The Science of Everything Podcast - Episode 82: Racial Differences in Intelligence

Episode Date: March 11, 2017

A discussion of the evidence and methodological issues pertaining to explaining the causes of racial differences in intelligence, with a focus on evaluating the hereditarian hypothesis that such differences are largely genetic. Recommended pre-listening is Episode 81: Intelligence Part 2.

You're listening to The Science of Everything podcast episode 82, Intelligence Part 3. I'm your host, James Fodall. So in this third and final component of the intelligence series of episodes, we're going to look at the controversial issue of group differences in intelligence. And we're going to focus on differences between black and white intelligence in the United States, because that's where most of the research has been done. I'll say a little bit about gender differences in intelligence as well, but mostly we'll look at racial differences, as I said, with a focus on the black-white differences in the United States.

Though I think it's likely that many of the general methodological and conceptual issues that we touch upon will be applicable in the analysis of racial differences elsewhere, both for other races and in other countries. Recommended pre-listing for this episode is the previous episode, on Intelligence, Episode 81, Intelligence Part 2. So basically, in this episode, we're going to critically analyze some of the literature and public debate about group differences in intelligence. And particularly what I want to do is focus on some of the claims that have been made

and the evidence that have been provided for them, because what I want to really focus on is the types of evidence that could support claims made either way, what the evidence actually does look like, but also how we could interpret the evidence and what sort of evidence we would need in order to substantiate some of the claims that have been made. So let's begin, and I think a good place to start is by talking about the 1994 book called The Bell Curve, Intelligence and Class Structure in American Life, by Richard Harnstein and Charles Murray. So this was a very controversial book when it was published, resulted in a lot of discussion and public outcry and media coverage and so on, many responses.

Now, I've not read this book, although I've looked at some parts of it, but the primary reason to start with this is I think it sets the scene for the sort of controversy that exists. So the basic purpose of this book was to write about the role that intelligence plays in American society, its consequences for political and social outcomes. Some of their primary claims that they made were, first of all, that intelligence, exists and can be accurately measured across racial and language national boundaries. And that, I think, is fairly well supported by the mainstream psychological evidence. We've discussed that in the first episode of the series on intelligence.

So some of the criticism directed towards the book was directed against that primary starting point, for example, by Stephen Jay Gould, who wrote a criticism of this book, criticized the very concept of IQ and the ability to accurately mention. measure it across cultural or national boundaries. I don't think those criticisms are very well founded, and we've discussed those previously. So we won't talk any more about those here. A second key point that they raised is that intelligence is one of the most important factors relating to or determining economic social success in the United States in the present. And I guess by application in other developed countries as well, of course, they focused on the US.

And they presented a lot of data showing how intelligence is correlated. with employment success and social success, like rates of divorce or having illegitimate children, being incarcerated, high school graduation rates, things like that. So the idea is that intelligence is a crucial determining factor for these various outcomes, which we've also talked a bit about before. I think that's fairly well established, although what is established more specifically is a correlation between intelligence and these social outcomes. What's less clear is what the causal factors are

in determining this correlation. That is, why does the correlation exist? And that's something we'll come back to a little bit later. But I do think that one of the key problems with this book is that it makes too readily causal claims on the basis of correlation evidence. That is the fact that two variables are correlated doesn't mean that one is causing the other. And we discussed this in, I think, episode 79 about basic concepts and statistics.

A third major claim that they made is largely heritable. So that's the, you know, 50-60-ish percent that I mentioned before. That's the mainstream figure, although I think there are problems with that, as I discussed in the previous episode on intelligence. So that's episode 81, when I talked about the methods that are used to determine that, adoption studies and twin studies, and how that heritability figure is only meaningful in the context of a given amount of environmental variability. And the twin and adoption studies, I think, are going to have underestimated the amount of environmental variability because they primarily

select from relatively prosperous families, because those are the families who either adopt children in the adoption studies, or in the case of twin studies, are more likely to be selected into the studies and be retained in the data sets. Much harder to get poorer families to be enrolled in those long-term studies. So for these reasons, I think these studies underestimate environmental variance, and therefore overestimate the heritability of intelligence. So the first three claims, I think, are relatively well-evanced, subject to some of the caveats that I mentioned.

The next two claims are sort of the more controversial ones. and will lead us into the rest of the discussion. The last two claims that they made in this book, the bell curve, is that no one's really been able to manipulate IQ to any significant extent by changing environmental factors except for a child adoption, and that the evidence therefore points to the fact that these sorts of manipulations aren't very successful,

these sort of social programs designed to increase IQ. And that relates to then the fifth and final point, which is that the US society and media and government, so on, is largely in denial of this fact, that you can't really increase intelligence through environmental manipulation, and that we need to base public policy decisions on the basis of the fact that largely intelligence is heritable and not easily, or perhaps at all changeable through environmental alterations. So the basic idea that this book, at least, seemed to present to many people, and it seems to me that this was sort of

the basic message of the book, even though there were other nuances in the argument, was that A large proportion of the variance in social outcomes in the US is explained by intelligence, and intelligence in turn is largely genetically determined. It's largely heritable, and you can't change it or can't change it very much through social programs. So many of the social programs designed to improve, say, college graduation rates or reduce poverty or reduce incarceration rates and that sort of stuff, improve employment outcomes. Many of these are based on the false premise that we can.

and change circumstances or other behaviours to improve outcomes. They're arguing, no, it's actually largely IQ differences that are causing these outcomes, and in fact you can't change those or can't change those very easily. So we need to base government policy or social policies on these facts. So that's the, I think, the core idea promoted by the Belko book. They also talked a bit about race, which is how it links into our discussion here. And they didn't explicitly argue that there's a strong genetic component to racial differences in intelligence in the US, but they did mention it, and that was a key point of controversy surrounding their book. Now, other authors have been much more explicit about genetic differences leading to racial differences in IQ.

