Talk:Race and intelligence/Archive 73

Archive 70

Archive 71

Archive 72

Archive 73

Archive 74

Archive 75

→

Archive 80

Overall balance

In addition to making the article’s “outline” more accurately reflect the consensus among researchers, as Distributivejustice is suggesting, I think the overall balance of the article ought to be improved. As per WP:UNDUE we should be providing more space to the environmental viewpoint than we do to the hereditarian one, but that doesn’t mean we should prevent environmental perspectives completely uncritically, particularly when there are well-known criticisms of some of these positions which aren’t currently mentioned at all. Similarly, the “criticisms of hereditarian positions” shouldn’t completely overwhelm the evidence presented for genetic factors, of which the article (the title of this section notwithstanding) presents almost none.

In addition to the replacing T34CH’s outline with the new one proposed by Distributivejustice, I’d like to propose a couple of other changes. First, I’d like the article to include some of the criticisms of environmental hypotheses, which for the most part are currently missing. And second, I’d like the “evidence for genetic factors” to actually include some of the evidence that readers would expect from the title of the section for it to contain. I might base this on Legalleft’s proposed edit here, although since his explanation of this is longer than what’s necessary for the article, I’ll definitely need to summarize.

Keep in mind, I’m not going to change the general proportions that these sections have to one another, which are about what they should be. The goal here is just to provide an explanation of what the evidence is both for and against each viewpoint, which is definitely appropriate given what the academic consensus is on this topic (as Distributivejustice pointed out), which is that the evidence is not strong enough one way or the other to determine with certainty that either hypothesis is correct. --Captain Occam (talk) 07:12, 18 October 2009 (UTC)

The key problem with the article seems to be that the important details are buried among many unimportant ones. --Distributivejustice (talk) 07:53, 18 October 2009 (UTC)

I agree that what you mentioned is a problem also. What would you suggest doing to fix it? --Captain Occam (talk) 08:04, 18 October 2009 (UTC)

One key suggestion comes to mind -- mostly kill the material with primary literature citations. If the only available citation is to the primary research result, then it's probably not a well supported conclusion. The emphasis should be on material that you find in multiple reviews. Thus, stick to the canonical examples. I find that book reviews (e.g. Hunt 2009 doi:10.1016/j.intell.2009.05.004) are a good source, probably because they are at the top of multiple of levels of filtering (the info passed from the primary literature, then into a book, and then into a review of that book). --Distributivejustice (talk) 08:24, 18 October 2009 (UTC)

That idea sounds like it could work, but I think it needs to be weighed against the balance point that I made.

For example, the “racial discrimination in education” section currently cites two primary literature papers about structural equation models that were used to determine whether the development of children in any ethnic group was affected by discrimination. These studies (and others like them) were considered important enough to be included in Jensen’s and Rushton’s 2005 overview of the research on this topic, and they’re also talked about in Jensen’s book The g Factor, but I’m not aware of whether they’ve made their way into any book reviews on this topic. Even if they haven’t, though, this is a pretty significant criticism of the theory that IQ differences are caused by discrimination in education, so it seems like it would be unbalanced to talk about this theory without mentioning the most important criticism of it.

In any case, you’re welcome to try cleaning up the article via the route you’ve suggested, as long as you keep the balance issue in mind also. I don’t read book reviews on this topic as often as you apparently do, so if you want to try this it’s probably best for you to do it rather than me. --Captain Occam (talk) 08:46, 18 October 2009 (UTC)

Sure, it sounds as if there are multiple reviews describing the results you mentioned. Re: book reviews -- I was just suggesting that they are a good way to find the absolutely most critical concepts. --Distributivejustice (talk) 08:56, 18 October 2009 (UTC)

Nisbett (2005); Dickens (2005); Fryer & Levitt (2006)

Under the heading "Evidence for environmental influence" are three paragraphs presenting the findings of three pieces of literature: Nisbett (2005), Dickens (2005) and Fryer & Levitt (2006). While I do not doubt the importance of their findings and their relevance to this topic, I'm curious to know why they have been placed under this heading. A reader would expect to find what this heading indicates, i.e. "evidence for environmental influence". Instead, what we have are three studies focused on attempting to disprove the hereditarian position. Disproving the hereditarian position is not the same as providing evidence for environmentalism. It would seem these studies are being misused, and that, if anywhere, they should be used as criticism of the hereditarian position. If they do provide positive evidence for environmentalism, then perhaps we should be looking for those portions of the studies and including them here. Am I alone in perceiving this as a problem? --Aryaman

Agreed. David.Kane (talk) 13:42, 18 October 2009 (UTC)

If the content refers to the references they should stay. Where the reference occurs is of little import. That it's being cited is. Aprock (talk) 15:03, 18 October 2009 (UTC)

I'm not quite sure I understand what you mean, Aprock. This isn't about removing specific references. It's about there being a sizable section under the heading "Evidence for environmental influence" which does not provide evidence for environmental influence. While it may be good criticism of the hereditarian position, as it's being presented now, it's not what the article says it is. I'm not saying the material should be removed from the article entirely. I'm saying this content under this heading is misplaced. Something needs to be done here. Seeing as T32CH has managed to get Occam blocked for 72 hours on a ridiculous 3RR "violation", I'm less than willing to undertake any significant changes to the article myself without first consulting everyone else. Provided other editors can see the problem and agree, we can either re-title it, move it to a more fitting section or simply remove it for now until it can be incorporated in a more accurate way. If I'm not understanding your point, please do explain. --Aryaman (talk) 16:32, 18 October 2009 (UTC)

I see what you're saying. I had read the talk comment as being about the references, not the article content. Too early for me I guess. I believe this content is there because the order of environment and genetics was reversed. I certainly think that cleaning up the opening to that section should be done to reflect the current structure of the article. Aprock (talk) 16:58, 18 October 2009 (UTC)

Trying to explain the situation, this paragraph dates back to a time where anything that was critical of the hereditarian position was considered an argument for the environmental position (an either/or situation, if you will). As long as the section isn't modified without consensus beyond positioning and titling, I have no objection to its being moved and appropriately retitled.--Ramdrake (talk) 16:37, 18 October 2009 (UTC)

I figured that's what had happened, and that it was probably simply being overlooked. Seeing as this material is really directed at demonstrating a lack of evidence for the hereditarian position, my proposal for now would be to move these three paragraphs down to the criticism section and either giving all three one all-encompassing title, or perhaps giving each one a title of its own. Rather than simply coming up with titles, however (to avoid WP:SYNTH or WP:OR), I would prefer if we could show that, whatever title(s) they get, the sources are really making a pointed attack on a specific aspect of the hereditarian position. If not, or if they are making an attack on several points summarily, then I would favor putting them all under one general heading. Comments? --Aryaman (talk) 16:55, 18 October 2009 (UTC)

No objection, as long as the content is preserved until we get around to discussing it specifically.--Ramdrake (talk) 17:07, 18 October 2009 (UTC)

I've moved the content in question. Others can change the (now provisional) title as necessary. Also, I made minor changes to the order, as I think substantial criticism should be presented prior to any tangential criticism. Unless others object, I shall consider this thread concluded. Thanks to all the participating editors. --Aryaman (talk) 17:25, 18 October 2009 (UTC)

Just did that. I want wo stress that I'm not married to the new title, but thought it was marginally more to the point than the original one. Please feel free to change to a more appropriate one if found.--Ramdrake (talk) 17:27, 18 October 2009 (UTC)

Overview section inaccurate

Restarting from above. It is clear to me that the overview section is an inaccurate summary of Neisser et al. Instead it seems to merely summarize Sternberg. The text of Neisser is quite dense and is making a complex argument. Please reread carefully. They make multiple, related conclusions about nature-nurture, and the context matters for what those conclusions mean. I had to read quite a few other reviews to figure it out. -Distributivejustice (talk) 18:07, 18 October 2009 (UTC)

to summarize Neisser on nature-nuture:

• no one knows the cause of group differences
• some kind of environmental explanation is plausible becuase we know that the Flynn effect exists
• good candidates for a specific environmental cause are caste and culture
• the plausibility of a genetic cause hinges on the answer to the between group envirnment question, to which there is no answer
• there is no direct evidence for a genetic cause

the key thong to notice is that the Flynn effect isn't itself a cause but rather a model for resolving environmental causes in general with high heritability --Distributivejustice (talk) 18:17, 18 October 2009 (UTC)

Thank you for sharing your opinion, but I respectfully disagree with it. From where I say, your overview is less accurate than T34CH's.--Ramdrake (talk) 18:19, 18 October 2009 (UTC)

Could you support your alternative reading here? The text is really quite complicated, and these are critical distinctions to understanding their consensus. --Distributivejustice (talk) 18:25, 18 October 2009 (UTC)

Though I can't comment on the accuracy of this as a summary of Neisser, I would say that I find it to be a good overall summary. It also accords well with the passage I quoted some time ago from Snyderman & Rothman. (Why does it not surprise me to find that Wikipedia has that article? lol.) --Aryaman

[outdent] to clarify my aim relative to the article -- the current overview isn't a summary of Neisser et al who were making quite different arguments and conclusions (which I'm inclined to believe that I have accurately summarized). That's the key error. --Distributivejustice (talk) 18:47, 18 October 2009 (UTC)

Ah, sorry for misunderstanding. In that case, this should be relatively straightforward. Just provide citations for those points (meaning page numbers), and we should be able to move on. If it's disputed, you could provide direct quotes on the talk to clear up any doubts. --Aryaman (talk) 18:49, 18 October 2009 (UTC)

Summary of Neisser et al

I outline below the 2 conclusions and 3 additional findings in Neisser et al related to the cause of group differences in test score performance.

Start in section 5 "Interpreting Group Differences"

The introduction section is a summary of what is to following (there are no closing summaries of each section). The introduction is just one paragraph and states their two main conclusion.

• Main Conclusion 1: we do not know what causes test performance differences despite a variety of hypotheses:

"""If group differences in test performance do not result from the simple forms of bias reviewed above, what is respon- sible for them? The fact is that we do not know. Various explanations have been proposed, but none is generally accepted.""" (page 94)

• Main Conclusion 2: The Flynn effect demonstrates that there must be some (yet unidentified) environmental effect powerful enough to produce difference in test scores of the needed magnitude to explain the Black/White differential, but it's not known what those factors are:

"""It is clear, however, that these differences-- whatever their origin--are well within the range of effect sizes that can be produced by environmental factors. The Black/White differential amounts to one standard devia- tion or less, and we know that environmental factors have recently raised mean test scores in many populations by at least that much (Flynn, 1987: see Section 4). To be sure, the "Flynn effect" is itself poorly understood: it may reflect generational changes in culture, improved nutri- tion, or other factors as yet unknown. Whatever may be responsible for it, we cannot exclude the possibility that the same factors play a role in contemporary group differences. """ (page 94)

Those are their two key conclusions, stated directly in the introduction.

Now consider section 6 "Summary and Conclusions"

They summarize 5 points here, 2 of which are their main conclusions:

• Finding 3: No test bias:

"""The cause of that differential is not known; it is apparently not due to any simple form of bias in the content or administration of the tests themselves.""" (page 97)

• Main Conclusion 2 restated:

""" The Flynn effect shows that environmental factors can pro- duce differences of at least this magnitude, but that effect is mysterious in its own right.""" (page 97)

• Finding 4: On the evidence for specific environmental causes:

"""Several culturally-based explanations of the Black/White IQ differential have been proposed; some are plausible, but so far none has been conclusively supported.""" (page 97)

• Finding 5: On the evidence for genetic causes: """There is even less empirical sup- port for a genetic interpretation.""" (page 97)

• Main conclusion 1:

"""In short, no adequate explanation of the differential between the IQ means of Blacks and Whites is presently available. """ (page 97)

They close section 6 with a list of open questions, in which they restate main conclusion 1 and 2 and highlight findings 3,4,5:

Main conclusion 2 (the Flynn effect):

"""Mean scores on intelligence tests are rising steadily. They have gone up a full standard deviation in the last 50 years or so, and the rate of gain may be increasing. No one is sure why these gains are happening or what they mean. """ (page 97)

Finding 3:

"""The differential between the mean intelligence test scores of Blacks and Whites (about one standard devia- tion, although it may be diminishing) does not result from any obvious biases in test construction and administra- tion, nor does it simply reflect differences in socioeco- nomic status.""" (page 97)

Finding 4: """Explanations based on factors of caste and culture may be appropriate, but so far have little direct empirical support.""" (page 97)

Finding 5: """There is certainly no such support for a genetic interpretation.""" (page 97)

Main conclusion 1: """At present, no one knows what causes this differential. """ (page 97)

--Distributivejustice (talk) 19:40, 18 October 2009 (UTC)

Nice work, Distributivejustice. :) --Aryaman (talk) 19:45, 18 October 2009 (UTC)

Nice work too, but I don't read your Main Conclusion 2 the same way you do (i.e. I don't read in it the emphasis you seem to read in order to claim that it is a main conclusion).--Ramdrake (talk) 19:59, 18 October 2009 (UTC)

It might be worth while to take the following into consideration: Never a Dull Moment (Neisser, 1997). --Aryaman (talk) 20:18, 18 October 2009 (UTC)

I commend you on your work, but you seem to have missed an entire page of Summary and Conclusions. Specifically see the four paragraphs beginning with "Like every trait..." followed by the three paragraphs starting with "One of the most..." (page 96). Because it is probably the most relevant quote, it might also be worthwhile to just include all of point (6) instead of slicing it up:

The differential between the mean intelligence test scores of Blacks and Whites (about one standard deviation, although it may be diminishing) does not result from any obvious biases in test construction and administration, nor does it simply reflect differences in socio-economic status. Explanations based on factors of caste and culture may be appropriate, but so far have little direct empirical support. There is certainly no such support for a genetic interpretation. At present, no one knows what causes this differential.

