Jason Malloy on P.Z. Myers

For normal IQ tests, the stereotype threat possibility definitely doesn’t explain why scores would differ between 100% African children raised by white families and mixed-race children raised by white families, especially when the mixed-race children believe themselves to be 100% African. I don’t think the stereotypes are any different—as I said before, 50% African people such as Obama are still just thought of as being “black”—but even if the stereotypes are different, people aren’t able to be affected by stereotypes of a heritage that they don’t know they have. So any explanation for the IQ difference that ascribes it to inaccurate testing will have to explain why stereotype threats for mental chronometry, and whatever’s causing the difference for normal IQ tests in the adoption study, would both affect the results in exactly the same way.

True, but the original study doesn't seem to separate the mixed children into 2 groups of 1."White mother & black father" and 2."Black mother & white father", it simply says "One white and one black biological parent". So let's consider the possibility that prenatal exposures was a significant contributor to the measured differences. Then mixed children with a white mother and black father would score an IQ more similar to children with 2 white biological parents, and mixed children with a black mother and white father would score an IQ more similar to children with 2 black biological parents. But since the study has lumped these scores into the same group, the average score they present would wound up somewhere in the middle.

What Plindboe said. Also with regard to the bolded statement you make, how was this tested? You made an earlier statement that the IQ differences held up for children who did thought they had both an African-American mother and father, while one of them actually was Caucasian. But how was this tested? According to the wiki article, there were 12 kids in this group and the difference did not differ significantly from the rest of the interracial group. I need to see the numbers here, but I'm very skeptical of how much this statement says based on a sample of 12 (on a larger note, I'm often very skeptical about significance testing without having insight in means and confidence intervals of the different tested groups and everybody should be).

In order for prenatal factors to explain why a group that half had African fathers and Caucasian mothers, and half had Caucasian fathers and African mothers, had an average IQ exactly halfway between a group of 100% Africans and a group of 100% Caucasians, prenatal environment would have to be the only environmental factor contributing to the IQ difference. If any other environmental factors were playing a role, the mixed-race kids who believed themselves to have two African parents would not have scored exactly halfway between the other groups like this. If that’s really the reason for it, then if you look at the individual IQ numbers for each mixed-race child, there should be a very visible division in IQ between those with white mothers (higher) and those with African mothers (lower).

This would be a significant enough element of the results that I suspect the Wikipedia article would have mentioned it if it had been the case, but I guess we ought to look at the original paper to make sure. Tomk80 offered to try and get the paper through his library, and if he isn’t able to, I suspect that at least Jason Malloy has it. I can try asking him for it if Tomk80 can’t get it.

In the meantime, though, here’s something I’d like to ask about this: what prenatal factors are you aware of that could vary with race regardless of socio-economic status? Is there any specific one that you think could be contributing to these results?

You’ll be able to see more of the actual numbers if you manage to find the paper through your library. Do you think you’ll be able to? If not, I can ask Jason Malloy whether he has it.

In order for prenatal factors to explain why a group that half had African fathers and Caucasian mothers, and half had Caucasian fathers and African mothers, had an average IQ exactly halfway between a group of 100% Africans and a group of 100% Caucasians, prenatal environment would have to be the only environmental factor contributing to the IQ difference. If any other environmental factors were playing a role, the mixed-race kids who believed themselves to have two African parents would not have scored exactly halfway between the other groups like this. If that’s really the reason for it, then if you look at the individual IQ numbers for each mixed-race child, there should be a very visible division in IQ between those with white mothers (higher) and those with African mothers (lower).

In the meantime, though, here’s something I’d like to ask about this: what prenatal factors are you aware of that could vary with race regardless of socio-economic status? Is there any specific one that you think could be contributing to these results?

Wait. At least if I go from what wikipedia states, the statement is that the average score of the 12 mixed children who think they have two black parents is not significantly different from the other interracial children. This does not mean that they scored exactly halfway in between and confidence intervals of these 12 might have easily overlapped with the other two groups, in which case they wouldn't differ significantly from the other groups either.