For example, two psychologists Jay Rushden and Arthur J. Ronson have published a lot of articles over the years discussing this, and there have been some others as well who, seem to be pretty fixated on this idea that the group differences in intelligence in the US are largely genetically determined. So let's look at what those group differences are. The most research has been done, as I said, on the difference between white and black IQ test scores in the US. And the best figures that I can get are that in the US at present, the average IQ score for African Americans is something around 85, whereas for whites, it's around. 100. There's some evidence that Asian and Jewish Americans score higher than 100, maybe even up to 115 for Jewish Americans, although the sample selection issues are a bit of an issue there.

Latinos also seem to score a bit lower somewhere between blacks and whites, but here I'm mainly focus on the difference between blacks and whites, which seems to be about one standard deviation, so about 15 points. Some of this data is a bit old, so there's some evidence that that's been reducing in recent years, so I tend to think it is. Maybe it's only a 10-point gap these days. It really does depend on which data you're picking and it's kind of controversial, but the basic finding that African Americans score lower in IQ test on average than white Americans, I think is fairly well established, maybe a difference somewhere between 10 to 15 points, which is quite a difference. Okay, so if we take that as our empirical fact, then the question

is, what is the cause of this? And most psychologists and intelligence researchers are generally prefer to stay away from this question and just say we don't really know what the cause is, and I think in part that's justified because there's not really that much good literature or high-quality research that's been done to answer this question as to why this difference exists. And so we can't really make very strong conclusions. There is, however, some evidence, and that's what I want to discuss in this episode. So what are the factors that could determine this or could cause this difference in IQ between blacks and whites? Now, there are two main paradigms or approaches to this.

One that Rushden and Jensen and some others propound is what they call the hereditarian model, which is that roughly 50%, or maybe even a bit more, of the difference in group intelligence achievement, so the black-white achievement gap, is attributable to genetic differences between racial groups. So that's the hereditary model. 50% or more is genetic. Whereas the culture-only model, which has been defended by other researchers, says that essentially there's no or only a trivial genetic contribution to group differences in intelligence, and almost all of it or all of it is due to environmental differences. So the key point, one key point here is that both approaches do admit to some environmental factor, at least some environmental factors in the group differences.

the question is whether there is any genetic component or not to the group differences. Now, another important point to note is that in the previous two episodes, I did talk about the heritability of intelligence and noted that a large proportion, probably about half maybe more, of the variation in intelligence is genetically determined. Now, but doesn't that automatically sort of show that these group differences must be largely genetic? Doesn't that automatically prove the hereditarian hypothesis? No, it doesn't. And even Russian and Jensen admit this, that explaining why individuals within a population differ in intelligence is not the same as explaining why two different populations, say blacks and whites, differ on average in their intelligences.

These are actually different claims. One is a claim about individuals in a population. The other is a claim about two population averages. So the fact that individuals within a population vary in intelligence largely because of genetic factors, or at least substantially because of genetic factors, does not. imply it, and you can't just jump from that to say that two different population means are different because of genetic differences. So, no, it doesn't follow from the individual evidence. In fact, this is a separate issue which needs to be investigated separately. And so we need to look at the specific evidence for this claim that the average group differences in blacks versus whites is due to genetic differences between those groups. Another point that needs to be raised here is that there are genetic differences between blacks and whites if you look for them.

So the genetics of race is itself a very controversial issue, and we'll talk a little bit more about this later on. For the moment, all that needs to be established is that when you take a sample of people who are considered African-Americans and take a sample of people who are considered white, and you look at various genetic loci or allilic proportions in the different populations, yes, you do find differences. the overall difference is actually not very large compared to, say, the difference between humans and chimpanzees. So all different humans from all different racial backgrounds are actually very closely related to each other compared to many other species that have more internal variability within them, genetic variability.

So humans are actually quite a homogeneous bunch. But nevertheless, you can find some differences between different races. So it's at least possible that some of these allelic differences, some of the differences in, the relative proportion of certain alleles and the different populations could account for differences in intelligence across those populations, blacks and whites. So it's at least conceivable that the hereditarian hypothesis could be true. You can't just reject it on a priori grounds. However, the fact that it's conceivable that it's true doesn't mean that it is actually true.

So we have to look at the specific evidence for the hypothesis, and that's what I'm going to talk a bit more about. So what evidence have people like Russian Indianson put forward for their hereditarian hypothesis? That is, that a significant fraction, probably more than half, of the intelligence difference between blacks and whites is due to genetic differences between those two groups. Well, one piece of evidence they point to was consistent IQ gaps across the globe. So it's not only in America that there's a difference between blacks and whites and whites in IQ, but also if you look at sub-Saharan Africa, obviously, mostly black people live there, people from an African heritage. That's generally how they define black people in their analyses. And you can quibble with that, but let's accept it for the moment.