Aprock (talk) 20:50, 18 October 2009 (UTC)

Oh, yes, I definitely skipped several related sections, such as within-group heritability and mean test score differences. I just wanted to emphasize the points relevant to showing that the current overview section doesn't match what Neisser et al wrote. They made a very different argument than what's currently written in that section of the article. --Distributivejustice (talk) 21:59, 18 October 2009 (UTC)

Thank you for clarifying that. I agree that the overview needs a lot of work. In fact, I would even suggest that you go ahead and actively improve it. I'd suggest keeping it at a high level and well organized, but beyond that, I don't think any of the basic issues you raised should be categorically excluded. Aprock (talk) 22:20, 18 October 2009 (UTC)

I just worked through the "Summary and Conclusions" section and have myself summarized the first 12 paragraphs for quick reference. The remaining part of the conclusion is itself numbered, so I felt there was no need to summarize its content - really, the 7 points offered in the text summarize the 12 paragraphs which precede them. My summary of the first 12 paragraphs follows. Please note that each paragraph has been summarized as one sentence:

• §1. Psychometric testing, though one of the most fruitful approaches to studying intelligence, has yet to produce answers to many questions regarding intelligence. (pg. 95)
• §2. Psychometricians have devised ways to measure the distinct yet intercorrelated abilities believed to play an important role in the development of intelligence, though the correlations between those abilities remains unclear. (pg. 95-96)
• §3. Although intelligence test scores correlate moderately well with educational measures, and thus measure important skills, educational achievement is not primarily determined by intelligence. (pg. 96)
• §4. Intelligence test scores also correlate significantly with adult occupational status and modestly with undesirable and antisocial behavior. (pg. 96)
• §5. While both genetic and environmental variables are involved in the manifestation of intelligence, the role of genetics increases in importance with age. (pg. 96)
• §6. Why this happens in not yet understood, and neither is the environmental contribution to this increase. (pg. 96)
• §7. One important environmental factor is formal education, which positively affects intelligence, though the positive impact of intervention programs usually fade at their conclusion. (pg. 96)
• §8. Although not yet fully understood, it is clear that some environmental factors which negatively affect health can also have negative affects on the development of intelligence. (pg. 96)
• §9. The "Flynn effect", which refers to the striking worldwide mean IQ increase of 15+ points over the last 50 years, may be the result of environmental factors such as improved nutrition, cultural changes, improvement in the administration of tests, changes in educational practices or some other hitherto unrecognized factor. (pg. 96)
• §10. Although sex differences in overall intelligence are minimal, there are substantial differences in the distribution of specific abilities between the sexes, which may be caused by a combination of factors such as hormone levels and specific social environment. (pg. 96)
• §11. As the measured differences in intelligence between various ethnic groups is the result of complex patterns, broad generalizations should be avoided, though intelligence test scores in some minority populations are reasonably good indicators of educational achievement levels in later life. (pg. 96-97)
• §12. The long-standing 15+/- point difference between the intelligence test scores of African Americans and White Americans, though it may have narrowed somewhat in recent years, remains unaccounted for despite proposed explanations claiming bias, differences in culture or socio-economic status, or genetics as the underlying cause. (pg. 97)

I don't know if that will help resolve this minor dispute, but I think §5-§12 (with the exception of §10) should be summarized in this way. --Aryaman (talk) 21:25, 18 October 2009 (UTC)

I appreciate your summary VA, I think it's very clear and well done. Aprock (talk) 21:47, 18 October 2009 (UTC)

Thanks, Aprock. If you or Distributive Justice would like to use any part of that in paragraph form to improve the "Overview" section, I think the article would benefit greatly. I think Neisser et al. is giving a very good overview of what's known and what remains unknown, and seems to be about as non-partisan as one can hope for. I have several reservations about the current "Overview", as I feel that it does not accurately reflect the current state of the research, and is more concerned with partisan preempting of the hereditarian position rather than simply stating which issues are involved. I hope these issues can be resolved in an atmosphere of collegiality. Thanks, --Aryaman (talk) 01:12, 19 October 2009 (UTC)

Suggested change, comment on the proposed 'change

The consensus among intelligence researchers is that IQ differences between individuals of the same racial-ethnic group reflect real, functionally and socially significant, and substantially heritable differences in intelligence.

My problems.

• "The consensus"- which source says it's a consensus?
• "IQ differences between individuals"- i.e. variation
• "of the same racial-ethnic group"- at least one of the sources doesn't say this, "Intelligence knowns and unknowns" doesn't say this, it says "It is clear (Section III) that genes make a substantial contribution to individual differences in intelligence test scores, at least in the white population." which is not the same at all.[1] BTW the article also states, as I do above, that "So defined, heritability (h2) can and does vary from one population to another. In the case of IQ, h2 is markedly lower for children (about .45) than for adults (about .75)" and "The value of h2 can change if the distribution of environments (or genes) in the population is substantially altered."
• "reflect real, functionally and socially significant"- I don't know what this part is supposed to be saying. I don't see why it can't just say "reflect".
• "and substantially heritable differences in intelligence"- this is just a convoluted way of saying that the differences mentioned earlier in the sentence have a substantial genetic component, right? But it's written in a confusing manner.

I'd suggest this:

It has been estimated that genes contribute substantially (40-80%) to IQ differences between individuals within some human groups.

I think it says pretty much the same thing, but in a simpler and easier to understand way. I also think it more accurately reflects what at least one source says.

Cheers Alun (talk) 08:08, 15 October 2009 (UTC)

This is clear and precise and accurate - I am all for it! Slrubenstein | Talk 12:53, 15 October 2009 (UTC)

I prefer this to what is there, but it's not clear what "within human groups" means in this context. Aprock (talk) 14:24, 15 October 2009 (UTC)

We could change "human groups" to "population" or even "breeding population" if you prefer. What is confusing about the word "group?" Slrubenstein | Talk 16:00, 15 October 2009 (UTC)

I just don't know what "within human groups" means in this context. Maybe if I read the cite I'd know. But that's that basic problem I guess. If I have to read the cite to understand the sentence, it's not a good summary. Aprock (talk) 16:21, 15 October 2009 (UTC)

You are justt repeating yourself, I was hoping you would explain why you don't understand it. I have read some of the sources but even if I had not the meaning seems pretty clear to me. There is a definable group of human beings. Members of this group scored differently in IQ tests. According to one study, genetic differences accounted for forty percent of the variation among these members of the group. According to another study genes accounted for 80% of the variation in IQ score among members of this group. Which part of this is obscure to you? Slrubenstein | Talk 16:41, 15 October 2009 (UTC)

As I said this is better than what was before, but I don't know what groups means here. How large are the groups, how are they defined, how were they formed, are they from clinical trials, are they racial groups, are they families. Aprock (talk) 16:59, 15 October 2009 (UTC)

I see your point Aprock, but it really means any group of humans, however defined. So it could be a group of children, or a group of people living on a specific street, or a group of people who worship at the same church, or a group of people who identify as belonging to the same nation, or who share the same citizenship, or even who identify as belonging to the same ethnic group. Essentially it doesn't really matter how we define the group. I can't think of a better way to say this, unless we say something like "within some human groups, however defined)"? Alun (talk) 17:17, 15 October 2009 (UTC)

That clears it up more for me. But the sentence is particularly confusing because of the topic of the article. A cursory reading of the sentence gives me the impression that groups here refers to racial groups. Maybe it should be "any group" instead of "some group"? Or maybe "any collection"? Aprock (talk) 17:38, 15 October 2009 (UTC)

This is where it is especially important to make a distinction between "race" and group. A race is a category, it is a way of thinking of people. A group is an aggregate of people who interact or are interconnected in some way. In the United States, segregation ensured that for over a hundred years (the era of Jim Crow) the "Black" race and the "White" race were also groups. I may not agree with Alun here - if the White and Black races are expanded to include everyone living in Europe plus Arabs, and everyone living in Africa and Melanesia, I am not sure they are "groups" any more. Someone just decided to classify people into two classes. I bet if you go back to the articles that measure heritability, there is some criteria for group. In sociology a group and a class are not the same thing. In practice, they may overlap, but they are still two different things. For the purposes of determining heritability, you are looking at a group. But race is a social category. In the US, when studies are made and the group is defined as white or black, it is very important to attend to scale - they are usually talking about American whites and American blacks, and these indeed may be groups but the results do not automatically transfer to other people who for one reason or another are classified as "white" or "black." Slrubenstein | Talk 18:08, 15 October 2009 (UTC)

I think I understand that. Is there a citation which shows that heritability is related to distinct groups of people, as opposed to the group of all people? Aprock (talk) 22:43, 15 October 2009 (UTC)

You could use the word "set" instead of group I suppose. By group, what we really mean is that set of individuals that supplied the data that we have used to calculate the heritability estimate. This estimate will vary from group to group, depending on how environmentally homogeneous the group is. So, for example let's take a group of individuals who share many environmental characteristics, say they all have parents who are both graduates and who both worked, they all went to the same school at the same time, and grew up in the same village, and all of their parents stayed married etc. etc. The amount of environmental variation is smallish, so we will probably find that our estimate of heritability is high. Now let's take the same number of people of the same age, but make them a random sample taken from the whole of the USA, this group will almost certainly be much more environmentally heterogeneous, and well probably find that our heritability estimate is much smaller. The thing about heritability of IQ is that we always find some within group genetic contribution to the variation in intelligence, which just means that some of the difference between the high scoring ones and the low scoring ones is due to genes (remember it's the difference between individuals within group). So how one defnes one's group is very important for heritability estimates. The reason it's "some groups" is that we can't say "all groups", because we simply can't know how heritability estimates will look for all possible groups. Alun (talk) 05:30, 16 October 2009 (UTC)

(outdent) The fact that people are assuming this statement refers to "any group" is proof enough that the change was most decidedly not an improvement. The tests that were performed to determine heritability were performed on people of the same racial/ethnic group - most reliably on twins. To obscure this fact reduces the rest of sentence to gibberish. If you guys really want to have the statement as it reads now, then you need to get rid of that whole string of citations and go find new "sources" claiming that some "universal heritability" of .40-.80 exists so that, regardless who you take, any group will give you that result, like Slrubenstein says - which you won't find, because it defies the concept of heritability altogether. This is just bizarre, lol. --Aryaman (talk) 18:09, 15 October 2009 (UTC)

The old sentence was no more clear about this than the new sentence. Maybe you could propose an improvement? Aprock (talk) 18:30, 15 October 2009 (UTC)

How about this? I'll give you a standard explanation as presented by Flynn (2001:155-156), and you guys can decide how the summary should read, and whether consensus supports the old statement or what's up there now:

There are two principal kinds of evidence pointing to the conclusion that intelligence is substantially genetically determined and from which its heritability can be calculated. The first of these consists of studies of monozygotic twins reared apart. It has been found that these have highly similar IQs, represented by a correlation between adult twin pairs of .72 (Bouchard, 1993). This figure needs to be corrected for the unreliability of measurement. Assuming the test has a reliability of .9, the corrected correlation between the twin pairs in .80. This correlation is a direct measure of heritability. The second method consists of comparing the degree of similarity between identical twins and same-sex, nonidentical twins brought up in the same families. Because identical twins are genetically identical, whereas nonidentical twins have only half their genes in common, if genetic factors are operating, identical twins should be more alike than nonidenticals. The simplest method for quantifying the genetic effect was proposed by Falconer (1960) and consists of doubling the difference between the correlations of identical and same-sex nonidenticals. Studies of the intelligence of adult twin pairs have been summarized by Bouchard (1993, p. 58). He found correlations of .88 for identical twins and .51 for same-sex, nonidenticals. The difference between the two correlations is .37, and doubling this difference gives a heritability of .74. This figure needs to be corrected for the imperfect reliability of the tests. Using a reliability coefficient of .9, the corrected correlations become .98 for identicals and .56 for same-sex nonidenticals. The difference between the two correlations is .42, and doubling this difference gives a heritability of .84. This is very close to the heritability of .80 derived from the first method. This is why most experts on the issue estimate the heritability of intelligence as approximately of .80, or 80 percent.

As you can see, this is not derived from studies of "any group", but has a very specific meaning. Good luck with summarizing. --Aryaman (talk) 18:41, 15 October 2009 (UTC)

I see nothing in that blockquote about racial or ethnic studies. If I were to use that information as the source of the summary, I would strike all reference to the word group. Aprock (talk) 18:47, 15 October 2009 (UTC)

Arya is confusing many things. Race and ethnicity are two different things. Both are social ategories, not biological groups. Which of these studies is about "race" and e carefu. of they use the word race, what do they mean by it? Arya seems to have a racialist agenda, inserting race where it is not an issue. Slrubenstein | Talk 19:38, 15 October 2009 (UTC)

Judging from the section "Personal opinions regarding current areas of activity" on his User page, you may be right. Aprock (talk) 20:07, 15 October 2009 (UTC)

:-) Forgive me for laughing, but I have to. I quote a section from Flynn, and now I'm a "closet racist"? Oh, bother, lol. The only reason any of this is being discussed as far as I'm concerned is to explain the existence of "racial-ethnic group" in that original statement - which I most certainly did not write, I might add. If you want someone to "blame" for introducing "race" into the discussion, look to Jensen, Rushton, Flynn, etc. They are the ones using heritability to make claims regarding differences in the intelligence of racial-ethnic groups, and the fact that the original statement included that phrase was, as far as I could tell, motivated by nothing other than intellectual honesty on their part. As far as the use of twin-studies go: if you don't see the connection between doing studies on "white" twins and "black" twins and making statements about the heritability of intelligence in "white" populations and "black" populations, then this topic will require more explanation than I am willing to go into. I was perfectly happy with the original statement leaving the specifics of race out of the equation, and I was not the one who initiated changing this statement. So, please, leave the race card well alone.

With that aside, are you truly satisfied with the changes to the statement? Do you think that the sources you're using to support this new statement actually say what you portray them to say? --Aryaman (talk) 20:35, 15 October 2009 (UTC)

No one called you a closet racist. Defending yourself against such is probably not very useful. It would probably be better if you focused on constructive contributions instead. In this section we are discussing a specific sentence. I have asked what the meaning of group is in this context. Given that the blockquote in your response don't indicate any grouping whatsoever, I'm curious as to which source the grouping phrasing is from. I'm more than capable of examining the source myself, but I'm not going to read a dozen sources just to find this one (poorly defined and sourced) claim. Aprock (talk) 20:41, 15 October 2009 (UTC)

Someone insinuated I could have a "racialist agenda". OK, I'll work on not seeing that as ad hominem. ;-)

I really was hoping the Flynn passage would help clear this up. I'm doing the best I can with WP:AGF. Maybe I'm trying too hard to give the different views a fair shake here. Maybe I'm trying too hard to prevent uncritically applied changes from leading this article into blatant self-contradiction. Maybe I need to relax, let you guys do your thing, and wait until you've made the changes you see as fit, and then go through it. Because this is getting us nowhere. I'm startled at the proposition that the behavioral genetics research conducted on twins - chosen precisely because of their genetic similarity/identity - is being interpreted as applying equally to "a group of children, or a group of people living on a specific street, or a group of people who worship at the same church." This has nothing to do with my personal views on race, ethnicity or anything else. This has to do with reporting sources in a credible manner. What I personally think regarding heritability is entirely irrelevant. The question is: Is this what the sources are saying? Not that I can see. They are not taking groups of people off the street at random and testing them to determine heritability. But maybe Alun can prove me wrong there. If s/he can show that these researchers we've been discussing have in fact done their research on more or less randomly selected populations with no genetic or even ethnic affiliation (or is religious affiliation counted to ethnicity these days?), then I will stand corrected and promise to give you folks free reign. :-) --Aryaman (talk) 21:15, 15 October 2009 (UTC)

I was really hoping the Flynn passage would help too. Unfortunately, it didn't reference race, ethnicity, or groups of any kind. I share your apprehension about the use of the word group here. It's not at all clear where it comes from, or how the original sources use it. That is the specific thing I've been trying to clear up. As far as I can tell, and I haven't spent much time with those source, none of them are making any group level conclusions at all. Aprock (talk) 21:28, 15 October 2009 (UTC)

The problem with the Flynn passage is that all it does is explain how one measures heritability. That is not the point, no one disputes that heritability can be high. What the Flynn passage does not do is give estimates for heritability for different types of data. Assuming that Flynn's data are the usual data used for these analyses, then he's dealing with a very homogeneous sample group. The problem with the Flynn quote is that it doesn't really address the difference of different heritabilities between different groups. The quote you give is irrelevant to this discussion. Alun (talk) 06:11, 16 October 2009 (UTC)

If it would help, I could reference quite a few sources which make it patently clear that "groups" in this discussion are referring - particularly in the case of Jensen's, Flynn's and Lewontin's discussion of heritability - specifically to "white" and "black" groups. Would that help at all? --Aryaman (talk) 21:39, 15 October 2009 (UTC)

By all means, going over sources is what we're here to do. Aprock (talk) 21:41, 15 October 2009 (UTC)

(outdent) OK. I'm picking more or less at random, but here's one which, though slightly leaning in one direction - as is probably justified given the relative support the positions discussed enjoy - makes it perfectly clear that, in this discussion, "group" has a specific meaning (from: Snyderman & Rothman The IQ Controversy, the Media, and Public Policy (1990:123-124):

Compounding the difficulties in maintaining rational discourse about group differences is the fact that genetic influences on race and class differences in IQ are extremely difficult to estimate. Within-group heritability, because it is tied to the particular environment and genetic variation existing within the group, has little relevance to the causes of between-group variation. Jensen has argued that the high degree of heritability for IQ within the black and white populations makes it more probable that between-group differences have some significant heritable components as well. Richard Lewontin and others disagree, citing examples from genetic studies with other organisms demonstrating that high within-group heritability can be associated with almost any degree of genetic influence between groups."