Why regardless of socio-economic status? Even if they corrected for SES these factors might still be of influence. General factors would be diet and exposure to contaminants. Next to this, I don't know from the information I have so far what age the children would be adopted. Next, adoption experiences might also be of influence. Whether a child looks more African-American or more Caucasian could be of further influence. Again, I'll know more about correction etc once I read the paper and a number of commentaries (on both methodology and results) I requested.

I had none in mind. Just thought that you dismissed Tomk80's suggestion a bit prematurely.

I posted the actual IQ numbers in post #18. It’s a difference of 10-14 points between the 100% African children and the mixed-race children, and a difference of around 5-9 points between the mixed-race children and the white children. I guess I shouldn’t have said it’s exactly half, since the first difference is a little larger. But what I’m saying is that if any environmental factors other than prenatal conditions were playing a role in this, you’d expect the difference between the white children and the mixed-race children to be larger than between the mixed-race children and the 100% African children, not the other way around. I don’t think there’s any question that the difference in stereotypes between white people and 50% African people is greater than the difference between 50% African people and 100% African people, as is the case with just about any other cultural factor.

A lot of the research about this involving people raised by their biological parents controls for SES. I linked to this chart earlier, which shows that when you control for SES, the IQ difference shrinks but doesn’t disappear. So assuming that the cause of the IQ difference is the same in each of these studies, it isn’t caused by SES, or anything that strongly correlates with SES.

According to the Gene Expression article, the adoption study controlled for all of the most obvious environmental factors that can vary between races—parenting, income, nutrition (after adoption) and neighborhood. Regarding nutritional factors before adoption, the APA analysis states that diet doesn’t affect IQ except in cases of severe malnutrition, so it would have to be exposure to contaminants if it’s anything in that category. What contaminants might African-Americans be exposed to more than Caucasians, even when socio-economic conditions are equal?

As I said before, I don’t think it’s possible to be 100% certain that the difference in IQ is genetic, at least not without more research being necessary. So I’m not really dismissing his suggestion as being impossible. What I’m saying is just that as long as you can’t come up with a specific environmental factor which could be causing this, whereas a genetic pattern which could be causing it has been identified, an explanation that involves genetic factors is the more parsimonious of the two.

I wasn’t as knowledgeable about this topic when we debated about it around a year ago, so although my point was the same, I wasn’t able to support it as well. As you might remember, my main problem with an 100% environmental explanation isn’t that it’s not possible, but that it involves an argument from ignorance—you’re basically just saying “environment did it”, even though there isn’t any more specific explanation, which means attributing it to some unknown environmental factor is a hypothesis which can’t be tested. You agreed with this, but said there was the same problem with attributing the difference to genetics (or saying “genetics did it”), and at the time I didn’t really have a response to that. But during the time since then, I’ve learned that as the genes which influence intelligence are being discovered, they’re being found to be distributed in the ways one would expect to be the case if the genetic hypothesis were true. So while still not enough to be considered “proof”, this shows that the genetic hypothesis is testable, and that tests are supporting it.

Does that make sense?

The problem here is that SES can only be approximated. We can only use surrogate measures for these many of them don't cover the load completely. Income, achieved education, job status have all been used as indicators of SES, but none of them are perfect. For example, education achieved only covers part of the "education experience", namely how long someone has studied and which degrees he got. But if you take that measure and also input the quality of education, you explain more of the variation in IQ between races than if you only use the education achieved.

Which is one of my main gripes here. I'm very skeptical on what SES actually means when you use it. It depends on the study, the exact measures used. You never really input "SES". You input income, or years of education followed, or poverty index.

I didn't have any specific in mind, but I'm sure if one had the time one could go on a statistics hunt for environmental factors and come up with several suggestions. If one finds higher numbers of blacks compared to whites who drink, smoke, take drugs etc. during pregnancy, or more often give birth to babies with lower birth weights, or contain a higher amount of various harmful elements due to poor nutrition (I read somewhere that black people have greater amounts of lead in their bodies than white people, which can have negative consequences for future intelligence). I think several of these factors I mentioned, and probably several I haven't, have been shown to have harmful effects on development and intelligence, and I can't imagine it would be insurmountable to find some of these factors at higher frequencies in some groups compared to others.

Yep. I'd like to see more of this evidence. Is it that link in #22? If it is, it doesn't appear to work for the moment.