Score significantly lower on IQ tests compared to, say, white Europeans or North Americans. The results there are a bit controversial. Rushton and Jensen say that the average is about 70 in sub-Saharan Africa. A matter analysis, a more recent manner analysis that I looked at showed was closer to 80, which is consistent with the Flynn effect, actually, because we know that IQ scores are increasing, especially in places like Sub-Saharan Africa. Whereas in East Asia, there's maybe a bit of an increase in IQ compared to Europe or North America. So, you know, 106 is a figure that they give as to the average. So Russian Jensen particularly claimed that there's sort of a hierarchy in a sense of cognitive ability

across races, with people of African heritage at the lowest, with European heritage in the middle, and East Asian heritage at the top. And so they claim that it's consistent, not just within the US, but also across when you look at other countries, which supports the hereditarian hypothesis. Now, there's one immediate problem here, because, as I said, the way they define their racial groupings is basically where most of your ancestors come from. If most of them came from Europe, then it's European heritage or Caucasian, white, whatever. Most of them came from East Asia, then that's Asian heritage. If most of them came from Africa, sub-Saharan Africa, then that's African heritage or black. Now, one immediate problem with that, if you know anything about human evolution, is that if you go far enough back, everyone's ancestors came from Africa.

That's the out of Africa theory, that somewhere around 60,000,000 years ago, the first modern homo sapiens migrated out of Africa through the Middle East and into Europe and Asia, and later on to the Americas. So what does it mean to say most of the people's ancestors came from Africa? when everyone's ancestors came from Africa. So obviously there's a time span that we're going to have to look at here. We're talking about most of people's ancestors, somewhere between 1,000 to 10,000 years ago or something like that, not like 50,000 years ago or 60,000 years ago

when everyone's ancestors came from Africa. So automatically we have to be focusing on ancestors of a particular time period. And this is consistent with some of the other evidence that they put forward in favor of their hereditary hypothesis that is, they think their hereditary hypothesis is supported by environmental differences between Europe and East Asia on the one hand and the African savanna on the other. So they think basically that when the first modern humans began to migrate out of Africa,

they moved into new environments that were sufficiently different from the savannah that they came from in Africa so as to promote the evolution of, on average, increased cognitive abilities. So I'll read you a couple of quotes to you to illustrate this idea. So this is from Russian and Jensen. Evolutionary selection pressures were different in the hot savannah where Africans lived, then in the cold northern regions the Europeans experienced all the even colder Arctic regions of East Asians.

These ecological differences affected not only morphology but also behaviour. It has been proposed that the farther north of populations migrated out of Africa, the more they encountered the cognitively demanding problems of gathering and storing food, gaining shelter, making clothes and raising children successfully during prolonged winters. End quote. Now, what they're claiming here, though, basically, is that selective pressures were different in the colder northern regions to, as they say, the hot savannah, and that this promoted the evolutionary selection of greater cognitive ability. Now, it should be noted before we analyse this, that they don't necessarily have to provide an explanation for why genetic differences exist if they do between rates.

They could just say that, well, they do, and therefore they account for the IQ differences without saying why they exist. But at the same time, if we're to analyze this hypothesis about genetic differences, we would have to make it plausible. There needs to be at least some reason as to why this would emerge. Remember that around 60,000,000 years ago, all human beings have a common ancestor. So in that relatively short evolutionary time span, we're supposed to say that there have been very substantial evolutionary pressures to yield one standard of the or more differences between different groups of humans in terms of intelligence. And that's quite a big difference that's supposed to evolve in quite a short period of time. So in order to account for that, there must be some pressure, a selection pressure, that could be operative.

And so presumably we'd have to at least have a plausible mechanism as to what that could be. Otherwise, we're going to be scratching our heads a bit and thinking, well, how could this possibly have happened, biologically speaking? And at least to me, their argument does not seem very plausible at all. One issue is that we don't know very much at all about what the original environment of evolutionary adaptation was in Africa. They talk about the hot savannah. We don't really know what it was like living in those regions at that time, exactly what the climate was, or whether humans mostly lived on the savannah or more in the forest, or what exactly. So there's a great deal of uncertainty there.

They also mentioned the cognitively demanding problems of gathering and storing food, gaining shelter, etc., in the cold, which is certainly true, those would be cognitively demanding to figure out how to do that stuff. But it's not at all clear that there would be any more or less cognitively demanding than the same tasks in the hot savannah or in the jungle, or wherever else people might have lived in Africa or in the Middle East or anywhere humans have lived. So they don't really have any evidence or present any evidence that there's an objectively higher cognitively demanding level of cognitively demandingness in living in one environment compared to the other. And there would have to be in order to explain this selective pressure, which, remember, supposed to have occurred only over a few tens of thousands of years, which is not very long, evolutionarily speaking.

So it must be a significant pressure to result in that difference, not just a very slight pressure, and yet they present no evidence of that, and it's not at all plausible to me. Account they've given really shows that we would expect such a difference. There's another problem with a more general problem appealing to genetic differences as the basis for group differences in IQ, which is that although there are genetic differences on average between, say, African-American and white American populations, the differences that exists,

or the diversity and variability in genetic patterns that exists within the human species as a whole, does not very well map to racial groups as we would categorize them, especially in the West. So, as I mentioned before, this issue of genetically categorizing or clustering humans on racial lines is very controversial. But some of the reading that I've done, and maybe I'll do another episode on this in the future, but the bottom line is that some studies have found genetic clusters along more or less racial lines,

and Russian and Jensen cite this as basically, oh, they say it divides into Europeans, sub-Saharan Africans, South Asians and Pacific Islanders, Native Americans and Aboriginal Australians in the New Guinea. And basically those are sort of the classical races. But the big problem with some of these studies is that basically what they do is they just sample a handful of different groups in these regions. So, you know, like three or four from sub-Saharan Africa, a few from Europe, a few from Native American populations, and then do a cluster analysis on this. Now, when you pick populations in that way, of course you're going to find the clusters around the traditional racial groups,