[Then he gives a summary of Lewontin's corn analogy which I have removed.]

Lewontin's point is that even if we know that the within-group heritability of IQ is substantial, this tells us nothing about the possibility of genetic between-group differences; they are independent questions. Jensen demonstrates, however, that high within-group heritability necessarily implies substantial between-group genetic influences, as long as the sources of environmental variation are the same between as within groups. Large within-group heritability is only consistent with no genetic influence between groups if there is some source or sources of environmental variation that exists only between groups, like the variation in nutrients in Lewontin's example. But black and white Americans are two populations where such a source of environmental variation is very likely to exist: it falls under the heading of "racial discrimination". Thus, if the environments of black and white Americans differ in ways that are not generally seen between families or individuals within the black and white communities, estimates of within-group heritability are of dubious relevance to the between-group question.

The major obstacle to the study of the causes of group differences is that it may be impossible to randomize or control the relevant environmental factors, and thus separate genetic from environmental sources of variation. How can one be sure, for example, that a black and a white child have been raised in similar environments, when genetically based racial differences remain obvious? Even black and white children raised in the same home may be treated very differently because of their skin color. Many experts in the study of genetics have argued that the nature of the situation makes it impossible to adequately assess the question of the genetic influence on groups differences in psychological traits.

Thus, as I hope is clear, I'm not trying to push any "hidden agenda" here. I'm simply trying to see to it that an important aspect of the cited material is not mistakenly deleted. It is important to note that the groups under discussion - especially as it related to the Jensen/Lewontin debate on heritability - are "racial and ethnic" groups. If that is removed, the figures given become meaningless, as they were derived from race-ethnic group-specific studies. Does this help to clarify things? Or at least clear me of trying to push some "agenda"? --Aryaman (talk) 22:10, 15 October 2009 (UTC)

Arya, all this is irrelevant. My problem is not about between group differences. That is a different question and irrelevant to the change I want to make. My problem is that we need to make clear that heritability estimates vary between groups, however we define our group (and it can be defined by self described "race"). We also need to make sure that we do not try to make claims that the sources don't make. So the Snyderman study you quote from does not say "racisl"/ethnic groups, it say "black" and "white". Well that is not the same thing. We need to say what our sources say. Besides the Snyderman study is not a reliable source, it was not written by academics but by two right-wing journalists with a political point of view to push. We just had a discussion about citing the NYT, I don't think we should cite this for the same reasons, it's not peer reviewed, not published in a reliable academic journal, and not written by experts. Alun (talk) 06:11, 16 October 2009 (UTC)

It seems here that within-group and between-group are being used here with very specific meanings. Which citations refer to research results regarding within-group heritability? Aprock (talk) 22:40, 15 October 2009 (UTC)

Don't be fooled by the debate about within to between group variation, that is irrelevant to thie problem with the current sentence. The problem is that within group heritabilities are different for different groups. So perhaps for one set of data the heritability of IQ is 80%, but it can be as low as 25 or 30% for another set of data. This basically means that one set of data derives from people who have a homogeneous environemnt, and the other other set of data derives from a group of peope with a heterogeneous environment. Let's try to stick to the point. Everything Arya is saying is irrelevant to the changes I want to make. Alun (talk) 06:11, 16 October 2009 (UTC)

I have not researched the entire list of citations (I'm not even sure who originally found and cited them, as that was before my involvement with the article and I haven't searched through the very long history of this article), but I would say that the question is actually irrelevant. The original statement was simply stating that "the consensus among experts is that IQ differences between individuals of the same racial-ethnic group are significantly heritable", with it being assumed that "heritable" in this case is in reference to genes and not environment, as the very method being used to determine within-group heritability was specifically designed with the intent of isolating the role of genes in the expression of intelligence. Until someone shows me sources which specifically state otherwise, I don't even see this statement as contentious insofar as it's not advancing a claim which is under serious dispute, and if I were editing this article in isolation, I wouldn't even think it to be necessary to source it without an outside objection. What is under dispute is whether within-group differences can be used to make claims regarding between-group differences. (Jensen argues that it can, provided no "factor X" can be found (he did not find what he considered to be persuasive evidence for its existence); Flynn thinks that "factor X" exists, and many feel that "racial discrimination" is "factor X" (note that Snyderman & Rothman say "very likely"), with the debate as to the magnitude of discrimination taking place both within as well as beyond the domain of behavioral genetics.) However, regardless whether the sources cited are discussing within-group heritability or between-group heritability, and if the latter, regardless of which side of the debate they support, all of them must operate on the assumption that IQ differences between individuals of the same racial-ethnic group are significantly heritable, for that is the basis or starting point of the subsequent discussion. If this statement needs any qualification, in my opinion, it is in the direction of expansion, not simplification and reduction. Specifically, I think that there is equally little room for doubt - even from Jensen, as he was the one of the first to even discuss "factor X" - in the academic community that heritability can be demonstrated to vary with environment. This, too, is one of the fundamental assumptions upon which the discussion operates. But this is discussed as an "obstacle" at present, as there are currently too few reliable tools for isolating environment-dependent variation in heritability. That's not to say that better tools can't or won't be developed - I'm sure someone this very minute is working on exactly that somewhere in the world.

My point in discussing all this is that I think the current leading statements in the Genetic Factors section are misleading at best. Beyond the discussion of the first statement, I find the phrase "Lewontin suggests that some genotypes are more influenced by environments than others" in need of a much closer inspection. From my reading, his saying that some genotypes are subject to greater environmental influence than others is pointing to real-world, external circumstances, not inherent differences in genotype. But, like I said, this needs to be examined very closely.

Anyway, I hope I have been able to make clear why I think the original statement is far superior to what we have now. But, if the involved editors fail to perceive the genuineness of my concern and remain adamant, then I will refrain from continuing to try to explain myself. These discussions are proving quite useless as far as improving the article goes, and I do not enjoy wasting either my own time or the time of others. Thanks, --Aryaman (talk) 23:47, 15 October 2009 (UTC)

• "The original statement was simply stating that "the consensus among experts is that IQ differences between individuals of the same racial-ethnic group are significantly heritable", with it being assumed that "heritable" in this case is in reference to genes and not environment,"

Actually the original sentence says no such thing. What you say is far better, and I think I could live with something close to what you say here. I still think it is a bad idea to include the term '"racial"-ethnic group'. Even your Snyderman study doesn't use this term, it says "white" and "black". I'm also not sure which sources say this is a consensus. But what you say is actually very close to the changes I suggest. The problem is that you seem to think that the changes I suggest create a change in meaning, but they don't, the change I want to make doesn't substantively change the meaning of the sentence at all. That was never my intention, what I think the sentence actually is actually trying to say is correct with regards to the citations, I just don't think it does a very precise or accurate job of representing what the citations say. BTW heritability never refers to genes, it is a group level trait, it refers to the group in question. i.e. in this group the amount of variation attributable to genes is 80%-that is what heritability means. That is why a heritability estimate for one group cannot be applied to another group. When Jensen applies his 80% figure (calculated in a homogeneous "white" group, to all other groups, that is a mistake, and he has been rightly critisised for it. He is making the error of assming that heritability is a property of the trait (i.e. intelligence), and he was wrong. Now you are making a similar mistake. Hope that finally clears things up for you, I have after all explained this to you four or five times over the last day or so and you still appear to be confused about what heritability actually measures.

• "the very method being used to determine within-group heritability was specifically designed with the intent of isolating the role of genes in the expression of intelligence."

That statement is not correct at all. Heritability estimates cannot be used to "isolate the role of genes in intelligence", they are measures of variation within a trait. It is a statistical treatment of data. Seriously, the only way to to determine the role of genes in intelligence is to find the actual genes involved, that would mean molecular biology and not statistical analysis. I don't know where you get this idea from, but you are mistaken, no reliable source would make this claim.

• "What is under dispute is whether within-group differences can be used to make claims regarding between-group differences."

Well actually no, Jensen isn't stupid, he knows that he can't use within group variatio to make claims about between group variation. But he does claim that if the environments of African-Americans and "white" Americans are similar, then there can be no environmental explanation for the difference. But I digress, the discussion about within to between group variation is irrelevant to the changes I want to make to the article. The fact that you bring it up indicates to me that you don't clearly understand what my problem is. I suggest you take a look at the problems I have with the sentence again and respond to those. Thanks. Alun (talk) 06:11, 16 October 2009 (UTC)

Why has this sentence already been changed? The previous consensus that the sentence was fine in its original state can only be overruled by a new consensus that it should be changed, and there certainly isn’t one yet. VA’s points about this haven’t even been addressed yet.

I’m going to change it back to what it was originally. Everyone else, we need to continue following the original consensus about this until we’ve come to a new decision. --Captain Occam (talk) 00:52, 16 October 2009 (UTC)

I'm going to ask that you undo your revert. The working version is better than the version you reverted it to. If you'd like to be constructive, it would help if you described the problems with the way it currently is. Flying in and reverting a currently discussed topic is not constructive behavior. Aprock (talk) 01:06, 16 October 2009 (UTC)

Varoon Arya already described the problem with the new version of this sentence in his last few posts, and why he considers the earlier version of this to be better. So did I, in the earlier sections where we debated about this. Whether you choose to read our comments about this is up to you, but the explanations are there, and most of them haven’t been addressed yet.

In any case, that isn’t what matters here. What matters is that you and Alun are going against the previous consensus about what this sentence should say by editing it, when there is no new consensus for changing it. Consensus can change, but in this case it hasn’t yet. Until it does, we need to stick with what consensus originally determined that this sentence should say. --Captain Occam (talk) 01:18, 16 October 2009 (UTC)

Where was the current phrasing determined to be consensus? The version you reverted to has problems (see the root posting by Alun). I'll ask again that if you have a problem with the version you reverted from to state the problem. As far as I can tell, it was a working consensus. If it wasn't, what is the point of dispute? The only one who was raising issues about it was me, and I thought it was good enough until something better came along. Aprock (talk) 02:02, 16 October 2009 (UTC)

“Where was the current phrasing determined to be consensus?”

If you aren’t familiar with this article’s history, you need to look through the discussion archives to find this. I’ve been watching this discussion since 2007, and I remember this being resolved at some point during that time, but I don’t remember exactly when it was.

“I'll ask again that if you have a problem with the version you reverted from to state the problem.”

I already have, but I guess I’ll summarize what I’ve said already: Alun (and now you) have been trying to second-guess what the source material for this sentence says, based on what’s being reported in sources that aren’t currently being used for this sentence, even though those other sources do not appear to reflect the academic consensus. In fact, according to VA’s last post, the academic consensus appears to be the opposite for what Alun is claiming. Only one of the nine sources currently being used for this sentence does not support its original phrasing (the one from Neisser), so the obvious solution to this is to just stop using that source, and continue using the other eight.

I’ve explained this before, and it’s usually been ignored. That’s fine if you want to do that, but until the concerns raised by me and Varoon Arya have been addressed, the original consensus about what this sentence should say still applies. --Captain Occam (talk) 02:34, 16 October 2009 (UTC)

• "trying to second-guess what the source material for this sentence says"

Could you point out where I've "guessed" what the source says? Actually what I've done is to ask for clarification about what the sentence says, and to stste that at least one of the sources says something different.

• "based on what’s being reported in sources that aren’t currently being used for this sentence"

Actually I say that "intelligence knowns and unknowns" doesn't support what the sentence currently says, and that's the very first source cited to support the sentence. So I don't know where you get the idea that I'm only referring to sources that aren't used from. You appear not to have read my posts, or else to not know which sources are being used.

• "Only one of the nine sources currently being used for this sentence does not support its original phrasing"

So provide evidence that all of them say that heritability is high in "race-ethnic" groups, because the sections from the sources quoted above to support their use, none used this term at all. I repeat, if we are going to say that the sources claim that heritability of IQ is substantial in all racial-ethnic groups, thenw e need a source that makes this explicit claim. None of the sources make this claim in the sections quoted.

• "according to VA’s last post, the academic consensus appears to be the opposite for what Alun is claiming."

