Race and Intelligence: Science's Last Taboo

[liberalbiorealist]

It also comes before your failure to explain why you hypothesis explains variance over and above proximal factors like SES, parental and personal education, differences in opportunities.

Not particularly to whore my own blog (but not particularly to avoid it either), I think you might take a look at this post of mine which addresses this issue. The bottom line here is that the children of blacks of some of the highest educational and economic levels are outperformed by the children of whites of some of the lowest such levels.

It's a really hard result to explain away by environmental accounts of IQ.

[Lazar]

It also comes before your failure to explain why you hypothesis explains variance over and above proximal factors like SES, parental and personal education, differences in opportunities.

Not particularly to whore my own blog (but not particularly to avoid it either), I think you might take a look at this post of mine which addresses this issue. The bottom line here is that the children of blacks of some of the highest educational and economic levels are outperformed by the children of whites of some of the lowest such levels.

It's a really hard result to explain away by environmental accounts of IQ.

um this was not a general comment as much as it was a direct challenge to enigma to actually put up some evidence for his nonsense theory.

It is true that controlling for some environmental does not reduce this gap to zero. However, this is NOT the only source of potential bias* which might account for these differences. See my post for other potential sources of bias.In any case it is equally the case that research suggests gaps in IQ have closed from 15 point down to single figures in recent years and that gaps in achievement tests have gone from 1.1 SD to .65 SD.

In addition, the research on herditability of IQ is advancing rapidly (see here). The old models of linear combinations of inheritance, shared environment, and unique environment have been found to be lacking. In essence it seems more likely that inheritance and environment and their effect on IQ is non-linear. Indeed, in the model I pointed to the results suggest that the herditability of IQ may differ wildly depending on SES. "The models suggest that in impoverished families, 60% of the variance in IQ is accounted for by the shared environment, and the contribution of genes is close to zero; in affluent families, the result is almost exactly the reverse." - from linked article. This makes the whole idea of assigning causes for IQ differences to inheritance alone extremely dubious.

All of this however, is hand waving over the biggest problem of all. That is that variance in IQ is far greater within groups than between groups. As many have said already, race is a socially constructed phenomena and individuals within 'racial classes' are NOT homogeneous. It seems to me that until research can show group differences predict differences in IQ AFTER controlling for the individual level (and can construct a meaningful definition of race) it should be treated as a statistical artifact ONLY. As yet I have failed to find any research that provides such data.

Indeed the whole process of testing group differences as it stands is highly dubious.

By Zuckerman, Marvin
American Psychologist. Vol 45(12), Dec 1990, 1297-1303.
Abstract
The scientific premises for looking for statistical differences between groups designated as races (on somewhat arbitrary grounds) are questionable. The explanation of such differences in strictly biological-evolutionary terms is even more dubious. Studies of temperament, basic personality traits, disorders (such as antisocial personality), and specific genetic markers show that there is much more variation within groups designated as races than between such groups.

[Delvo]

Why does the "no biologically inherited difference" side keep harping on the "more difference within races than between them" theme? It doesn't support the conclusion you keep bringing it up as supposed support for. We all know that the existence of differences that are not due to biological inheritance does not indicate the non-existence of differences that are, or in other words that there's no reason why the two things can't coexist. We all know that the fact that
x>y does not mean y=0. We all also know that practically nobody who says that there are biologically inherited differences in intelligence has claimed that it's the only factor that exists or even that it's bigger than all others combined, which is the ONLY kind of claim that would actually be countered by this within/between thing you keep trotting out.

Yes, I know it's true, but what's the point of badgering on and on about it? It's like responding to someone who says that properly working freezers keep their contents below the temperature at which ice would melt, by protesting "No, no, no! They use electricity!" Truth alone doesn't mean it actually has anything to do with anything.

[liberalbiorealist]

Let me try to go down the list of issues you raise, although I do rather wish you would focus your objections to a particular point instead of the scattershot you've offered up.

You say at the beginning,

It is true that controlling for some environmental does not reduce this gap to zero. However, this is NOT the only source of potential bias* which might account for these differences. See my post for other potential sources of bias.