because you've picked your samples in accordance with those geographic areas. That's not a valid way of doing this sort of study. A better way would be to randomly sample from across the human population, not just including some sort of relatively isolated groups from each of the continents, but including samples drawn from all sorts of intermediate groups between, you know, like in the Middle East, for example, or people from mixed racial backgrounds and other people that are harder to classify along traditional racial lines. if we took that full sample of human genetic diversity and then did a clustering analysis,

and that hasn't really been done because obviously it's more expensive. Some studies do better than others. It's not at all clear that you get the traditional clustering analysis. To understand that argument, if I were to take genetic samples from one group of people, so one sort of narrowly defined population, so they take samples from these. If you were to do one of those from each continent and do a cluster analysis, what do you think you're going to get? You're going to get one cluster for each continent. That's not at all surprising. The more interesting question is if you do a sample across all of the human genetic diversity, including all the intermediate groups, more outbred populations, etc., do you get the same clustering?

And I think the answer is no. In particular, one study that I consulted, which looked at single nucleotide polymorphism, so that's a one amino acid difference in genetic sequences, found that there were larger differences within sub-Saharan African, within the group of sub-Saharan Africans than between Africans and Eurasians, so Europeans and Asians. That is, there is more genetic diversity within the group of people

who now live in sub-Saharan Africa compared to the difference between Asian and European populations and the average sub-Saharan African group. And that is not at all surprising, evolutionarily speaking, because the groups that migrated out of Africa and settled in Europe and Asia were only subsets from the wider sub-Saharan African population.

So essentially, those populations have only had 60 to 70,000 to diverge from what they started from, whereas sub-Saharan African populations have had much longer, because human homo sapiens have existed in sub-Saharan Africa for maybe 200,000 years. Depends on where you draw the cut-off. But the basic point is there's been more time for divergence of genetic differences between populations of different groups within sub-Saharan Africa than there has been for, say, European populations versus the Hubs-Saharan African population from which they came. So you really expect, evolutionarily speaking, that there's more diversity within sub-Saharan Africa than there is between sub-Saharan Africans and these other groups. So if you're going to do a fair and full clustering of analysis of the genetic variance within human populations, you would sort of have two races, really.

I think you'd get like sub-Saharan Africans and like everyone else, which would probably all be more similar to each other in terms of the variants than the great swath of variance that exists within the sub-Saharan African populations. So the long and the short of it is that this idea of clumping all people who have sub-Saharan African heritage together and calling them Africans or black and then arguing that they have genetically lower intelligence, say, than Europeans,

doesn't make a whole lot of sense because we expect a huge amount of variation within, relatively speaking, a huge amount of variation within the sub-Saharan African population. So, you know, there'd be some groups who would have higher. So if there is genetic difference between human racial groups that leads to intelligence differences, we would expect some groups within the sub-Saharan African munch

to be higher in intelligence, some to be lower. There's no reason that overall they would show this much lower level, and in fact we'd expect a large amount of variability within that. So I think there's a problem with the way that the racial clusters have been determined. Basically, they've been determined along social lines because black and white is something that is socially relevant in the US and in other cultures for the past few hundred years. But it's not, I think, the natural clustering that you would come to if you did a proper cluster analysis of human genetic variation.

So for these sorts of reasons, I think that the hereditarian hypothesis is on shaking ground to start with it. It's not clear what the mechanisms could be to produce this large amount of variation between this larger difference in intelligence between sub-Saharan Africans and Europeans and Asians over such a short evolution period of time. And it's also not clear that the clusters that we're talking about, that the racial groups even map too very well to actual clusters of genetic commonality. So those two things don't disprove the hereditary hypothesis, but I think it rests on pretty implausible assumptions genetically and evolutionarily speaking

when you consider these factors. But nevertheless, we still do need to look at some of the more specific evidence to see whether the hereditarian hypotheses can hold up. And as I mentioned, apart from the global comparison of IQ scores, so the blacks in America and blacks in sub-Saharan Africa, both scoring lower than Europeans and white Americans, they also, Russian and Jensen also cite worldwide differences between blacks, whites, and East Asians

in a variety of other factors as well, social factors and biological properties. And they have this table in one of their papers, which is, I have to say, the most bizarre table I've ever seen, that I can recall having seen it, at any rate, in an actual scientific paper. Because they list all of these traits and then compare them between blacks, whites and East Asian.

So they've got differences in IQ scores, which we mentioned before, cranial capacity differences, which we'll talk a bit more about later, differences in egg-twinning rates, two-eggitwinning, so that is dysargodic twins between the different races. They say that it's higher in blacks compared to whites and higher in whites compared to East Asians. They think that this is consistent with their hypothesis about genetic differences between the races. But all of the other entries in the table are not numbers.

They're just like higher, intermediate and lower, or shortest and longest and so on. They have these bizarre things like personality. They've got aggressiveness, cautiousness, impulsivity, self-concept and sociability. And most of it's like, you know, aggressiveness is higher for blacks, intermediate for whites, and lower for East Asians. And likewise for sociability, higher for blacks, intermediate for whites, and lower for East Asians. And they have other things like reproduction and social organization. So, sex characteristics, larger in blacks, intermediate in whites, lower in East Asians.