Actually VA's last couple of posts are irrelevant to my concerns. He appears to be interested in having a debate about within to between group variation. I am not interested in that debate, it is irrelevant. I say again, my concerns revolve about use of the term "racial-ethnic groups", which is not used by at least one of the sources, and the claim that heritability estimates are always "substantial". I'd like to remove the reference to "racial-ethnic groups" because I don't think anyone claims that they are substantial in "racial-ethnic groups", although they may claim that they are high in all groups that have been measured. I'd also like to say categorically that heritability estimates do vary between groups, and that they may be higher in some populations and lower in others, I'm concerned that we don't fall into the fallacy of saying that heritability is a property of the trait in question, it is a property of the population (or group) in question, and therefore if two groups have different environments, they may have different heritabilities for intelligence. Alun (talk) 07:56, 16 October 2009 (UTC)

The original statement is a correct reflection of the consensus. The terms "racial-ethnic", "substantially" and "heritable" are specific and necessary if the meaning is not to be distorted or easily misunderstood. (That misunderstanding is clearly possible with the suggested alternative should be clear from Alun's and Slrubenstein's comments above.) Including these terms in no way favors one bias or another. Also, please note that the original statement makes no attempt to quantify the role of genetics in intelligence - least of all for a "random" group. Different studies have yielded different results. But the consensus is that, even with those reported variations, the genetic contribution to intelligence is substantial, as opposed to trivial. That's just about as clear as I think this can be made. --Aryaman (talk) 03:49, 16 October 2009 (UTC)

I've asked several times for the citation which refers to the "racial-ethnic" group results. So far, there has been none. I'll add the appropriate tag till that gets sorted out. Aprock (talk) 04:18, 16 October 2009 (UTC)

It's in each of the citations, but you have to know what to look for. As pointed out above, the "groups" relevant to heritability are not arbitrary collections of people. There are actually two different kinds of groups which all of the sources emphasize, each sharing the property that sibilings/twins are always from the same group: SES and race/ethnicity. For example, from "Intelligence Knowns and Unknowns" they add this caveat after describing the heritability point estimates:

"These particular estimates derive from samples in which the lowest socioeconomic levels were underrepresented (i.e., there were few very poor families), so the range of between family differences was smaller than in the population as a whole. This means that we should be cautious in generalizing the findings for between-family effects across the entire social spectrum. The samples were also mostly white, but available data suggest that twin and sibling correlations in African-American and similarly selected White samples are more often comparable than not (Loehlin, Lindzey, & Spuhler, 1975)."

So emphasizing that heritability estimates refer to differences within races/ethncities (and also SES groups) is important. I also know why the original text says "real, functionally and socially significant", but that doesn't seem to be in dispute. --Distributivejustice (talk) 04:30, 16 October 2009 (UTC)

If it's in each of the citations, then someone should be able to pick one for me to go to the library and check out. As your blockquote points out, the conclusions in the first source weren't about racial-ethnic groups, but family groups. I suspect the same thing is true of the other sources. But if you can name one of the other sources which mention specifically that conclusions about with-racial-ethnic groups were made, I'd be happy to go look it up. Aprock (talk) 04:35, 16 October 2009 (UTC)

Sorry, maybe I'm not explaining myself well. It's a common and appropriate caveat added to descriptions of heritability that they are within racial/ethnic groups; and that caveat is universally applied when describing racial group differences. All of those sources include a similar description. This is mostly done because the majority of the data comes from "white" populations. It's been a long time since I read many of those books but I recall Brody (1992) for example saying that clearly. Could someone explain why that phrase elicits particular concern? Could it be said more clearly? --Distributivejustice (talk) 04:41, 16 October 2009 (UTC)

That is certainly a much better explanation than some of the others. I'll go look up Brody. And yes, it appears that the problem is that this introductory sentence is burdened by the use of overloaded jargon. As I mentioned above, it would probably be better to use a more straightforward description in the opening paragraph of the section, and discuss the very issue you mention directly in the body of the section. Aprock (talk) 04:45, 16 October 2009 (UTC)

• "It's a common and appropriate caveat added to descriptions of heritability that they are within racial/ethnic groups"

Really? I don't know if that is true. Usually they refer to a specific group. So they may say something like" the heritability of IQ within the "white" group is 80%" or "the heritability of IQ within the "black" group was 40%". But that's the point isn't it? We need a source that says that heritability of IQ is high for "racial-ethnic" groups, notone that says that it's high for a specific group that has been measured. Even if we find a source that states that heritability was measured as high for a white set of subjects and a black set of subjects, it doesn't follow that it is high generically for "racial-ethnic" groups. Mostly the high estimates have been derived from specific "white" populations. Alun (talk) 07:58, 16 October 2009 (UTC)

I think I understand what wasn't clear now. The key thing is that family-based IQ heritability studies only and always measure the heritabilty of "IQ differences between individuals of the same racial-ethnic group". It would take some other kind of study to measure the heritabilty of IQ differences between racial-ethnic groups, a value for which there definitely is not a "consensus among intelligence researchers". So, it seems that the very dense single sentence should be two or more sentences to give proper context because it's been misunderstood here. --Distributivejustice (talk) 16:31, 16 October 2009 (UTC)

Although what you say is absolutely true, that's not what I mean. The heritability of differences between groups is certainly an issue, but it is not relevant to my problem. My problem is more run of the mill. The article stated that IQ was highly heritable within "racial-ethnic" groups. But the sources don't say that, the sources say that IQ is highly heritable in specific self-identified groups that have been measured. So if one of the citations states that the heritability of IQ within a specific "white" set of samples is 80% and the same study states that the heritability of a specific "black" set of samples is 70% we can cite the source as saying that heritability has been measured as high whin these two groups as defined by the sources, but we can't use that source to support a sentence in the article that claims that heritability of IQ is generally high within all "racial-ethnic groups". That's what I mean. I think this has been sorted by use of the word "population" and clarification by saying that in "race and IQ" research these populations are nearly always self identifying "white" and "black" populations. Alun (talk) 06:14, 21 October 2009 (UTC)

Overview outline

Here's an outline. I'll call that my first draft and wait for comments before extending it to other nature/nurture hypotheses or other topics. --Distributivejustice (talk) 23:10, 17 October 2009 (UTC)

• Intelligence and IQ are complex.
  • IQ tests measure an important but not complete subset of intelligence.
  • IQ tests are themselves complex, measuring a variety of constructs, including the general intelligence factor g and other more specific mental abilities.
• Racial groups aren't homogeneous
  • Racial-ethnic groups are self-defined.
  • Groups such as "Asians" and "Hispanics" have complex and varied ancestral and ethnic origins.
• IQ scores vary within and between racial-ethnic groups.
  • The range of variation within each group is greater than the average differences between groups.
  • The "structure" of IQ differences between groups may be more complex than can be captured by simple average differences in overall IQ scores. Several long-standing questions are whether black-white IQ differences are due to differences in g, whether other specific mental abilities vary between groups, and whether IQ differences are the result of factors other than differences in ability (e.g. test bias).
• According to the APA, no one knows the cause of between-group differences in IQ. While various explanation have been proposed, none is generally accepted.
  • Jensen argues that a high heritability of intelligence within groups, combined with a variety of indirect empirical evidence, impose strong constraints on the plausibility of an entirely environmental explanation of black-white differences in IQ, and thus favors a hypothesis that both genetic and environmental factors are the cause. The APA concluded that there's no direct evidence supporting a genetic hypothesis.
  • Lewontin argues that Jensen's argument misses the fact that factors which vary between groups but not within groups, such as the effects of racism, caste and culture, can explain a large difference in average IQ regardless of the within-group heritability of IQ. The APA concluded that factors of caste and culture may be appropriate explanations, but at present there is no scientific answer as to whether the environmental difference between groups meet the criteria established by Lewontin.
  • Flynn rejects Lewontin's solution to Jensen's argument as untenable. Flynn discovered that average IQ scores can improve substantially over time despite the high heritability of IQ. Flynn proposes that a complex interaction between genes and environment is the cause of group differences in IQ, making environmental remediation of these differences possible through environmental intervention. The APA concluded that they cannot exclude the possibility that the environmental factors responsible for the Flynn effect are also responsible for between group differences.

---

At first blush, I would say that this is an overly complex overview, looking more like an article outline than a summary of our current understanding. I certainly think all the details you mention should be included in the article though. Maybe the overview is the proper place, but I'm a bit hesitant to include a he-said/she-said in an overview section. Aprock (talk) 00:25, 18 October 2009 (UTC)

Okay, I think that's a good point. It seems that the text from bullet 1 to bullet 3 is all consensus statement and the lead-in to 4 is also. Then the rest of 4 belongs elsewhere. How does that sound? That would be the first four paragraphs below. --Distributivejustice (talk) 00:56, 18 October 2009 (UTC)

Full prose draft

Intelligence is a complex concept. IQ tests measure an important but not complete subset of intelligence. IQ tests are themselves complex, measuring a variety of constructs, including the general intelligence factor g and other more specific mental abilities (such as spatial reasoning).

Racial groups are not internally homogeneous. Racial-ethnic groups are defined and self-defined based on ancestry and ethnic origins. Groups such as "Asians" and "Hispanics" have complex and varied ancestral and ethnic origins.

IQ scores vary within and between racial-ethnic groups. The average scores of African Americans are conventionally reported to be 15 points below those of Whites. The range of variation within each group is greater than the average differences between groups, resulting in overlapping distribution of overall IQ scores. The "structure" of IQ differences between groups may be more complex than can be captured by simple average differences in overall IQ scores. For example, Chinese and Japanese Americans tend to score higher on tests of spatial ability. Several long-standing questions are whether black-white IQ differences are due to differences in g, whether other specific mental abilities vary between groups, and whether IQ differences are the result of factors other than differences in real ability (e.g. test bias).

According to an APA consensus statement, no one knows the cause of between-group differences in IQ. While various explanation have been proposed, none is generally accepted. IQ tests do not appear to appear to be biased against African Americans, and simple models in which socieoeconomic differences directly cause IQ differences do not work.

Jensen argued that a high heritability of intelligence within groups, combined with a variety of indirect empirical evidence, impose strong constraints on the plausibility of an entirely environmental explanation of black-white differences in IQ, and thus Jensen favored a hypothesis that both genetic and environmental factors are the cause. The APA consensus statement concluded that there's no direct evidence supporting a genetic hypothesis.

Lewontin argued that Jensen's argument misses the fact that factors which vary between groups but not within groups, such as the effects of racism, caste and culture, can explain a large difference in average IQ regardless of the within-group heritability of IQ. The APA concluded that factors of caste and culture may be appropriate explanations, but there is no scientific answer as to whether the environmental difference between groups meet the criteria established by Lewontin.

Flynn rejected Lewontin's solution to Jensen's argument as untenable. Flynn discovered that average IQ scores can improve substantially over a few decades despite the high heritability of IQ. Flynn proposed that a complex interaction between genes and environment is the cause of group differences in IQ, making environmental remediation of these differences possible through environmental intervention. The APA concluded that they cannot exclude the possibility that the environmental factors responsible for the Flynn effect are also responsible for between group differences.

---

here's a full prose implementation of that outline. --Distributivejustice (talk) 00:35, 18 October 2009 (UTC)

This looks pretty good to me; it's definitely better than the outline T34CH came up with. As you pointed out from your literature review, the consensus among researchers is that there is not enough evidence to conclude one way or another about whether genetics contribute to the IQ difference, so any "outline" or statement of "academic consensus" in the article should reflect that consensus.

Something I wonder, though, is whether having an "outline" that's separate from the lead section is necessary at all. The lead section is itself an outline, and having both of them seems like it might be redundant. I don't have a strong opinion about this, though. Do the other editors here think that having both a lead section and an outline is necessary? --Captain Occam (talk) 06:27, 18 October 2009 (UTC)

I object to some of this outline. For example, the APA statement does say that an environmental explanation is entirely possible, but that it hasn't been found at this point. It also says that whatever little evidence there is does not support the genetic hypothesis. That's a far cry from saying that there isn't enough evidence to conclude one way or another. Definitely, T34CH's outline is much closer to waht the consensus references say.--Ramdrake (talk) 06:51, 18 October 2009 (UTC)

I share Ramdrake's concern and appreciation of T34CH's outline. Slrubenstein | Talk 06:59, 18 October 2009 (UTC)

"It also says that whatever little evidence there is does not support the genetic hypothesis. That's a far cry from saying that there isn't enough evidence to conclude one way or another."

As I understand this, Distributivejustice's conclusion that this is the consensus was not based only on the APA statement, but also on the literature review he conducted which was discussed earlier. --Captain Occam (talk) 07:16, 18 October 2009 (UTC)

I opened a section on what the report says below. --Distributivejustice (talk) 08:52, 18 October 2009 (UTC)

Reading the prose, I have to say that it is exactly the sort of thing to avoid in an overview. It comes across as a mish-mash of facts and numbers without being a clear summary. What might be the main sentence of the overview appears in paragraph four, and I'm not sure that it actually represents current academic work. Aprock (talk) 08:27, 18 October 2009 (UTC)

Is the aim of the overview (http://en.wikipedia.org/wiki/Race_and_intelligence#Overview) to present a primer to bring readers up to speed on background or to summarize the rest of the article? --Distributivejustice (talk) 08:52, 18 October 2009 (UTC)

No. It is to give an overview of our current understanding of the relationship between race and intelligence. Simply said, what are the current conclusions. Aprock (talk) 08:57, 18 October 2009 (UTC)

This is an issue I was starting to wonder about... the overview should not take the place of the lead. We could expand the lead a bit, but "overview" was just someone's rephrasing of "academic consensus". Maybe we need a better section title. T34CH (talk) 17:58, 19 October 2009 (UTC)

more discussion of outline

I appreciate the thoughtful work that everyone here has put into this. I have an alternative view. I hope people will give it some thought. I am not by the way rejecting out of hand earlier proposals, but if we want to work towards a consensus I hope this will be considered. Note: I am deliberately vague in some areas because other proposals cover some areas so nicely -perhaps a compromisxe version would combine parts of this with parts of earlier proposals:

• race and IQ are complex
  • one reason is that race is studied principally by anthropologists and sociologists, and IQ is studied principally by psychologists. Researchers in different disciplines may speak past one another, or misunderstsand one another
  • Another reason is that this debate rages most strongly in the US, and much of the data used comes from the US. Some notable attempts to internationalize or universalize the debate have been controversial
  • a third reason is that measurements or models for variation within groups are sometimes used in discussions of variations between groups
• Race
  • The mainstream view among anthropologists and sociologists is that race is a social construction
    • on IQ tests and experiments, race is usually self-defined
    • sociologists and anthropologists have discussed a range of sociological and historical forces shaping how people identify the race of others and ultimately themselves
    • the social processes through which races have been constructed means that they are often defined differently in different countries
    • and that they are heterogeneous
  • among sociologists and anthropologists, a fringe view view of races identifies them with populations
  • in the general public, race is quite often confused with population
• IQ scores

well, here I would largely defer to the work others have done above. Slrubenstein | Talk 06:55, 18 October 2009 (UTC)

First question: is this a suggestion for an article outline or an overview section outline? Second: I don't feel qualified to make an assessment without reading further. Can you suggest some citations? --Distributivejustice (talk) 07:39, 18 October 2009 (UTC)

T34CH

I've requested that you please discuss your rather significant changes as per WP:BRD. Instead of discussing, you reverted me and decide to go ahead with your edits. There are ongoing discussions regarding nearly every aspect of this article, and being disruptive is not very helpful. Everyone let you get away with the huge 30+ edit streak of October 17th, but really, the same rules apply to you as to everyone else. Please discuss the changes you want to make beforehand. Thanks, --Aryaman (talk) 20:52, 20 October 2009 (UTC)

Yeah, like I said I wasn't done yet. Now which changes did you have a problem with? T34CH (talk) 20:55, 20 October 2009 (UTC)

I didn't know the world stopped for you, T34CH. As you've mentioned in earlier discussions ("When I get a chance I'm going to work on a rewrite of the whole article, and hopefully get consensus to rename it to someting more meaningful and discriptive"[2]), your involvement in this article is of a rather subversive nature. I don't think giving free reign to someone who is admittedly aiming to override consensus is a wise course of action given this article's history.