Now it would have been nice had you stuck with that point, and attempted to explain yourself. In any case, while you characterize my original point about SAT scores for whites vs blacks at different economic and educational levels as only controlling for "some environmental", it hardly does justice to the force of point I was making (made more fully in the post on my blog I linked to). Let me state it most starkly here: the children of white parents who make only $10K-$20K per year outperform on the SAT the children of black parents who make over $70K a year. Likewise, the children of white parents who have only a HS education outperform on the SAT the children of black parents who have a graduate degree. The environmental advantages of the black children here aren't just some smallish factors. They are the very heart and soul of the environmental explanation as its usually set forth. Seriously, are we to think that a child whose parents have graduate degrees, and all the sophistication that implies has no important advantage over a child whose parents who, in this day and age, have only a HS diploma? And yet the children of these quite sophisticated blacks can't do as well as the children of these quite unsophisticated whites? What kind of sense does that make, if environment is the ultimate explanation of performance? And, with regard to the issue of bias, shouldn't the far wealthier or far better educated black parents be providing a much, much richer environment for their children than the disadvantaged white parents, which would set any biases in their favor, if there really are relevant biases at all? (Consider even that classic example of supposed bias from the SAT -- an example that gets used again and again most likely because no other example seems particularly convincing -- in which knowing the meaning of the word "regatta" is required for the correct answer to a question. Is it really more likely that a child of white parents who make only $10K-20K will know the answer to that question than the child of black parents making over $70K a year? Or more likely that the child of white parents who have only a HS diploma would know the answer than the child of black parents with a graduate degree? What's left of the claim of "bias" against blacks after one reflects on what these advantages might really mean for these black children?)

Many of the same points apply to the research you link to from Turkheimer. Even granting the point (which others dispute) that genes contribute much less to differences among low SES children than to high SES children, how, still, does one account for the very facts I just mentioned about black children of parents with high incomes or high levels of education? Shouldn't these children particularly be enabled to excel, not suffering any environmental influence to depress their scores uniformly as with low SES children? How can they possibly fail to outcompete, and easily, white children with the very disadvantages that are supposed to make such an enormous difference?

In any case, I think a good summary of some of the objections to Turkheimer's work might be found in a comment that just appeared on my blog, which quotes from a paper from Jensen and Rushton.

You end by talking about how race is just a social construct, etc. I have to say, I find that line of talk -- I won't call it reasoning -- impresses me as just pure obfuscation. Whites, for the relevant purposes, are those whose ancestors are almost entirely (let's say 90%+) from Europe. Blacks, for the relevant purposes, are those whose ancestors are dominantly from Africa (let's say 70%+ -- a bit more generous since blacks in America apparently have 20-25% white ancestry on average). Evolution works along lines of descent, and it's clear that, for tens of thousands of years, the group of humans who occupied Europe have had very little crossbreeding with the group of humans who occupied Africa. How that gets cast out genetically in detail is mostly irrelevant to the point that there's every reason to believe that the two groups have gone their separate genetic ways over those tens of thousands of years. Given these separate paths, it would have been quite a miracle had such a crucial item such as brain functioning developed in exactly identical directions and speed in both groups. All the SAT scores are demonstrating, in my opinion, is that there was no such miracle.

And the point regarding the fact that there are greater differences within groups than between them certainly doesn't mean we can't measure or expect differences between groups. Indeed, I'd say that the very best reason to believe their might be differences between the groups is that there is an omnipresent variability within each group. That variability within a group is exactly what evolution requires in order to select for the more useful version of a trait. If the group of human beings in Africa were being selected for cognitive functioning at an ever so slightly slower rate than the group of human beings in Europe, then one would expect that the European group would have higher cognitive functioning on relevant dimensions over the course of tens of thousands of years.

[GodwinGrey]

LiberalBioRealist wrote - The bottom line here is that the children of blacks of some of the highest educational and economic levels are outperformed by the children of whites of some of the lowest such levels.

It's a really hard result to explain away by environmental accounts of IQ.

The bottom line here is that some of the children of whites in some of the highest educational and economic levels are outperformed by some of the children of blacks in some of the lowest such levels.

It's a really hard result to explain away by genetic accounts of IQ.