Same for hormone levels. Higher in blacks, intermediate whites, lower in these stations. And then they have things like law-abidingness, mental health and marital stability, lower in blacks, intermediate, and why it's higher in these stations. So this is so bizarre, because I don't know where they get this data from, because they don't actually present any numbers for most of these things. It's just higher, intermediate, and lower, or shorter and longer, and I'm wondering where these come from, and also how they were determined controlling for other factors, because obviously if you're

looking at people in different countries, for example, then things like law abiding this. It's going to be hard to measure that in a consistent way. across countries. So how is it exactly to determine that blacks have on average lower law-abidingness than whites, internationally speaking, because it says worldwide average differences? And probably the silliest thing in the table is cultural achievements. So that's labeled as low for blacks, high for whites, and high-free Asians. Now I think it's meaningful to take a selection of, say, the African-American population and take a sample from the white American population, give people IQ tests, and then look at the averages for

those two groups and compare them. So I think you could do that. But what on earth does it mean to take a sample of the white population of America and the black population, compare the cultural achievements of the people in that sample, and then rank them as high and low? It's completely baffling to me and what that could possibly mean. If that's not what they're talking about, maybe they're talking about cultural achievements as in like black civilizations compared to white civilizations, then I don't see how that's in any way relevant to current day differences in IQ scores or how you can somehow attribute the cultural achievements of a particular civilisation to a given race, what that could even possibly mean. So things like this don't

inspire me with a great deal of confidence as to the sort of research methods that they're that Russian Jensen are pursuing here. This table is really quite strange to me, where the data came from and how they control for other factors and how they're interpreting. this data. It's also important to note that there's no real way that they can distinguish from these data alone between the hereditarian and the environmental-only hypothesis. So, for example, let's suppose that aggressiveness is higher in blacks compared to whites. I don't know if that's true, but suppose it is. Now, is that genetically determined, or is it the result of social differences, cultural differences, that is the different average

socioeconomic position of blacks in the US compared to whites, for example. It seems quite plausible that it could be the latter. And likewise for pretty much all of these other things, rates of sexually transmitted diseases, intercourse frequencies, law-abidingness, mental health, they could all be the result of social differences, environmental differences, between the racial groups. So I don't know how they... It's not clear to me that this, even if taken at face value, the results of this table supports their contention. because both the hereditarian and the environment-only hypothesis could account for the results of this table.

So beyond those sort of worldwide differences and these race-behavior matrix, that they're called, that's the table that I mentioned, what other evidence they present in favor of the hereditarian hypothesis? Well, they do appeal to the classic twin studies and adoption studies, which I discussed a little bit in the previous episode in terms of the methodology behind them. Now, there actually aren't very many studies in total

that look at racial differences on the basis of, say, cross-racial adoption. That's ultimately what we'd like, right? Black children raised by white parents, black children raised by black parents, white children raised by black parents, and so on. And then we could look at the differences between how much of the differences in IQ scores are determined by the race of the parents and how much is determined by the race of the children. Under the hereditarian hypothesis, you'd expect a big difference between, you'd expect a big difference on the basis of the race of the children because their genetic

potential would be shaping their average IQ achievement. Whereas under the environment-only hypothesis, you'd expect the race of the parents essentially to do all the work. That is, there'd be no difference between black and white children raised by white parents, or black parents, for that matter, of the same socioeconomic standing. As I said, there have not been very many studies, either adoption or twin studies, that have followed that sort of methodology. You know, you count the number on one hand, maybe two hands at most, and many of them have been fairly small sample size, and there are questions about the representativeness of the samples and so on.

One of the issues is always, when you're placing, say in adoption studies, when you're placing children to be adopted, black children and white children, is there a difference in the age of the white children when they're adopted versus a black children? If white parents have a preference for adopting white children over black children, then it may be easier to adopt out the white children. and therefore they may be adopted out at a younger age, therefore the black children may be adopted out at a later age, meaning that there's a longer time period in which they were experiencing the relatively, presumably, and I think this is correct, relatively deprived environments of an orphanage or foster parents or wherever they were before they were adopted out. So if there's an age difference between the children when they're adopted out, then there could be environmental differences for a significant proportion of the developmental period such that that could account for a difference in IQ later on. So that's one of the issues with these sorts of studies. You really have to try and control for all of the factors. Likewise, is there a difference in socioeconomic status

or education outcomes or other things of the parents who adopt black children versus white children or some difference there that is of the adoptive parents? You want to try and control for that as well. And many of these studies have these sorts of issues. Now there is one of these adoption studies that they cite, one main one that they cite in their favour, which specifically involved adoption of black children, white children, and I think children of mixed racial parentage by middle-class white families. And they found that basically the more white heritage the children had, the higher their IQ scores were. But this study has been criticized on the basis of some of the factors that I just mentioned, particularly that it's not clear if there were differences

in either the adoptive parents and the age of the children when adopted across the racial categories. So this would invalidate the results if there were differences other than than essentially racial differences between the children when they're adopted out and or between the environments of the adoptive parents. There are also a number of other of these racial adoption studies which are either not mentioned or severely downplayed by Russian and a Egyptian which don't support the hypothesis. For example, there was one study of children born to African-American servicemen who were serving in occupying Germany. after the Second World War. So born to German mothers and an African American, or also some