What don't I like about your recent edits? Your POV is obvious, your language is unclear, and your omissions are embarassing.[3] I'm all for improving this article, but I'd like to see more editors involved. I won't revert you again, as I know how eager you are to get other editors blocked, but I certainly wouldn't complain if another editor took it upon themselves to do the same. --Aryaman (talk) 21:30, 20 October 2009 (UTC)

You know, you have a strange habit of avoiding specifics. It's a little sad that it took you almost an hour to come up with nothing. If you have problem with something, say it. If you can't point to something (and let's not start with the type of fallacies you tried to pull off above but got called out for by other editors), then stop hampering genuine efforts to improve the article. T34CH (talk) 21:34, 20 October 2009 (UTC)

An hour? Are you looking at some clock somewhere? LOL. Like I said, T34CH, the world doesn't stop for you. Well, rather than sit here spelling it out, I suppose you won't object to my directly editing the article in the same spirit. --Aryaman (talk) 22:10, 20 October 2009 (UTC)

I fail to see a difference in tone from this edit. If you can't explain what you don't like about my edits at least explain what you like better about yours. T34CH (talk) 22:55, 20 October 2009 (UTC)

That, dear T34CH, is the very reason why I think you should keep from editing this article as much as possible. You simply don't see the difference. But, trust me: It's there, in a very big way. Am I being too harsh? Perhaps. Perhaps you mean to say that you don't see that big of a difference in POV. Assuming that you understand yourself to be neutral, then I take that as a compliment. I'm not here to push a POV. I'm here to try and make sure that no one pushes a POV, and that positive, constructive and informative changes are made. My last edit to this article is, I believe, solid proof that I have no "ulterior motives" in participating here. I could easily edit this whole article in the same manner, but it is necessary to get consensus first, or else everything devolves into needless reverts, bickering, etc. For example, I would delete the whole second paragraph of the lead, as the current first paragraph makes it entirely redundant and superfluous. But just image what would happen if I did that. ANI complaint, RfC and possibly the threat of a topic block thrown in for good measure. Thanks, but no thanks. --Aryaman (talk) 23:08, 20 October 2009 (UTC)

I see this is difficult for you. Let's start from the easiest method. Take the one dif that bothers you the most and explain in excruciating detail what's wrong with it. T34CH (talk) 02:19, 21 October 2009 (UTC)

"Overview" proposal text

Editors have voiced several concerns regarding to current "Overview" section, myself included. In light of the immediately preceding discussion, I would like to propose a new text for this section, based upon the summary found in Neisser et al. (1996).

Psychometric testing, despite being one of the most fruitful approaches to studying intelligence, has yet to produce answers to many questions regarding intelligence. Though psychometricians have devised ways to measure the distinct yet intercorrelated abilities believed to play an important role in the development of intelligence, the correlations between those abilities remains unclear. As intelligence test scores correlate moderately well with educational measures, it is apparent that such tests measure important skills. However, educational achievement is not primarily determined by intelligence, though intelligence test scores do correlate significantly with occupational status later in life.

While both genetic and environmental variables are involved in the manifestation of intelligence, the role of genetics has been shown to increase in importance with age. Why this happens in not yet understood, and the question as to what role the environment plays in this increase remains unanswered. Nonetheless, there are several important environmental factors which are known to affect the development of intelligence, such as formal education and general health. The much-discussed "Flynn effect", which refers to the striking worldwide mean IQ increase of 15+ points over the last 50 years, may be the result of similar environmental factors such as improved nutrition, cultural changes, improvement in the administration of tests, changes in educational practices or some other hitherto unrecognized factor.

As the measured differences in intelligence between various ethnic groups is the result of complex patterns, any conclusions which require broad generalizations run the risk of oversimplifying the issue as well as misrepresenting the available data. At the same time, intelligence test scores in some minority populations are reasonably good indicators of educational achievement levels in later life. The long-standing 15+/- point difference between the intelligence test scores of African Americans and White Americans, though it may have narrowed somewhat in recent years, remains unaccounted for despite proposed explanations claiming systematic bias, differences in culture or socio-economic status, or genetics as the underlying cause.

I feel that this gives a solid and balanced summary of Neisser's findings, and could serve as a very good overview of the issues involved. Please comment or make suggestions for change below. --Aryaman (talk) 02:20, 19 October 2009 (UTC)

Major issue: this totally ignores the issue of "race" in the "race and intelligence" debate. Many have pointed to the fact that causation of a biological/psychometric phenomenon (the IQ gap) by a social construct (race) is not possible and have thus completely rejected the "race and intelligence" studies. That part would come from anthropologists, typically. The review by Lieberman (an anthropologist) [4] and/or the one by Sternberg (a psychometrician) [5] would be adequate. Also, here are some links to the AAA's statement on race and intelligence [6] and the AAA's statement on race [7]. I believe once all these viewpoints become integrated in the overview, we'll have a much better rounded overview.--Ramdrake (talk) 07:52, 19 October 2009 (UTC)

"Totally ignores"? The whole of the third paragraph deals with it. As for incorporating other "official" statements, I think the one on "Biological Aspects of Race" from the AAPA is a bit more on-topic, even than the AAA statement on "Race and Intelligence". The latter is a minute (3 sentence) social policy piece pleading to "the academy, our political leaders and our communities" to reject the notion of race. Fine, but I'm more interested in what the experts think, not in how the AAA in its role as a pseudo-political instrument wants to direct social policy. The AAPA piece has substantially more "meat" to it, though I think we have to ask whether this article is going to benefit from its inclusion. The question as to whether or not the concept of "race" in reference to intelligence is a valid one which can be substantiated by data is one best left answered by the people who measure intelligence. The opinion of an anthropologist is going to be dependent upon the data delivered by fields such as behavioral genetics, not vice versa. That's the key point. Anyway, if we look at the AAPA statement, approximately 3 of their 11 points might be relevant to this discussion. I list them below:

§2. "Biological difference between human beings reflect both hereditary factors and the influence of natural and social environments. In most cases, these differences are due to the interaction of both. The degree to which environment or heredity affects any particular traits varies greatly." (pg. 714)

§10. "There is no necessary concordance between biological characteristics and culturally defined groups. On every continent, there are diverse populations that differ in language, economy, and culture. There is no national, religious, linguistic or cultural group or economic class that constitutes a race. However, human beings who speak the same language and share the same culture frequently select each other as mates, with the result that there is often some degree of correspondence between the distribution of physical traits on the one hand and that of linguistic and cultural traits on the other. But there is no causal linkage between these physical and behavioral traits, and therefore it is not justifiable to attribute cultural characteristics to genetic inheritances." (pg. 715)

§11. "Physical, cultural and social environments influence the behavioral differences among among individuals in society. Although heredity influences the behavioral variability of individuals within a given population, it does not affect the ability of any such population to function in a given social setting. The genetic capacity for intellectual development is one of the biological traits of our species essential for its survival. The genetic capacity is known to differ among individuals. The people of the world today appear to possess equal biological potential for assimilating any human culture. Racist political doctrines find no foundation in scientific knowledge concerning modern or past human populations." (pg. 715)

§2 tells us pretty much the same thing as what we learned from Neisser et al.

§10 is perhaps tangentially relevant, but really addresses the question as to any "racial" basis for elements of culture (language, economy, political structure, etc.) which, of course, there is none (the AAPA is simply stating the obvious here, as no one is arguing that such things could have a basis in biology). At the same time, it's trying to avoid saying that the common conception of "race" does, to some degree, correspond with the observed physical variation found in human populations - which is commonly accepted - but doesn't quite want to say it, as that's too close for comfort to sounding like there actually are some biological distinctions between "races" - which there are, though the whole point here is to de-emphasize them to the point of meaninglessness. Fair enough. I can see what the AAPA is trying to accomplish, and as a social policy, I can recognize its value.

§11 is the closest this statement comes to our topic. Interestingly, it describes intelligence as "one of the biological traits of our species". Did they screw up a little here? I'm not going to harp on it. Notice that they simply skirt the issue, though, without getting right down to it. But really, what should we expect from anthropologists? They've taken the "right" overall position, which is a vast improvement over the "science" of the late 19th century, and I suppose they are to be commended on that point.

Can the overview be improved through the inclusion of this kind of material? I highly doubt it. But I submit this to the discussion of the other editors. --Aryaman (talk) 13:29, 19 October 2009 (UTC)

Please remember that the nominal subject of this article is Race and Intelligence. While psychometricians are admittedly experts on questions of intelligence, anthropologists are the acknowledged experts on questions of race. It wouldn't do to let psychometricians define what a race is. Also, I would like to caution you again about second-guessing what expert are trying to say or not to say, just as you did for your interpretation of paragraph 10 of the AAPA statement.--Ramdrake (talk) 14:38, 19 October 2009 (UTC)

Ramdrake, I'm not doubting that anthropologists are the experts on race. I am doubting, however, whether they are to be consulted as to the role of race in matters pertaining to the measurement of intelligence. I have no problem with referring to their position if we're going to make a broad statement regarding race - though we have to wonder if doing so is really in the interest of the article. But to pull in a statement such as that made in §11 - which deals specifically with race and intelligence - is not only of questionable propriety but also of questionable value, seeing as it provides a less-differentiated view than that of Neisser et al. ("The genetic capacity for intellectual development is one of the biological traits of our species essential for its survival." - not a statement I think we want introduced to the overview, do you?)

As far as my "second guessing" goes - thanks for your well-intentioned concern, but it's my opinion on the value of that material. Other editors can see that, and I'm not distorting the text. The fact that I view it critically and within the sociopolitical climate in which it was conceived and composed does not make it necessary for you to "caution" me. I'm fully capable of keeping my opinion out of the article itself. And again, I don't see how including anything from either the AAA or the AAPA is going to give us a better introduction to the reigning consensus than we could already have from Neisser. Thanks, --Aryaman (talk) 14:53, 19 October 2009 (UTC)

As I said, this article is entitled "Race and intelligence". If we skip the opinion of the race experts (the anthropologists) on this subject, we are presenting a one-sided view of the subject. I cannot agree to such a slanted presentation, and I believe you'll find that several other editors will share my position. As is, I oppose your version of the overview as incomplete and somewhat slanted.--Ramdrake (talk) 15:00, 19 October 2009 (UTC)

I'm not consciously trying to slant this one way or the other. If you can find a way to work in the AAPA's statements regarding race and intelligence, then by all means, please do so. But I think you'll find that, rather than improving things, it will make things less clear. Read exactly what it says in §11. "The genetic capacity for intellectual development is one of the biological traits of our species essential for its survival. The genetic capacity is known to differ among individuals." Though I think there's a little too much "genetic capacity" and "biological trait" in that statement, I can see it being reflected in Neisser. Then they follow up with "The people of the world today appear to possess equal biological potential for assimilating any human culture." Well, that's fine too, I guess, as no one but the racists would disagree. (Though, what's with the "appear to possess"? Is the AAPA saying there is any shred of doubt concerning this?) But why aren't they saying "the people of the world possess equal biological potential for intellectual development" or "there are no biological or genetic reasons which can account for differences in intelligence between racial or ethnic groups" or "biology plays no role in the differences in intelligence between racial or ethnic groups"? That's what we'd like them to say. But that would not quite reflect the current consensus based upon psychometric studies, and I can only assume the folks at the AAPA knew their material well enough to avoid making such a statement. Does it bother you to describe this statement as more of a political instrument than a useful piece of academic literature? I'm sorry, but I can't help viewing it this way. Regardless, as I said, I'm not trying to slant this at all. You are of course invited to make suggestions or emendations which could prevent this turning into another "I don't like your version, so let's just keep what we have". Several editors have voiced concern with the current version, and all are invited to help in resolving the issue. Thanks, --Aryaman (talk) 15:19, 19 October 2009 (UTC)

At a fundamental level, I think it is important to mention that (nearly?) all of the psychometric studies have not been done with respect to genetic race. Aprock (talk) 20:25, 19 October 2009 (UTC)

I note that Aryaman seems really to misunderstand science. She wonder's why the AAPA uses the word "appears" and I wonder if she thinks that means that anthropologists are divided, some believing that many group inequalities in intelligence are caused by genetics. Aryaman, the reason any scientist will use words like "appears" is because science does not deal with absolutes. Even so-called laws (e.g. entropy) are probabalistic statements. Even if all the scientific evidence up to now suggested that atomic particles like neutrons are indivisible and the smallest forms of matter that can exist, a good scientst would say this appears to be so because more data can always lead to a new model. Even if this "fact" sustains all challenges for a hundred years, good scientists will continue to use the word "appear" because science just does not deal with absolute truths. You may wish it to, but my advice is try some religions instead. Then again, you have just made it plain that you do not have a good-faith view owards science. As you say "Does it bother you to describe this statement as more of a political instrument than a useful piece of academic literature? I'm sorry, but I can't help viewing it this way." Well, okay, if this is your bias I think we can all agree that you simply should not be allowed any role in editing the content of this article with regard to what scientists think. Parhaps there is some article here on a topic you do know about, why not edit that one? Slrubenstein | Talk 13:25, 21 October 2009 (UTC)

I really am laughing out loud right now. :) Thanks for providing the sometimes painfully needed comic relief around here, Slrubenstein! --Aryaman (talk) 20:07, 21 October 2009 (UTC)

Lede and overview

Some time ago Slrubenstein asked me to comment here. I have been following what has been going on here and on noticeboards.

• At the moment I find the lede poorly written and unhelpful.
• The overview on the other hand explains fairly well some of the problems/misconceptions in the popular debate in the US, although it is a little shakey on historical context.

In particular the overview highlights (a) problems in the use of intelligence tests as a measure of intelligence (b) problems with the use of the term race, clarified now as a social and cultural construct by anthropologists and sociologists in academia. Historically intelligence tests have been carried out on population groups (the correct neutral term to use, particularly in the lede) and the results interpreted. What is not made clear in the lede is the significant change in the academic approach to both race and intelligence in the century or so since measurements were first made and interpreted. I feel the lede has to be written to be far closer to the 2 caveats on the terms "intelligence" and "race" in the overview and should avoid the use of terms like "racial and ethnic groups", which have not been properly justified by academic references. It should make clear that psychologists like Lynn and Rushton have been heavily criticized for their misuse of statistics in academic journals: their fringe viewpoint should not be given undue weight. It should also make clear that the so-called "ongoing debate" is largely in the US and not elsewhere. Wikipedia is an international online encyclopedia, supposedly based on academic journals and books. It is not a place for US citizens to engage in unscientific navel-gazing.

In summary:

• The lede needs to be completely rewritten: Aryaman has not done a good job in summarising the issues; nor has he exercised care with terminology.