[Lazar]

Why does the "no biologically inherited difference" side keep harping on the "more difference within races than between them" theme? It doesn't support the conclusion you keep bringing it up as supposed support for. We all know that the existence of differences that are not due to biological inheritance does not indicate the non-existence of differences that are, or in other words that there's no reason why the two things can't coexist. We all know that the fact that
x>y does not mean y=0. We all also know that practically nobody who says that there are biologically inherited differences in intelligence has claimed that it's the only factor that exists or even that it's bigger than all others combined, which is the ONLY kind of claim that would actually be countered by this within/between thing you keep trotting out.

Yes, I know it's true, but what's the point of badgering on and on about it? It's like responding to someone who says that properly working freezers keep their contents below the temperature at which ice would melt, by protesting "No, no, no! They use electricity!" Truth alone doesn't mean it actually has anything to do with anything.

Note that there is no "no biologically inherited difference" side. Only a side which says that the genetic component of IQ does not account for the differences between 'racial' groups. Importantly, as I have posted before this is what the evidence to date suggests.

As for your complaint, the point is to show that groups are heterogeneous and thus it is quite possible that defining them as a inherent group may simply be false and thus group differences are merely be an artifact of the way an individual arbitrarily decides who goes where and who they exclude from analysis. That is that typically we would want groups to maximise in-group similarities and out-group differences. This is clearly not happening in the case of race and IQ and thus one wonders whether it makes any sense to talk about such groups at all.

CJ will have another perspective on this than me but from where I am sitting I am asking your side of the debate to show that after controlling for individual level differences there is ANY significant different left to account for. That is those who are saying these are real groups that really differ from each other need to show that a significant percentage of variance is explained in IQ after controlling for the individual level. Do that and we can start having a debate on what group differences mean.

Again this is the problem I have been harping on about. The moment you do an ANOVA or T-Test type analysis you as the researcher make a decision that the group level is and implicitly (or explicitly as the case may be) suggest it is the most important level at which to conduct analysis - in all the group differences I have seen to data these a-priori assumptions are the ones being made. Essentially this means that any group level analysis squashes individual level variance (or sweeps it under the rug). We can however, empirically test the percentage at which individual and group level account for the variance in a construct. As yet I have found no evidence that this has been done and thus no evidence on which to say that the group level explains anything AT ALL. I have, however, placed calls for this data to which no one on this thread has responded. As such I will continue to respond that there is more variance between groups than within groups until such evidence is forth coming. Want me to stop saying it, well bring on the evidence (random effects model is what I am after thanks - that is something that explores individual and group level variance).

[Delvo]

What do you mean by "control for individual differences"? "Controlling", in every instance I've seen, is a group issue. For example, you can "control" for parental income by making sure that you only compare people whose parents had equivalent incomes, or "control" for education by making sure that you only compare people with the same level of education. But then that's creating other groups: income level groups and education level groups. What is "controlling" when it isn't about some kind of group?

[Lazar]

What do you mean by "control for individual differences"? "Controlling", in every instance I've seen, is a group issue. For example, you can "control" for parental income by making sure that you only compare people whose parents had equivalent incomes, or "control" for education by making sure that you only compare people with the same level of education. But then that's creating other groups: income level groups and education level groups. What is "controlling" when it isn't about some kind of group?

It refers to analysis in which both individual level and group level contributions to X are simultaneously calculated. Similarly, one could look at the effect of SES on IQ at both the individual level and the group level. I use control for individual differences to make it easy to understand (clearly it was not). What it means in essence is that you can empirically determining the relative contributions of individual level and group level rather than deciding a-priori which level to focus on.

[nac]

What has this got to do with my point? If you wish we can take it down to just two groups - those who left Africa at that time and those who did not . Show me statistics that demonstrate that the contemporay progeny of those left behind
have done better in IQ tests in the USA than the progeny of those who migrated.

For a systematic study, we'll need dozens of individuals from each group, all raised in the same culture and environment, plus control experiments.

[Darkchilde]

MOD NOTE

liberalbiorealist has been banned for racist bigotry, as seen in the only 3 posts in this thread. Please continue the discussion, as before this individual joined.

END MOD NOTE