French-African soldiers, who were subsequently adopted by a white German, and no difference was found in terms of the IQ scores of the children on the basis of whether they had black parents or white parents. So I won't go through. There are a few other studies like this as well, adoption studies that have looked at transracial adoption. Basically, whether you want to believe in the hereditary hypothesis on the base of this evidence, it seems to me, it comes down to which studies do you want to pick and emphasize on which do you want to ignore? Because if you read papers by Russian and Jensen and other people like that, they emphasize the studies that show differences. If you read papers by other people, they emphasize the studies that don't show

differences. There seem to be more studies that don't show differences than that do, but there are a few that do. Overall, though, I think the evidence is just not persuasive, either way, really. That is, the adoption studies are just, there are too many factors that can't be properly controlled for, in particular age at adoption, whether the race of the children has been properly determined, because that's sometimes unclear, and whether there are differences in adoptive parents, as well as the factors that I mentioned before, that is in the previous episode, about the limitations of adoption studies as a whole. So overall, I just don't find this evidence very persuasive. It's just not very good quality. There aren't very many studies. Most

of them are pretty small. They're based on samples of convenience. And with all of the other problems that I mentioned, I don't think that it's possible to make any strong conclusions about cross-racial differences in achievement on the basis of these adoption studies. I think that on balance, the adoption studies show negative results, but there are a couple that show positive. I don't think they're very good, but the overall, I think, fair conclusion of this transracial adoption literature is just it doesn't say, it doesn't tell us very much. There are too many weaknesses and too many problems. And that is consistent with the mainstream psychological position, which is that we don't know what the cause of racial differences in intelligence is.

Certainly I don't think we know on the basis of adoption studies. Another type of study that Russian and Jensen look at is racial admixture. That is that if it's true that people with white European heritage have higher average IQs than people with more sub-Saharan African heritage, then we would expect people with mixed racial heritage to have an average IQ that's somewhere in between black and white averages. So the basic result from this literature seems to be that if you base the analysis on skin tone, that is how white is someone's skin versus how dark is it,

then you do find some correlation between skin tone, whiteness of skin tone and higher intelligence. However, that itself, the correlation doesn't seem to be that high, but the more important point is that that itself doesn't tell us anything about whether hereditary hypothesis is true, because, of course, people are treated differently on the basis of skin tone. In fact, that's the primary means people use to classify you in terms of race, at least in the West. So that's obviously not going to provide any way to distinguish between the genetic versus the environmental hypothesis. So a better study doesn't look at skin color, but looks at self-assessed degree of black ancestry. And that does not always correlate with skin color, because you can have someone who's half black and still be mostly white, or vice versa.

just depends on the luck of the genetic drawer, essentially. You can have two children from the same family. One looks mostly white. One looks mostly black, even though they have the same parents. That happens. So that's why there's a difference between self-assessed degree of black ancestry versus skin color. And when you look at these hypotheses, these types of studies, rather, it doesn't seem like there's any relationship between degree of black ancestry and intelligence scores.

And this particular finding also raises another significant problem with the hereditary hypothesis, at least in the US case, which is that almost no African-Americans today actually have 100% African-Arican ancestry. In fact, a large portion of African-Americans have mostly European ancestry, and nearly all of them have, nearly all African-Americans, have at least some degree of European ancestry. So there's a large variation within the African-American population, so the degree of European ancestry. And the converse also applies to the white American population, the large proportion. I've read someone like 20%, it might be more than that,

of people who consider themselves white in the US and look white, actually have black ancestors. The point is there's actually a lot more mixing between these populations than we might think. And so efforts to try to explain black and white differences in Ikea, or intelligence score tests on the basis of genetics are problematic because there's actually less genetic difference between those groups than you might think. That is, two people who could both be classified as black and both look black may have very different amounts of,

like we do have very different amounts of actual sub-Saharan African ancestry. It seems that the environmental hypothesis is much more discriminating in this respect, that is, it's able to pick out why people who, so look black, are likely to have worse scores rather than people who just actually have the ancestry. Because it doesn't seem socially that the ancestry actually matters. What matters is, you know, who your parents are, how much education they have, what type of neighbourhood you come from, what sort of opportunities you had at school, at preschool, and so on, how much time your parents had to care for you,

the resources they had available, discrimination, all these sorts of factors, which are based much more on what you look like rather than your actual ancestry. So it seems to me that the environmental hypothesis maps much better to the actual sort of social reality that in terms of factors we know affect intelligence scores than the hereditarian hypothesis because of this mixture of racial populations in the US. Another related line of evidence, which I think cuts against the hereditary hypothesis, is that if you have one black and one white parent, then the average IQ of mixed racial births in this sense would expect it to be the same, regardless of which parent is black, because you would on average have 50% percent.

black, 50% white. It doesn't matter whether the mom or the father is black or white. However, on the environmental hypothesis, it's plausible that the mother is much more important in the socialization of the child, in particular in terms of speaking to the children, for example, looking after them, ensuring that they engage in more enriching cognitive processes or activities, these sorts of things. Quite plausibly, the mother is more important than the father on average. So if that's true, and the environmental hypothesis is true, we would expect to see that, so 50-50, that is mixed racial births, who have a white mother, would, on average, have high IQ scores than those who have a black mother, but a white father. And at least according to one study that, I found

this is, in fact, what we observe. In fact, according to this, a nine IQ point difference between children of white mothers and black fathers compared to black mothers and white fathers. So that's impossible to account for on the hereditarian hypothesis, which would say, well, 50-50 in both case, therefore should have the same IQ scores, but makes much more sense on the environmental hypothesis. Another piece of evidence that might support the hereditary hypothesis are blood group analysis. That is, there are actually fairly large differences as far as things go in blood group markers between European populations and African populations. And you can still observe those in the US population today. So you would expect to see correlations between these

and IQ scores if, in fact, there was a substantial genetic component to IQ, to the racial differences in IQ. That is, we would expect to see that the more of the European blood group markers you had, then the high, on average, your IQ would be because the way genetic inheritance works, the more of these blood group, European blood group markers you had, on average, you would expect those people also to have more of the supposed high IQ genes, or whatever they are exactly, that European populations are supposed to have. And therefore, you should observe a correlation between European-like blood groups and higher IQ. But in fact, we don't observe this.