• The overview should be slightly rejigged to provide a more careful historical context for the multiple issues discussed in the article.

Mathsci (talk) 14:06, 21 October 2009 (UTC)

I've had a go at the lede. Not sure if it helped any. Feedback most welcome.--Ramdrake (talk) 14:28, 21 October 2009 (UTC)

Thanks. I think these changes solved many of the problems with the lede. The cautious tone of the current lede is quite appropriate. I made a small change, substituting "population groups" for "racial and ethnic groups". Mathsci (talk) 17:45, 21 October 2009 (UTC)

It's looking better now. I wonder what we can do with the "overview" vs. the "lead". The lead is supposed to be the overview, but we had tried having a section on "academic consensus" which was then objected to and transformed. Aprock, are you still thinking that these should be different sections, or should we just merge them? Distributivejustice, I think that the section you had mentioned about the different POV's would go well in place of the overview, if the current overview is merged into the lead. Not an in-depth examination of them, but a quick "overview" that later sections could refer to for context (again with the caveat that I think we really need to avoid a long list of supporters' names under each POV). T34CH (talk) 19:31, 21 October 2009 (UTC)

Mediation, anyone?

Seeing as the task of improving the article seems to get bogged down with personal conflicts on either side, I'd like to suggest submitting this dispute to informal mediation. I believe that a neutral third party may be helpful in attaining a properly balanced article. However, I would suggest a mediator who is himself versed in the sciences to help us resolve this dispute.--Ramdrake (talk) 14:10, 21 October 2009 (UTC)

Actually, I would be willing to leave this article alone for good if the entire batch of currently involved editors were to agree to the same thing, particularly T34CH, Ramdrake, Slrubenstein, Alun and Mathsci. (Though, Alun does appear to be editing in good faith.) No need for "mediation". This article is being systematically manipulated by a group of editors with a very long history of cooperation on related subjects, all of them take the same POV on these matters, and most of them use the very same tactics when it comes to disrupting another editors' work, best typified by Ramdrake's "BRD" approach, i.e. revert first, demand discussion, wear down your "opponent" on the talk page by refusing to see any merit in the point s/he is making, and if s/he rears up, report it to the noticeboards. That is the central problem which has dogged this article from the beginning, and will continue to dog it as long as they are involved. Now, I can surmise easily enough that the reactions would be: (1) "Of course we take the same POV - we're reporting what the mainstream, credible sources say. You, on the other hand, are distorting everything, trying to push fringe material, and generally inept enough to deserve a topic ban"; (2) "WP:BRD is part of Wikipedia policy. If you don't like it, go edit some blog somewhere"; (3) "We've all heard the 'cabal' theory before - utter nonsense. The fact that we edit the same articles, share the same views, call each other in for support in arguments, fabricate consensus, share our email addresses for off-wiki discussion, and collectively hound anyone who stands in our way does not make us a "cabal". How ridiculous!" So, spare me the rejoinders. As for myself, I see no light at the end of this particular tunnel. I don't care enough about Wikipedia to do what would be necessary to clean this up, nor do I have the clout. I admire users such as Captain Occam for having the determination to stick up to the lot of you, even in the face of multiple blocks and hoots of ridicule. If you really want to get a mediator, then get dab involved. We have had our conflicts over things in the past, but I have to hand it to him: that man knows how to build an encyclopedia. He upholds policy at every turn, is brutally honest in his evaluations, is open to justified criticism, can admit when he makes a mistake, and is possibly one of the best, most productive editors in the whole project. Otherwise, my advice to the rest of you is to watch out for the cherry-picked "mediator". --Aryaman (talk) 20:51, 21 October 2009 (UTC)

So I take it you decline mediation?--Ramdrake (talk) 20:57, 21 October 2009 (UTC)

Varoon Arya (talk · contribs), what you have written is unexpected and perplexing. Please stick to content rather than spouting conspiracy theories about other editors. You do not seem to be assuming good faith at the moment. Mathsci (talk) 00:04, 22 October 2009 (UTC)

Worldwide View tag

I removed the worldwide view tag. Ramdrake added it back in and noted "Still needs more worldwide literature." I am fairly familiar with the literature on this topic. No important "worldwide literature" is missing from the article. So, the tag should go unless Ramdrake can cite some specific books and articles that are missing. (He does not need to add them to the article himself. I, and others, will do that. But he can't simply assert that something is missing without providing specific examples.) David.Kane (talk) 13:13, 11 October 2009 (UTC)

It's been over a day since you asked about this, and nobody has provided an example of worldwide literature that's missing from the article, so I think the tag can be removed. --Captain Occam (talk) 22:11, 12 October 2009 (UTC)

May I remind you that Wikipedia is under no deadline? I'm still assembling refs on this.--Ramdrake (talk) 22:17, 12 October 2009 (UTC)

Most elements of Wikipedia are under no deadline, but WP:TAGGING states that tags shouldn't be left on articles indefinitely if nobody is providinig a justification for their inclusion. "If an argument on the talk page has been made as to the reason for the tag, but someone still feels that the tag is inappropriate, he or she should explain the reasoning on the talk page. If there is no reply within a reasonable amount of time (a few days), the tag can be removed." --Captain Occam (talk) 22:41, 12 October 2009 (UTC)

If Ramdrake says he is assembling refs, it is important for us to wp:AFG that this is the case. The above response can be perceived as wp:Wikilawyering. T34CH (talk) 23:17, 12 October 2009 (UTC)

Fair enough. But in the interest of fairness: tagging without actionable criticism can be perceived as tag bombing to promote a point of view. If Ramdrake has a case - and I'm assuming he does - he can easily rectify the matter within a few days. --Aryaman (talk) 23:25, 12 October 2009 (UTC)

From your edit, it seems you are proposing to remove the tag now and replace it later. Can you wait a few days and AGF as T34CH asked?--Ramdrake (talk) 23:28, 12 October 2009 (UTC)

(Assuming the question is directed at me:) If a person thinks a tag is needed, then, assuming good faith, they have a good reason for adding it. In this case, you said that the literature was not "worldwide" enough. It is safe to assume you're basing that on some kind of knowledge as opposed to a hunch or gut instinct, i.e. you personally know of some literature which is not included at present. I think the above editors are simply indicating that, if you know of literature which is missing, you should add it. If you can't, then one must assume that you were mistaken, and such a one would be justified in removing the tag. Of course, you should be given ample time to do so. But a few days really is enough in my opinion. Thanks, --Aryaman (talk) 23:53, 12 October 2009 (UTC)

It has been 4 days, so I have removed the tag. If anyone would like to revisit this issue, please feel free to. Also, anyone with knowledge of research articles (whether from outside the US or not) that have not been included should feel free to add them. David.Kane (talk) 12:21, 15 October 2009 (UTC)

Aprock, I notice you've just put the tag back. How long do you think we need to wait for its presence to be justified before it can be removed? WP:TAGGING says that "a few days" is a reasonable amount of time. It certainly shouldn't be there without justification for more than a week.

Fairly soon now, this discussion will become so old that it will be moved to the discussion archives. If Ramdrake has not replied by that point, it will become too late for him to do so, because archives aren't supposed to be modified. I hope you'll agree it's not reasonable that after a tag is challenged, it would continue to remain in the article for so long that the person who wants it there fails to justify it within the maximum amount of time that Wikipedia's format allows for replies on talk pages. --Captain Occam (talk) 00:58, 16 October 2009 (UTC)

I went through my old list of WW references. Turns out that many of them were integrated since in the article. However, a friend of mine was kind enough to send me these also:

[8][9][10][11][12][13][14]

Sorry for the delay, but I'm in the hospital most days these last couple of weeks, dealing with ESRF and related issues. Therefore, I don't have as much time for this as I used to.--Ramdrake (talk) 01:04, 16 October 2009 (UTC)

Thanks for the references. Good luck with your health issues. I will review these citations. If any of them seem useful and relevant, I will add them to the article. (But note that we already have 100 or so references, so I won't be adding them to just to add them. And, of course, other editors are free to add them (or delete them) as they see fit.) Once I have done so, I will delete the tag. David.Kane (talk) 01:21, 16 October 2009 (UTC)

David, the question isn't whether all the refs we find are in the article or if there are too many refs already (not really a valid arguing point in any situation that I can think of). The question is whether the topic is addressed from a worldwide view or not. It would seem to me that the case has been made that the article is US centric. That means the article should be developed to remove this bias. Simply integrating the refs above (presumably wherever they fit) would not solve that problem. The tag currently serves to inform readers that they should not assume the article is fully representative of the world view on issues surrounding race and intelligence. T34CH (talk) 18:26, 16 October 2009 (UTC)

Relative to what ideal? Using such a standard, one could reasonably apply that tag to almost every article in Wikipedia. Now, obviously, this being English Wikipedia, there is a bias toward sources in English. I believe that this article is "representative" of world views on this topic. To show that it is not, you must provide examples of specific views that it does not represent (and make a case that this are worth including in an article that is already on the long side). Am I unreasonable to be concerned about WP:Tag bombing in this case? Note also WP:OVERTAGGING: "It is very rare that more than two or three tags are needed, even on the worst articles." If the Worldwide tag were the only one for this article, it might be reasonable to keep it. But there are 3 others, all more important/relevant, in my view. David.Kane (talk) 15:45, 17 October 2009 (UTC)

Let's consider a specific (hypothetical) example> I don't know Chinese, but I am happy to believe that there are articles in the Chinese academic literature about race and intelligence, articles that address the issue from a non-US centric focus, perhaps comparing the intelligence of Chinese and Koreans. If someone reads Chinese, knows this article, and wants to use the information in this article with an appropriate citation, then great! More power to them. But the article does not merit a Worldwide tag until this happy event (for the academic literature in China, Japan, Malaysia, Brazil . . .) occurs. David.Kane (talk) 15:49, 17 October 2009 (UTC)

It has been 10 days since this discussion started and 4 days since my last comment. Does anyone object to removing the Worldwide tag? David.Kane (talk) 12:33, 21 October 2009 (UTC)

10 days in which a lot of other issues have been discussed. Why not wait until we all come around to that issue rather than to keep pressing this single issue? I don't see that this issue is any more important than the rest.--Ramdrake (talk) 15:30, 21 October 2009 (UTC)

It does not matter whether you (or I) see this issue as more or less important than any other issue. Is this tag justified or is it not? If you believe that it is, then you should engage in discussion about it. Tags must be defended. In particular, you must provide an argument for overruling standard Wikipedia practice that "It is very rare that more than two or three tags are needed, even on the worst articles." If you do not have the time and/or inclination to debate this now, no worries. We can simply remove it and come back to the topic a week or a month or a year from now. David.Kane (talk) 19:39, 21 October 2009 (UTC)

I just strongly object to this philosophy of "work on it, or else..." The case has been made that this article lacks a worldwide view. Currently, editors are engaged resolving other issues. Your demand that this issue be adressed now or else you're going to remove the tag doesn't strike me as appropriately constructive behaviour, especially under the current circumstances.--Ramdrake (talk) 19:47, 21 October 2009 (UTC)

I do not mean to argue "work on it, or else..." I believe that this tag is not necessary. To keep it, you should explain why it is a more critical issue than the other three tags on the article or you should explain why the standard Wikipedia practice --- "It is very rare that more than two or three tags are needed, even on the worst articles." --- does not apply. David.Kane (talk) 12:31, 22 October 2009 (UTC)

I didn't say it is more critical, but I'm saying it is roughly as critical. And what you are quoting is just a guideline. There are thousands of articles out there with four or more tags.--Ramdrake (talk) 12:14, 23 October 2009 (UTC)

Another break

I repeat my problems, include my suggested change, and Arya's suggested compomise (thanks for that Arya, your suggestion about what the sentence actually means was far better than what the original sentence said):

Original text:

The consensus among intelligence researchers is that IQ differences between individuals of the same racial-ethnic group reflect real, functionally and socially significant, and substantially heritable differences in intelligence.

My problems.

• "The consensus"- which source says it's a consensus?
• "IQ differences between individuals"- i.e. variation
• "of the same racial-ethnic group"- at least one of the sources doesn't say this, "Intelligence knowns and unknowns" doesn't say this, it says "It is clear (Section III) that genes make a substantial contribution to individual differences in intelligence test scores, at least in the white population." which is not the same at all.[15] BTW the article also states, as I do above, that "So defined, heritability (h2) can and does vary from one population to another. In the case of IQ, h2 is markedly lower for children (about .45) than for adults (about .75)" and "The value of h2 can change if the distribution of environments (or genes) in the population is substantially altered."
• "reflect real, functionally and socially significant"- I don't know what this part is supposed to be saying. I don't see why it can't just say "reflect".
• "and substantially heritable differences in intelligence"- this is just a convoluted way of saying that the differences mentioned earlier in the sentence have a substantial genetic component, right? But it's written in a confusing manner.

My suggestion:

It has been estimated that genes contribute substantially (40-80%) to IQ differences between individuals within some human groups.

I don't think I am in any way changing the meaning of the sentence, only making it more precise and accurate. But Arya has also offered an alternative, which might provide the basis for a compromise.

Arya's suggestion:

the consensus among experts is that IQ differences between individuals of the same racial-ethnic group are significantly heritable

I can live with Arya's suggestion with some minor changes, how about

My version of Arya's proposed change, I include a footnote to clarify the nature of heritability:

There is a consensus that IQ differences between individuals within the same population (usually self identified "Black" or "White" in studies of race and intelligence) are significantly heritable.

• Footnote reads

How does that sound? Cheers Alun (talk) 07:56, 16 October 2009 (UTC)

In his book "Intelligence" Nathan Brody states:

There is little or no doubt that IQ is a heritable trait among white individuals. There is considerably less data on the heritability of IQ in black samples. And, these studies do not provide ideal data for an analysis of heritability.[16]

Which is what others also say. Alun (talk) 09:05, 16 October 2009 (UTC)

I claimed Arya has a racialist agenda, because even when she provides a quote where heritability is not about race, she insists that we must explain it in terms of race. The key line in the lengthy quote above is, "Within-group heritability, because it is tied to the particular environment and genetic variation existing within the group, has little relevance to the causes of between-group variation." Arya challenges us to summarize the point. In fact, Alun did a fine job. Some people may not be sure what "group" means. Well, the solution ot htat is not to delete what Alun wrote, the solution is to add another sentence defining "group." But Alun's sentence is a vast improvement over what someone else wrote because it is clearer and more accurate. Arya provides a long passage where Flynn talks about studies of heritability based on twin-studies. Understandably so. Now, the studies do not make any claims about race, or even about ethnicity (which is not the same thing as race). Why does Arha insist that these are about "racial/ethnic groups" which mixes together two different kinds of identity? The bottom line is in the quote, above: heritability measures the proportion of variation within a group owing to genes. It says nothing, and cannot be twisted to say anything, about diferences between groups. Now, if you want to demand that because heritability is not about "race" then really, we should just remove all discussion of inheritance and genes from this article, okay, that is a position that is at least consistent with the science.

