results was out of courtesy to his colleagues who believed something completely different. It is incoherent, but it is precisely what you would expect from the radioactive material that is the issue of racial IQ.

Even if you were a researcher who seriously believed that differences were due mostly to genetic factors, would there be any gain to making that opinion known? No. It is fuzzy enough that you would be branded racist, and would immediately lose your job and livelihood. This is no joke, and has happened to several researchers in the past. That is why psychiatrists hardly even respond to surveys asking the question. Who wants that headache? Who wants to become the next Watson, Pesta, Carl, Rushton and Jensen?

The more you read this, the easier you may understand why this lecture is coming from Elewa’s Substack, not a university lecturer. There are thousands of people whose life work is to ensure results from this study — or studies that threaten the blank slate orthodoxy — are coddled and caveated and misinterpreted to death. Anyone without background knowledge would leave that article unsure of what he read, despite the evidence piling up enormously on one side of the floor. Can we imagine what the people who beat this article to death would have said if the study found something entirely different? If it had found that all of the kids scored around the same score, which would be consistent with the uniformity of their environment? Would there be any ambiguity whatsoever about what the study actually says? Would they care about cofounding factors that made this possible? But since we find the opposite in these results, it is the duty of these activists to footnote it to death.

That is the current state of public discussions of IQ in general and racial differences in IQ in particular. This, of course, makes research horrible for unlearned rubes like me. Every “general” article concerning the matter carries the spin of the ideological leanings of the writer. A writer who leans to the extreme left may look at the wealth of data on the matter and still insist that the studies do not prove what they seem to prove. A writer who’s on the extreme right may look at it and argue that they prove more than they do.

The only way to know what the paper really says is to look at the original arguments themselves and come to your own conclusions. Given that, let us look at the paper that Scarr and Weinberg wrote in response to Michael Levin, a fellow who argued that the study proved the assumptions of hereditarians.

The Reply

In his article, Comment on the Minnesota Transracial Adoption study, Michael Levin says this:

If environmental differences account for all the between-race difference in IQ test performance, one would expect the difference in IQ between the B/B cohort and the birth children to be no larger than the deviation of the birth children from the white genetic mean.

I could be wrong, but this represents a fairly reasonable conclusion given the results of the study. He goes further to say this:

Environmentalists would presumably expect preadoption experience to be a significant source of variance, but Weinberg et al. (1992) report that, so far as could be determined given its limited range of variation, time of adoption explained only 7% to 17% of the variance in adoptee test scores.

Scarr and Weinberg, in this article, argue that isn’t the case. And they present some evidence for their position. The first is the Asian and Indian adoptees, whom Levin does not discuss. Genetic theories of racial IQ predict that Asian children should score at or above Whites. In this study they scored between the interracial and Black groups. The authors argue this is difficult to reconcile with the genetic hypothesis, and that Levin’s silence on it is not a small omission.

The second point is that the groups were not, in fact, raised in identical environments. Black adoptees had lower placement quality, more placement disruptions, later ages at placement, and less time in the adoptive home than White adoptees. The rank ordering of these early experience variables mirrors the rank ordering of IQ outcomes almost exactly. The authors argue that Levin’s dismissal of early adoptive experiences as explaining only seven to seventeen percent of IQ variance is therefore misleading — because the groups also differed on those experiences in racially patterned ways, meaning the confound is more serious than a single variance figure suggests.

The third point follows from the second. When they statistically controlled for early adoptive experience differences, racial group differences in IQ shrank substantially and became non-significant — a trend rather than a finding.

The fourth is a logical objection to Levin’s inference structure. The authors argue that demonstrating the insufficiency of environmental explanations does not confirm a genetic one. Other environmental variables — racism, social classification as a minority despite White adoptive parents, stigma in schools and communities — were never measured. Their absence from the model is not evidence of their absence from the children’s lives.

The fifth is a technical point about heritability. Levin draws on twin study evidence of high within-group IQ heritability to support the genetic hypothesis for between-group differences. The authors argue this is a category error — within-group heritability says nothing about what causes differences between groups. The two questions are statistically and logically independent.