And even Russian and Jensen admit this, that there just see no evidence from the blood group studies. And basically, Russian and Jensen just say that, well, the genetic markers that we have from blood groups don't have large enough allele frequency differences between Africans and Europeans in order to tell any difference. That justification isn't really very persuasive because even if the differences in allele frequencies are only relatively small, you should still see some difference. You would have to say that the markers were completely worthless in determining racial heritage in order to justify there being no IQ difference between people with these markers. But in fact, they're not worthless. You can tell something probabilistically about where someone's come from, that is where their ancestry is, predominantly from, on the basis of blood group.

a population level. That is, you can tell if there's a certain proportion of certain blood markers in a population, this is an African population versus European population, you can make those determinations. Yet it doesn't seem that these markers are useful for predicting IQ, which seems to indicate that there's no support for the hereditarian hypothesis of IQ differences between the groups being associated with genetic differences between the groups. So, there are a few other pieces of evidence that I want to touch on before we finish out this episode. One is that Russian and Jensen argue that

environmental differences can't account for the differences, the large differences in outcomes between racial groups because adoption and twin studies show that shared environmental factors, like degree of parental education, for example, don't make any difference in adult IQ scores. That is the proportion of the variance of IQ that these shared environmental factors explain, goes down to zero by the time people reach adulthood and they've got one of these graphs in there in their paper. Now, I don't believe this result. I think it's very implausible on the face of it that's a parental education and the opportunity you've been given as a child and the number of words that your parents speak to you and how enriching your child environment is and so on, all of which we know make a big difference

to children IQ scores. It seems very unlikely to me that that somehow just doesn't make a difference to adult scores. How that could actually work socially, that is your early child. childhood and adolescence somehow just doesn't matter when it comes to adult IQ scores. I haven't seen explained anywhere. That just seems completely baffling to me. So I think we need pretty strong evidence in favour of this hypothesis to that there's no shared environment effect on adult IQ scores in order to take it seriously. But there just isn't good evidence for it. Basically the evidence is a lack of evidence.

There have been no adoption studies that have studied primarily adults over the age of 20. So we can't say anything on the basis of adoption studies about this adult effect. It does seem to be that the shared environment is less important for older children, say adolescent incentives for younger children. That seems plausible that as children grow older, factors outside of the family become more significant in their development, but that's not the same as saying that the shared environmental factors go down to zero when they're adults. And so there's just no evidence from those from adoption studies.

in terms of twin studies, the main evidence comes from the fact that, as I mentioned, the difference between dyszygotic and monosogotic twins that I talked about in the previous episode. The problem there is assuming that the shared environment for monosogotic and dysogotic twins is the same, which is demonstrably false. Identical twins have much similar shared environments than monazogotic twins. So I talked about this in the previous episode, so I don't want to rehash all of that. But I don't consider that evidence very persuasive. There's also the problem that I mentioned in the previous episode that environment and genetics aren't independent of each other, that if you have genetic predisposition for higher intelligence,

you're likely to pick enriching environments, which help you then to further increase your intelligence and practice those cognitive tasks and so on, and vice versa for people who might be genetically disposed to in the opposite direction. So they interact with each other all along the process from childhood through adolescent into adulthood, and I don't think you can meaningfully separate them

and the way that these studies try to do. So the fact that the shared environment coefficient goes to zero to me doesn't actually tell us very much about the importance of the shared environment, especially, as I mentioned, because of this crucial assumption between the difference in monosogotic and diagogic twins, which I think is false.

The final set of studies that have been put forward in favor of this hypothesis that shared environment doesn't matter for adults is identical twins raised apart. And there's a great paper that talks about this. Basically, there haven't been very many studies that look at this, Most of them aren't very good. In particular, a lot of these studies try to increase the sample size, because obviously there aren't very many identical twins who are raised apart. So they try to increase

the sample size by including raised apart, including as raised apart pairs of twins who actually aren't really raised apart at all. Maybe they're sort of partly raised apart. So like they shuffle between houses or they're raised in different houses but that are next door to each other, things like that. So they're not completely together, but they're not completely separate either. So you can't classify that as being raised apart. Their environments are obviously going to be highly correlated there. And when you actually look at, you can't do rigorous statistical analysis in this way, but when you pick case studies of certain pairs of twins who we have life history information about,

we know that these are identical twins and we know that they were raised very independently, like in two completely different households, one that has clearly a higher socioeconomic standing and more enriching environments and so on than the other, and there are a few case studies that we have of this. In every case, it's very clear that the twin in the better environment scores substantially higher in IQ tests, you know, by about 16 points in this particular set of cases that I'm looking at than the one in the impoverished environment. So that's not a rigorous statistical study because we're looking at case studies here, But I think basically the problem with the identical twins raised apart studies is that many of the studies have grouped together too many cases that are too different.

Many of the cases that twins are actually not really raised apart. They're raised in very similar environments that are connected to each other. What you really want to focus on are the identical twins who are raised completely separately in very different environments, which are clearly, one is clearly better than the other in terms of what we would expect to be conducive to development of intelligence. And in those cases, it seems very clear that there's a very substantial effect of the environment. This is also consistent with some of the adoption studies results that shows that when you adopt a child from low-soceconomic status into a middle-class family, you can increase their IQ by something like 15 or more points. That's measured as a child.