The problem is simple: some people start out believing that because there are some genetic differences between people of different races, genes must explain all differences, or at least one that matters a good deal in our society (and perhaps even moreso to people who would like to write an encyclopedia): inferior or superior intelligence. So they go about looking for anything that they can distort in order to support the argument that yes, some people are just born smarter. The irony is, this is just such a dumb and ignorant position! Anyway, NPOV and all, let's include multiple viws. But heritability remains a measure of variation within a group and NOT between groups, let's at least get the definition right. Slrubenstein | Talk 09:30, 16 October 2009 (UTC)

Believe it or not, I do not want to start a big edit war. I have no objection to the article saying that some intelligence researchers (I see no evidence of a consensus among intelligence researchers, which describes a large number of people doing many different kinds of research) believe that differences in IQ between races is largely explained by genetics. What I do objct to is the word heritability being misrepresented. As long as the proper definition of the word is made clear, I do not object to presenting a view about intelligence differences. Let's not get bogged down in the wrong debate. What characterizes the view of Rushton is not his view (right or wrong) of heritability, but his view (right or wrong) or differences among races. Let's be clear about that. We need to go through thoe sources and make sure that we use only those that specifically say that the difference in IQ scores among races is due to biology. To include citations of twin studies that say that heritability of intelligence is high ... and that do not discuss differences between races ... violates NOR. This is an issue we need to resolve, and we do it simply: articles about heritability that do not mention race support the sentence Alun proposed. Articles on differences between races can support a sentence on the biological cause view. Slrubenstein | Talk 10:41, 16 October 2009 (UTC)

I don't think this sentence is about differences between groups. I think this is about whether the sources say that heritability is high amongst individuals within "racial-ethnic" groups. I don't think the sources say this. I think the sources either say that heritability is high within specific defined groups (e.g. "white" or "black") or they give a general comment about the heritability of intelligence being significant. But here's the thing, even if heritability is high within one group, it can be low in another group, and it's wrong for our article to say that it's "substantial" generically within "racial-ethnic groups" unless that's what the source states.

Interestingly it might be more important to specifically include the word "adult" in this sentence. From a reading of much of the literature, there is no disagreement that heritability is always lower in groups of children than in groups of adults. It therefore seems to me that we need to emphasise, when we talk about a high heritability for intelligence, that this applies to adult populations. I think that's particularly important when we take into account that childhood environment has a massive effect on educational achievement, and that environment is much more important for IQ in children than it is for adults. Effectively it might mean that it's the childhood environment that dictates the average IQ differences. This is certainly the belief of people like Geoffrey Canada, and the results of his Harlem Children's Zone certainly give credibility to the huge effects of childhood environments. Alun (talk) 13:13, 16 October 2009 (UTC)

Slrubenstein, you say that you don’t want to have an edit war over VA’s proposed changes, yet you continue to change this sentence in a way that he’s clearly expressed a problem with? Something isn’t lining up here.

Alun, let’s take a look at this sentence you proposed: "There is a consensus that IQ differences between individuals within the same population (usually self identified "Black" or "White" in studies of race and intelligence) are significantly heritable."

I don’t think this is any better than the sentence as it was originally phrased, but at least it isn’t significantly worse either. One thing I would suggest changing, though, is to leave either the term “intelligence researchers” or “experts” in the sentence—just saying “there is a consensus” doesn’t mean anything if it doesn’t say among whom. I think we should wait for VA’s input about this before the sentence is changed, though. --Captain Occam (talk) 19:47, 16 October 2009 (UTC)

I appreciate your attempt to be more constructive, but I do find your "revert first, ask questions later" approach to be rather rash. I think the sentence above is certainly better than what you keep reverting it to. It's still not clear to me what it is about that sentence which you think is worse than what you keep reverting it to. Could you explain that? Aprock (talk) 19:53, 16 October 2009 (UTC)

The main thing I preferred about the original version of this sentence was that it also mentioned another aspect of the consensus among researchers, which is that IQ is a socially and functionally significant measurement of intelligence. However, now that this is explained by the first paragraph of the “test score difference” section, I suppose that mentioning it in the first sentence of the “genetic factors” section is no longer essential.

I’ve now changed the sentence to what I quoted above, since this is at least better than what Slrubenstein changed it to. I still think we should wait and see what Varoon Arya has to say about this, though. --Captain Occam (talk) 20:05, 16 October 2009 (UTC)

Thank you for clarifying that. Also, thank you for updating the sentence. I don't doubt that this may not be the final version, but it is better than what it was before. Aprock (talk) 20:08, 16 October 2009 (UTC)

"There is a considerable body of research which has shown that differences in IQ measurements in individuals of the same population (typically "white" or "black" in studies correlating race with intelligence) are significantly and substantially heritable. Concordance rates for IQ from studies conducted on both identical and same-sex nonidentical twins indicate a correlation of approximately 80% in adults, and most experts in the study of intelligence view this as a reliable estimation of the heritability of intelligence." -- That would be my suggestion, but I'm not particularly dissatisfied with Occam's change. I think the finding that heritability may vary with age should be mentioned a little later, particularly with a discussion of the role of environment, but I'm not set on that. --Aryaman (talk) 21:00, 16 October 2009 (UTC)

At a fundamental level, I think information like this should be in the article. The main issue here is that this sort of detail really isn't appropriate for the lead paragraph of the section. I think this basic view may be where a lot of the friction is coming from. From my perspective, the intro paragraph of a section should be as straightforward and jargon free as possible. As it stands right now, it's better in that regard, but is still awkward. Aprock (talk) 21:05, 16 October 2009 (UTC)

Well, I don't think it's contested that most experts agree that, in all groups, intelligence is to some degree heritable. No one is arguing for 100% environmental conditioning. So, if you'd rather start off with a general statement such as "There is consensus among experts that intelligence is a partly heritable trait, though the degree of that heritability remains an area of intense research." and then follow with what I or others have suggested, I don't think that would set the house on fire. :) --Aryaman (talk) 21:11, 16 October 2009 (UTC)

This sentence has just been changed again, by Muntuwandi, to an entire paragraph. His new version is pretty different from anything that we've discussed, and I don't believe that it's supported by the sources being used. Since Aprock approved of what I most recently changed this sentence to, and Alun is the one who suggested this change originally, I this consensus probably goes against this newest change.

For these reasons, I think Muntuwandi's change needs to be reverted. If he wishes to replace this sentence with his new paragraph, he'll need to justify to me, Alun, Aprock, and Varoon Arya why his new version is better than the one that the four of us have agreed on. --Captain Occam (talk) 01:56, 17 October 2009 (UTC)

For the record, I agree with this revert. While a rewrite should probably occur (to make it clearer, and ensure that important information into the body of the text) I think it's clear by now that that will have to happen through talk. Aprock (talk) 07:01, 17 October 2009 (UTC)

Agreed. I think anyone involved in this discussion would have made that revert, regardless whether Muntuwandi made it in good faith or not. (Which makes me wonder whether the article couldn't benefit from a temporary "Under development" template while these discussions are ongoing. It gives a clear warning to other editors who may not be aware of the current discussions that they should check the talk page before making any substantial changes. It's worth considering, even if it accomplishes nothing other than to prevent editors just like Muntiwando from becoming frustrated if their well-intentioned edits get reverted.)

"There is overwhelming consensus among experts that intelligence is a partly heritable trait, though the degree of that heritability remains an area of intense investigation. There is a considerable body of research which has shown that differences in IQ measurements in individuals of the same population (typically "white" or "black" in studies correlating race with intelligence) are significantly and substantially heritable. Concordance rates for IQ from studies conducted on both identical and same-sex nonidentical twins indicate a correlation of approximately 80% in adults, and most experts in the study of intelligence view this as a reliable estimation of the heritability of intelligence."

Is this something we could agree on? --Aryaman (talk) 02:11, 17 October 2009 (UTC)

That sounds pretty good as far as content is concerned, but it seems like it might be a little longer than is necessary. Do you think that information could be condensed into two sentences, rather than three? --Captain Occam (talk) 04:43, 17 October 2009 (UTC)

(reset indent) I have made some rather minor changes to the present formulation. First, most of those IQ studies were not aimed at studying the relationship between race and intelligence. Second, if we are to start quoting heritability numbers, we need a good source, such as a review or better (secondary or tertiary source).--Ramdrake (talk) 03:51, 17 October 2009 (UTC)

Arya, I think your suggestion is not a bad starting point for discussion. It covers many of the positions, but I think it does need some work. Here are my problems with it, but let's not get side tracked or draw "red lines in the sand", we're here to cooperate after all.

• "There is overwhelming consensus among experts that intelligence is a partly heritable trait"

You don't need the word "overwhelming", consensus is consensus, the word "overwhelming" is hyperbole and is anyway redundant.

• "the degree of that heritability remains an area of intense investigation"
• "Concordance rates for IQ from studies conducted on both identical and same-sex nonidentical twins indicate a correlation of approximately 80% in adults, and most experts in the study of intelligence view this as a reliable estimation of the heritability of intelligence."

These two passages contradict each other. How can it be an area of "intense study" and also be the settled view of "most experts"? I don't follow this.

• "Concordance rates for IQ from studies conducted on both identical and same-sex nonidentical twins indicate a correlation of approximately 80% in adults, and most experts in the study of intelligence view this as a reliable estimation of the heritability of intelligence."

I'm also a little confused by use of the word correlation here, is this 80% a measure of r², or is it a heritability estimate? That's not clear. If it is 80% heritability, then we should use the word heritability and not correlation. I also think it's extremely important to say which group this 80% estimate was taken from. As I have stated repeatedly, it is not ever valid to claim that an 80% heritability from one set of samples represents a heritability estimate for all populations at all times. Heritability is a population level trait, and so that 80% measured only applies to a specific set of samples drawn from a population (even then there may be some dispute as to whether it represents an unbiased sampling). But I digress. Very important to say which group this 80% is measured from. Indeed if we are going to give this 80%, which is the upper measure, then we should also give some of the lower estimates used by other researchers, I've seen estimates as low as 30%, why don't you want to include these? My feeling is that we should give a range (e.g. 30%-80%). You can't say that "most exerts view this as a reliable estimation of the heritability of intelligence." 80% is an estimate from a specific sample set drawn from a specific population. It is not an estimate of the "heritability of intelligence" it is an estimate of the "heritability of IQ within this sample set". What's the evidence that "most experts" believe that heritability is a trait level measure? I think most experts understand absolutely that heritabilityy is a measure that only applies to a specific set of samples drawn from a specific population.

So I'd say something like this:

"There is a consensus among experts that IQ is partly heritable, though the degree of that heritability remains an area of intense investigation. There is a considerable body of research which has shown that differences in IQ measurements in individuals of the same population (typically self-identified "white" or "black") often have a large genetic component. IQ studies conducted on both identical and same-sex nonidentical twins indicate a heritability of between 30%-80% [or we could say "a heritability of well over 50%..] depending upon the population studied"

I give here a range of 30-80% because that is the range given in "Genetic foundations of human intelligence", Ian J. Deary, W. Johnson, L. M. Houlihan Hum Genet (2009) 126:215–232 doi:10.1007/s00439-009-0655-4. It should be noted that the 30% is probably an estimate for children, but it's not clear in the text. This paper also gives a "general" estimate for the heritability of g as "well over 50% in adulthood", which seems to be a general estimate for heritability for all humans, though it's not clear where they derive this figure from, it is probably an estimate based on a biased set of data. By that I mean a meta-analysis of studies carried out over the last 100 years or so in mostly US or European countries, but as I say it's not clear. I'm happy to use either the 30-80% range, or the "well over 50%" estimate. At least this source does give a general figure and a set of ranges, which I think is more useful that giving a single figure that can only be applied to one population.

What do you think? Alun (talk) 07:17, 21 October 2009 (UTC)

I'm pretty sure the low figures on the range are derived from child testing scores, as you mention, which is why I said "in adults" above. I mentioned earlier that I thought this difference should also be mentioned, but slightly later, as it tends to confuse the fact that heritability in adults is generally well above .50, and as Flynn (2001:155-156) said after his explanation of the two main methods of measuring heritability: "This is why most experts on the issue estimate the heritability of intelligence as approximately of .80, or 80 percent." This was also indicated in Neisser et al., where they mention that the genetic factor increases in importance with age (i.e. heritability estimates increase with the age of the group). If there are considerable differences in heritability (e.g. more than a 5 or 10 point difference which would preclude us from using a blanket "high" descriptor) between same-age groups, then of course this needs to be well-researched and adequately explained. I certainly have no problem with that as long as it's factually accurate. But, based on what I've read so far, I'd hesitate to say .30-.80 as a "range", as this may be conflating group comparison with the age-related phenomenon. --Aryaman (talk) 09:46, 21 October 2009 (UTC)

OK, I follow your logic. What do you think about this then?:

"There is a consensus among experts that IQ is partly heritable, though the degree of that heritability remains an area of intense investigation. There is a considerable body of research which has shown that differences in IQ measurements in individual adults of the same population (typically self-identified "white" or "black") often have a large genetic component. IQ studies conducted on both identical and same-sex nonidentical twins indicate a heritability for adults of well over 50% depending upon the population studied"

I hope we're close to agreeing a wording here. Alun (talk) 11:29, 21 October 2009 (UTC)

Well, this is the part where I ask for references. Flynn gives .80, but does not specify whether this result has been gained from "white" subjects only. My impression is that, while twin studies have been conducted on twins belonging to other racial and ethnic groups, the overwhelming majority of twin studies have been conducted on "white" twins. This may be a blind spot in Flynn, I don't know. But I think it's time we start looking at estimates of heritability derived specifically from non-white groups of comparable age, i.e. studies which compare the relevant results. If we're going to say "well over .50", then we need to back that up - especially if it conflicts with Flynn and others. Do you follow me? As far as the rest is concerned, we're moving in the right general direction. --Aryaman (talk) 20:02, 21 October 2009 (UTC)

You make some good points Arya. We have to think about what we are trying to say. Are we trying to say that heritability of IQ has generally been estimated as "high" in adult humans? (btw are you sure you mean Flynn and not Lynn?). The 80% figure has been bandied about ever since Jensen's 1969 article, and it has been constantly criticised as unrepresentative for several reasons.

• Some say that the data are derived from a homogeneous group of "white" middle class people, with the diversity of even the "white" US community under-represented, and I can probably find cites for that.
• Others have criticised Jensen's overly simplistic method for calculating heritability estimates.
• There are also some very good critiques of the use of monozygotic and dizygotic twins, most notably the observation that monozygotic twins share a great deal more of their environment than dizygotic twins. Indeed I have one paper that estimates that as much as monozygotic 40% of the variation that is usually attributed to monozygotic twins' genes may be explained by their greater shared environment compared to even dizygotic twins. That means that the difference between monozygotic and dizygotic twins is more than just their genes, mothers are more likely to treat monozygotic twins alike than they are to treat dizygotic twins alike. This paper estimates that if we adjust the heritability of IQ based on the shared environment of monozygotic twins, we find that it is as low as 28% (I'll have to double check that figure, but I can find the paper).