In the reply, the authors (this article was written in 1994, 4 years before Scarr admitted that the evidence could also prove the genetic case) conclude that their data favor a predominantly environmental explanation, while acknowledging that the study’s design cannot definitively settle the question in either direction.

However, these points by Scarr and Weinberg seem very defective to me. Consider the first point, for instance. The “mixed race” cohort of the group was not homogenous, as they included Asians and Indians in the same group. This is problematic because IQ studies predict higher performance for East Asians, not South Asians. So if the study had a higher share of Indian subjects, then their scores would be perfectly in line with what we may expect.

The second point merely begs the question, as we do not know if the confounding factors in the background of the Black subjects were due to characteristics correlated with their genetic background. For instance, agencies may have placed children they assessed as having lower potential into lower quality homes, or the children’s own early behavioral profiles may have influenced placement quality. If genetic factors influence the very early experiences that are then used to explain away genetic factors, the argument is circular.

The third point is that controlling for these early experiences within the cohort shrinks the gap. First of all, that seems like a post-hoc rationalization, as the reason the study looked at children adopted so early was to eliminate the possibilities of early experience confounding the experiment. After all, one of the points of environmentalists throughout the paper is that since the IQ scores taken at 7 were higher (by a few points) than we would expect of their ethnic background, it proves that environments do indeed change IQ. It seems incoherent to make that argument, and then later insist that the differences in IQ scores obtained at 17 were due to confounding variables from 10 years ago.

In the fourth point they argue that their experiment does not take into account all environmental factors, and there are other factors that were not controlled for and would still explain the IQ gap. This is plausible, but there is no reason to believe it is the case here. If Scarr and Weinberg believe that these confounding factors have the explanatory power they suspect, they have to prove it in black and white, not just offer it as a plausible reason for why their data did not show what they expected it to show.

The fifth is a more interesting point, as it gets to the heart of the “academic consensus”. No psychiatrist seriously debates the existence of a Black-White IQ gap. None would also seriously debate the idea that IQ is heritable. The only question, it seems, is whether this heritability explains IQ gaps between groups. You would think it does, but there is an interesting analogy that environmentalists use to prove their point.

Imagine you had two similar genetically related seeds and planted them in good soil, and you planted another group of similar seeds with similar genetic relationship in bad soil. The two seeds in good soil would be genetically related, and would grow strong and healthy. The seeds in the bad soil, on the other hand, will grow up to be stunted and weak. The two groups of seeds would be genetically similar to the other, but that genetic difference would do little to explain their stunted growth.

This analogy makes some sense, but I disagree with it for the simple reason that it begs the question. Where did this insanely rich soil that elevates genetic differences so wildly come from? Out of thin air? Most psychiatrists now argue that good education is a better predictor of life outcomes than IQ scores. But where does good education come from? Is it bestowed by God upon nations unequally? Importantly, If we assume this to be the case, we are taking an implicitly totally environmental view about IQ, and that isn’t supported by any evidence. In fact, an only environmental position is the one least supported by the evidence of this study. It is by far the worse spin on this topic, as you can see that it is deeply incoherent with the rest of the study.

Next on Understanding IQ

One more way to really understand what causes the effect here might be to look at early intervention studies — that is, studies that expose kids to intensive educational programs in childhood. If these interventions eliminate IQ gaps, we could still say these gaps are totally environmental, and that we just need to expose young kids to these intensive educational programs in early childhood.

The good news is that this study has also been done. In my next article, I will be looking at what that study found and what it proves. The results were interesting, to say the least.

Like, Subscribe, Share

If you have read this far, you are probably one of two people. The first is the kind of person who agrees with everything I have written and wants to share it with a friend who needs to hear it. Please do that. The second is the kind of person who thinks I have gotten something badly wrong and wants to tell me about it in the comments. Also please do that — I am genuinely interested in being corrected, as long as you are correcting me with evidence and not with feelings. Both of you should subscribe.