The effect as adults probably a bit lower than that, but nevertheless, it does clearly show that the immersion environments do matter. And there doesn't seem to be any good reason for me to think that there, effect goes to zero in adulthood. So, Russian and Jensen's claim that shared environment between whites and blacks couldn't make a difference in to adulthood, I don't think is very well established. Finally, one other piece of evidence that Russian and Jensen cite in favor of their argument is that basically Europeans and Eastations have larger brains and more neurons compared to African Americans, and that within races there's a correlation between brain size and intelligence.

and we did talk about that in the previous episode. Now, I'm not really sure about the quality of their data about cranial capacity here. I haven't looked into that great detail. So for the sake of argument, I'm just willing to accept that East Asians have larger brains than whites and whites have larger brains than blacks on average, although the differences aren't very large even according to their data.

But the problem with interpreting that in the way they do is that, as we mentioned at the beginning of the episode, you can't infer from the fact that within racial groups or within the population as a whole, brain size and intelligence are correlated, that therefore when you look at brain size differences between racial groups, that those will also be indicative of higher differences in intelligence. It's the same problem as arguing that within a race, intelligence is largely heritable, therefore cross-racial differences in intelligence scores must therefore be largely heritable as well, largely genetically determined.

It doesn't follow it because differences in averages between. different groups are a different question entirely to explaining variances within a group. So I don't, even if you take these cranial data and neuron counts at face value, I don't think they support their contention, or at least they provide at best a very weak evidence for their contention that whites have, the black white achievement gap in terms of IQ is largely due to genetic factors. What we'd need to establish really is that this difference in cranial capacity was truly predictive of IQ cross-racially. All in all, I think that the hereditarian hypothesis is deeply implausible.

That's my conclusion from reading this research. I don't think there's any really good evidence for it. At best, there's a few studies here and there which may hint at it, but I think there are many more studies of the various types that I mentioned that count against it. And then when you add that to its baseline implausibility and I think also the baseline high plausibility of environmental factors, that is we know that there are lots of environmental factors that do have a strong effect, demonstrating adoption studies in particular and twin studies. Those studies can't prove that all of the achievement gap is due to culture or environment,

but at least it shows that we know that there are cultural and environment factors that do play an important role, whereas we don't know that there are any genetic factors cross-racially that play an important role. So I think overall there's much more reason to think that there's no genetic component, or at least no substantial genetic component, like it could be a fraction of 1%, but no genetic component of any substantial nature that explains the difference in IQ score or test score achievement between blacks and whites in the US. And by extension, I think across the world, at least in large racial groups, may perhaps be different in the case of more specifically defined racial groups,

like Ashkenazi Jews, for example, who, at least according to some studies, have much higher average IQs even than whites. It seems at least more plausible that are narrowly defined genetic group like that, relatively defined genetic group like that, could have, there could be average genetic differences relating to intelligence than, say, African Americans, or sub-Saharan Africans, which is, you know, much larger group and much more heterogeneous. But even there, I'm skeptical. I tend to think it's probably mostly cultural or social differences that account for, say, high Jewish IQ.

But if there were any, I might think that it would lie there in some of those more genetically homogenous groups, say, as Ashkenazi Jews are compared to, say, sub-Saharan Africans, certainly. But that wasn't the primary focus of this episode. So overall, I think that we'll leave it there. So my conclusions are that there's no really good evidence for the hereditary hypothesis that racial differences in IQ are due to, genetic factors. I think that the evidence points that are being due to social and cultural factors that we've talked about in the previous two episodes that we know relate to effects, that have effects on IQ. And that overall, I think we can conclude from the literature that there's a lot we don't know about intelligence, particularly how it's improved and what interventions or social interventions

would work to help increase the intelligence of children from marginalized or low economic backgrounds, low-shosterous backgrounds. However, it does seem that intelligence is not this sort of fixed and immutable thing that sits in the brain somewhere, that it is a generic problem-solving and reasoning ability that is partly genetically determined, certainly, but also substantially is developed through practice. And that practice largely comes through exposure to the right environments and settings that allows you to practice in an appropriate way, which begins in very early childhood and increasingly compounds over the life as over the lifespan, with the genetic one's genetic predisposition then iteratively interacting with the environments that you

select and are selected for you by your family and so on and your school and your peers, etc. that allow you to practice those skills or not practice them and therefore get better or worse at that sort of generic, abstract reasoning ability. So that's my broad conclusions about sort of what intelligence is and how we should think about it. We should think about it as something that's mutable and that can be improved through practice, but not that's easily mutable, not that's something that you can just improve

by playing some internet game or something like that, like those puzzle games that you find to increase your intelligence. It's somewhere in between those extremes of easily mutable by playing a game, but completely immutable and genetically determined. And we still don't know,

there's still much we don't know about exactly how it works and what causes individual and group differences between them, but nevertheless, there's still a great deal we can learn by looking at the psychological literature with a critical eye focusing on the quality of the evidence presented and the research methods used. So I'll finish up there. Hopefully you enjoyed that episode and hopefully wasn't too controversial. I think my claims are fairly well based on the evidence, but if you disagree, feel free to send me an email at Fods12 at gmail.com. FODDS1.com. Also, we have a

Facebook page for the podcast, which you can find if you just type in the Science of everything podcast into Facebook. And if you give that a like, that'll give you information about when new episodes are being released or some other show announcements I occasionally put up there. Also helps to spread the message about the show. Otherwise, thanks a lot for listening, and I'll talk to you next time.