Given the amount of criticism of this 80% figure I really don't think we can include it alone and say that this is the accepted figure. It is one estimate of many, and it is the most roundly criticised of them all.

Take a look at this paper "A Twin Study of the Genetics of High Cognitive Ability Selected from 11,000 Twin Pairs in Six Studies from Four Countries" Haworth et al. Behav Genet (2009) 39:359–370 doi:10.1007/s10519-009-9262-3. Now we need to emphasise that this paper takes data from six twin studies from four different countries (Australia, the Netherlands, the United Kingdom and the United States), and there is no way of knowing how representative these samples are. But what they do is to take the top 15% of achievers (i.e. those who do the very best on cognitive tests) and estimate the heritability of intelligence for this top 15% from the four countries. They conclude that

In this first adequately powered analysis of the genetic and environmental etiology of high general cognitive ability (g), defined as the top 15% of the distribution, we find evidence for substantial heritability (0.50 with 95% confidence intervals of 0.41–0.60) and moderate shared environmental influence (0.28, 0.19–0.37).

So their estimate for heritability is 50% and although their samples included adults and children they claim no detectable difference for heritability by age for this high cognitive ability group. In their other paper of this year they give a heritability estimate for the whole sample set, here they do see a difference between children and young adults. In this paper their estimate for young adults is 66% "The heritability of general cognitive ability increases linearly from childhood to young adulthood" Haworth et al. Molecular Psychiatry (2009), 1–9 doi:10.1038/mp.2009.55.

It's fair enough that we cite a paper if we are going to use a figure, but then currently we use citations that didn't really support what we were saying before. The paper I quoted yesterday does support my suggestion of the words "a heritability for adults of well over 50%", I actually quoted it yesterday didn't I. Here it is again: "Genetic foundations of human intelligence", Ian J. Deary, W. Johnson, L. M. Houlihan Hum Genet (2009) 126:215–232 doi:10.1007/s00439-009-0655-4.

"The heritability of g is substantial. It increases from a low value in early childhood of about 30%, to well over 50% in adulthood, which continues into old age."

I think giving a figure that is a range (here well over 50%) is fine, especially when a reliable source gives it, this sort of figure is obviously quoted by experts who are familiar with the state of research in the field. So one can assume that this group of experts are happy to say that heritability is going to be over 50% for any adult group. I think that sounds reasonable. It's also worth noting that this is a review paper, and so attempts to draw in all of the most up to date research in the field.

My feeling is that the "well over 50%" is the least contentious and almost certainly does represent a consensus amongst researchers into cognitive ability because it represents a very broad range of posibilities, which is probably why the authors of this research paper used it.

What do you think? Alun (talk) 06:03, 22 October 2009 (UTC)

I do mean Flynn (I quoted two paragraphs from him one section up on this talk page). While "well above .50" is a bit "loose", it's not inaccurate, and you're probably right that it would be the least contentious. If I were writing this, I would put that phrase ("well over 50%") in the main body, and then add a footnote which gives more specific numbers attributed to researchers such as Flynn, both higher and lower, followed by a solid source which summarizes this part of the discussion, i.e. the various estimates of heritability, provided one can be found. For example:

[1]Deary et al. (2009:215–232). Experts have made various attempts at specifying the degree of heritability, though with little success: Flynn (2001:155-156) cites .80, Expert X (2004:1) cites .yx, and Expert Y (2008) cites .xy. For a general overview, see: GoodReporter (Year:pg).

How would that sound to you? --Aryaman (talk) 12:53, 22 October 2009 (UTC)

Sounds great. BTW can you give me a fuller cite for the Flynn source? You give "Flynn (2001:155-156)" Is this a journal paper or a book or what? I can't find it in his list of publications on his Wiki page (I'm assuming it's James Flynn), and I can't see it cited in the R&I article as it currently stands. Cheers. Alun (talk) 05:45, 23 October 2009 (UTC)

(outdent) Actually, you were completely right and I was completely wrong. I had something like 7 open books on my workspace at once, and had Lynn and Flynn on top of each other. I even had the page numbers from a passage from Flynn. Mea culpa. I apologize. Before my break, I was slightly overworked, both online and off. The section I was referring to comes from: Lynn & Vanhanen (2002:24). The full citation is: Lynn, Richard; Vanhanen, Tatu (2002). IQ and the Wealth of Nations. Westport: Praeger Publishers. ISBN: 0-275-97510-X. As a side note, that passage is followed by:

The heritability of intelligence among children is somewhat lower, probably because there are environmental effects largely from parents acting on children that wear off during adolescence. It is by including the lower heritability estimates derived from children with the estimates of around .80 for adults that some writers put the heritability of intelligence at around .70. The third principle method for estimating the heritability of intelligence is to examine the correlation between the IQs of unrelated children who are adopted and reared in the same families. The magnitude of the environmental impact is expressed in the correlation between the twin pair. The summary of the research by Bouchard (1998) concludes that among children the correlation is 0.22. This is a measure of the environmental contribution, indicating a heritability of .78. Among adults the correlation of 0.04, indicating a heritability of .96.

I'd like to find Bouchard (1998) and see exactly what he says on this, but apparently these high estimates are not only based upon genetically related twin studies. --Aryaman (talk) 14:26, 25 October 2009 (UTC)

No problem, I've done it myself. Thanks for giving the info. Cheers. Alun (talk) 15:46, 25 October 2009 (UTC)

Here's a (42 page long) review paper to check for the state of the art on heritability from the Minnesota folks: http://www3.interscience.wiley.com/journal/101526004/abstract --Distributivejustice (talk) 17:38, 25 October 2009 (UTC)

Thanks for that DJ. I've had a quick glance at it. It more or less supports the "greater than 50%" statement we are discussing. The relative quotes are

Devlin has written critically of the behavioral genetic literature (Devlin et al., 1995, 1997b), so that we may expect him to report a conservative estimate of genetic influence. The most important findings in their article are that: the broad heritability of IQ is about 50%.... the result of this article is to center the debate on whether IQ is 50 or 70% heritable... Given the need for cautious interpretation, it is especially noteworthy that the Devlin et al. (1997a) estimate of 49% for the heritability of IQ is in close agreement with the estimate of 51% reported by Chipuer et al. (1990) and the estimates of 47 and 58% reported by Loehlin (1989), who fit different although clearly converging models to the IQ correlations

On a different but related note, it's worth mentioning that this paper also states that

One of the most unfortunate misinterpretations of the heritability coefficient is that it provides an index of trait malleability (i.e., the higher the heritability the less modifiable the trait is through environmental intervention). Research on IQ provides an effective counter example to this false conception. As reviewed above, all available evidence converges on a moderate (in adolescence and childhood) to strong (in adulthood) heritability for IQ. Nonetheless, there is an equally strong convergence of evidence indicating that the population average IQ has increased substantially over the past 50 years (Flynn, 1998). The observation of significant trait heritability along with substantial secular increases, although perhaps paradoxical, is not a peculiar feature of IQ. Adult height is also highly heritable, but yet mean height appears to also have increased substantially in the same populations and over the same time period for which increases in IQ have been observed (Fernan dez-Bellesteros et al., 2001). Because the populations studied must be relatively genetically stable, environmental influences must constitute the source of secular increases in IQ. Dickens and Flynn (2001) recently proposed a theoretical model that would account for the high within-population heritability of IQ while allowing for environmental modifiability of the sort that would produce strong secular changes.

Which I believe was the mistake Jensen made in his 1969 paper. In his paper this was Jensen's exact claim, that a hight heritability meant that there is no point in trying to educate people from low IQ populations because this trait is not maleable because of it's high heritability. In this regard Jensen clearly fell into the misinterpretation. Alun (talk) 18:24, 25 October 2009 (UTC)

It seems that the lower bound estimate under the most conservative models for the heritability of IQ (in adults) is .50 and the upper bound is about .85 for other models. Note that the range is due to choice of models rather than uncertainty in the point estimates from any single model. I disagree with the details of your analysis of Jensen. The issue is more complicated, but certainly related. In other words, Jensen didn't make a _simple_ mistake. The issue is not well explained in the article now, but could be. Recursing back, to rephrase Bouchard's point: you inherit a norm of reaction (the relationship between your environment and your phenotype). The range of examined and plausible environments is central to any analysis of malleability and heritability. --Distributivejustice (talk) 18:34, 25 October 2009 (UTC)

Might it be worthwhile for the article to at least point out the three models/methods? Speaking of which, does the piece referred to above specify which are being used to obtain those results? And are these adult figures, child figures, or general mean estimates? --Aryaman (talk) 20:01, 25 October 2009 (UTC)

You ask hard questions. :) In short, yes, the models matter a lot. In a very important sense, all of these myriad debates can be distilled down to disagreements about which models are appropriate because they are taking place at a high level of statistical abstraction. Because statistical abstraction is entirely foreign to most audiences, most summaries gloss over this. Also, because the models of genotype-phenotype relationships are important across a range of topics far removed from this one, there's a large literature tangentially related to this. (But notably this topic draws enough academic attention that at times it has been leading the debate.) I'll try to find time to break out the details of your other questions, but it may be some time till I got to that. --DJ (talk) 21:09, 25 October 2009 (UTC)

• Actually DJ, the article doesn't say that the upper range for heritability of IQ is 85%, it says it's 70%. Here's the quote (the article is discussing the 50% result from the quote I give above):

  As McGue (1997) notes in an accompanying commentary, the result of this article is to center the debate on whether IQ is 50 or 70% heritable.

  So we could actually give this range. And yes, you're right, how one measures heritability is very important, as are things like sample size, how representative the sample is etc. The paper also goes into some discussion of the limits and strengths of twin studies.
• "I disagree with the details of your analysis of Jensen." I haven't given an analysis of Jensen. I merely pointed out that Jensen claims that the high heritability of IQ means that it's a waste of money trying to educate some populations, that is exactly what Jensen says, but this paper states that this position is a misinterprtation of heritability.
• "you inherit a norm of reaction", and? Heritability doesn't measure what you inherit, it is a property of the variation of the trait within the population, it is not a measure of what you inherit, i.e. the trait itself. You are conflating two different things. That's another mistake hereditarians often make, they think that heritability is somehow measuring the contribution of genes to the trait at hand. But heritability doesn't measure the contribution of genes to intelligence, it measures the contribution of genes to the variation of intelligence within a population. That is, if the highest measured IQ in a sample set is 150, and the lowest is 90, then a 50% heritability means that 30 points of this 'difference can be due to genes. That is that of this 60 IQ point difference, 30 points are due to environment, and 30 points are due to genes. What it doesn't tell us is how much of the 90 IQ points or 150 IQ points are due to genes or environment. But it's not as simple as that because there is a gene environment interaction that is often ignored, such that someone with "clever" genes who grows up in a non-stimulating environment will not perform well in an IQ test, which is almost certainly why environment has so much of a bigger influence during development. Or to put it another way, genes can't teach you algebra. Alun (talk) 06:53, 26 October 2009 (UTC)

I'm not sure where we disagree on your third point -- heritability is a point estimate of a population parameter but development is another matter and malleability depends on the (unobserved) norm of reaction and the range of plausible environments. On the first point look at Figure 5 and 7. Those h^2 values are around .85 or more in some cases. On the second, if Jensen said that "it's a waste of money trying to educate some populations" then he would be horribly and obviously wrong. He hasn't said that to the best of my knowledge. It would certainly be prudent for us to straighten that out. --DJ (talk) 07:27, 26 October 2009 (UTC)

(outdent) Even studies done on black MZ twins have yielded a heritability of ca. 80% (.77 according to Nichols' study of 76 pairs). I think it's becoming clear that it's pointless to throw numbers back and forth without more clarification as to the specific methods being used to obtain them. I also find it counterintuitive to provide a range without further clarification when we know full well that this range depends upon methods, samples, etc. Lynn makes a good presentation of three methods, and gives average figures for each of them. All of them average out to about .80 for adults. The reason I refer to Lynn is because he's presenting what he views to be consensus, not his particular findings. I'd like to see more literature which reviews a broader base rather than works which are concerned with narrowing down a particular number. --Aryaman (talk) 07:48, 26 October 2009 (UTC)

Can you explain how this is a fruitful line of development? No one denies that intelligence is significantly heritable. But what does this have to do with race and IQ? High heritability of IQ among blacks tells us nothing about why there is a difference in average IQ scores between blacks and whites. Heritability could be 90% and still the gap could be caused entirely by environmentally forces - could be. My point is the number tells us nothing. As someone else pointed out this article is not about heritbility of IQ, so much attention on this belongs in another article. This question is about differences in IQ between races. As far as I know nly a small fringe consider heritability measures such as tis salient. Can you explain how this helps us improve the article? Slrubenstein | Talk 10:36, 26 October 2009 (UTC)

Been wondering the same thing myself. We can report that "some believe that heritability of IQ is important to the debate and some disagree," followed by the reasoning (and in accordance with UNDUE). Anything beyond that is OR/SYN. T34CH (talk) 14:56, 26 October 2009 (UTC)

T34CH, Slrubenstein: Alun, DJ and I are working together towards improving one part of the text. Why should this bother you two? --Aryaman (talk) 15:28, 26 October 2009 (UTC)

Please stop being antagonistic. We bring up the issue of wp:SYN. You're spending a lot of energy and space here debating your personal understanding of a topic that is part of a different topic. Even if intelligence was 100% heritable from parents to adults, that still doesn't address group differences. This space should be devoted to how the sources interpret the literature, not what your interpretation of the literature is. T34CH (talk) 15:46, 26 October 2009 (UTC)

Yes, but maybe by explaining and discussing everybody's understanding of the literature, we could pave the way for an easier consensus?--Ramdrake (talk) 15:50, 26 October 2009 (UTC)

Folks, maybe it's better to work on improving the article one item at a time. Otherwise, things may get very confusing pretty fast. What do you say? However, Slr and T34CH do have a point that the discussions above may be somewhat outside the focus of the article, unless someone can explain to me how they relate?--Ramdrake (talk) 15:41, 26 October 2009 (UTC)

I thought we were still focused on the sentence that opened this section. That's an important and productive thread to close here. --DJ (talk) 18:51, 26 October 2009 (UTC)

Thanks for the clarification DJ. You're right that it needs to close. Just to recap quickly, most of the studies discussed above are investigations of heritability of IQ in individuals, are they not? That is to say, do they study the relationship of parent IQ to child IQ? T34CH (talk) 19:08, 26 October 2009 (UTC)

I'm sorry, I meant that it needs to be resolved satisfactorily, not that it needs to be terminated before it's completed. --DJ (talk) 19:17, 26 October 2009 (UTC)

That wasn't my intent. I'm sorry if it came off that way. Now about my question.... T34CH (talk) 19:26, 26 October 2009 (UTC)