Differences and similarities  

//Differences and similarities  

To deny that people differ from each other is patently ridiculous. We are all unique and our uniqueness is endlessly fascinating. Our physical differences are readily apparent; some of us are taller, heavier, stronger, paler, hairier and more appealing to look upon than others. No one denies these differences; to do so would be to deny the evidence of our own eyes.

We also know that people’s personality and character vary enormously – some people we are drawn to, others we instinctively dislike. We know too that not everyone has the same intellectual capability. Schools divide children into ability groups and teacher these groups differently. (This, I think, is part of the problem.)

When it comes to discussing physical characteristics, few people are so entrenched in their worldview as to argue that these qualities are, at least to some extent, inherited. Tall parents tend, on average, to have tall children. If parents exhibiting a particular phenotype – such as eyes with an epicanthic fold – were to have a child who lacked this distinguishing feature, their communities might be forgiven for drawing their own conclusions. Obviously, parents’ genes cannot be the complete story. We often think of a trait like height as completely inherited and yet, undernourishment during our formative years can make us shorter. Conversely, we’re prone to blame people for their weight, seeing it as entirely within our control and whilst of course food and exercise make a difference to our weights, it seems pretty certain that within a controlled environment, if two people eat and exercise in exactly the same ways, they will not necessarily put on the same amount of weight. At least some of the variance in our weights will be dependent on our genes.

But as soon as it comes to discussing cognitive differences, some people want to deny the differences between individuals could be in any way attributable to their genetic makeup. We want to believe that any such differences are the result of social inequality and that, if only society was just, all such difference would fade away. This strikes me not only as naïve, but as being in opposition to a great deal of scientific evidence.

Now, I’m the first person to accept that the kind of epidemiological studies that have produced this evidence are not the same thing as proof, but when so much data points in a single way, it’s bafflingly wrongheaded to ignore in favour of believing what we prefer to be true. As I reported here, there seems to be a correlation between the tendency to down play the role of heritability in human difference and a mistrust of the ability of science to answer empirical questions. An alarming number of academics appear to prefer identity politics and postmodernism to empirical data. As such, they are particularly prone to committing the moralistic fallacy and asserting that what ought to be is in fact true. What ought to be only ever exists in our minds and cannot be observed in the physical world.  Wanting a thing to be true does not make it so.

Recently, various commentators have seen fit to describe me as a racist because of the following exchange in the comments on this blog:

As you can see, I acknowledge that these findings are unpopular – I have to admit to being far from happy about them myself – but they do represent a mainstream view of the research. The link I provided was to a paper by Linda Gottfredson and colleagues entitled “Mainstream science on intelligence: An editorial with 52 signatories, history, and bibliography“. This was first published as a public statement intended to represent a scientific consensus to confront the misconceptions surrounding the concept of intelligence. Gottfredson lists 25 different points, some of which make for very uncomfortable reading. Here’s a taste:

Group Differences

7. Members of all racial-ethnic groups can be found at every IQ level. The bell curves of different groups overlap considerably, but groups often differ in where their members tend to cluster along the IQ line. The bell curves for some groups (Jews and East Asians) are centered somewhat higher than for whites in general. Other groups (blacks and Hispanics) are centered somewhat lower than non-Hispanic whites.

8. The bell curve for whites is centered roughly around IQ 100; the bell curve for American blacks roughly around 85; and those for different subgroups of Hispanics roughly midway between those for whites and blacks. The evidence is less definitive for exactly where above IQ 100 the bell curves for Jews and Asians are centered.

And later

Source and Stability of Between Group Differences

19. There is no persuasive evidence that the IQ bell curves for different racial-ethnic groups are converging. Surveys in some years show that gaps in academic achievement have narrowed a bit for some races, ages, school subjects and skill levels, but this picture seems too mixed to reflect a general shift in IQ levels themselves.

20. Racial-ethnic differences in IQ bell curves are essentially the same when youngsters leave high school as when they enter first grade. However, because bright youngsters learn faster than slow learners, these same IQ differences lead to growing disparities in amount learned as youngsters progress from grades one to 12. As large national surveys continue to show, black 17-year olds perform, on the average, more like white 13-year-olds in reading, math, and science, with Hispanics in between.

21. The reasons that blacks differ among themselves in intelligence appear to be basically the same as those for why whites (or Asians or Hispanics) differ among themselves. Both environment and genetic heredity are involved.

I can aware that bringing these comments to light is to open a can of worms. I also know that despite having been signed by 52 leading intelligence researchers, there are other academics who disagree with some of the 25 statements made. The IQ gap between different groups is a matter of fact. To deny that this is the case does no one any favours. The cause is. however, complicated and, as we’ll see, there’s reason to doubt that “[b]oth environment and genetic heredity are involved” in these differences.

So, is it racist to talk about race? Surely that must depend on one’s purpose. As I understand it, racism lies in promoting the superiority of one racial/ethnic group over another for prejudicial, discriminatory or aggressive ends. I abhor racism and see any attempt to discriminate against groups of people based on real or supposed racial differences as wrong. I know that race is, in large part, a social construct and that different groups’ experiences differ depending on all sorts of environmental factors.

And, most importantly, I believe that the differences between us pale into insignificance next to our similarities. What makes education possible is that despite our differences we all learn and think in broadly similar ways. But, as some groups of people seem to get a worse deal than other groups, I don’t think equality is great idea. Much better, I think, to treat everyone fairly.

In a truly equitable system – no one would be discriminated against in any way and the differences between students’ outcomes would be completely due to heritability. Although the goal of universal education is to create a totally fair environment for all, we’re a long way from achieving that aim.

That being the case, knowing that a) IQ apparently distributes disproportionately along racial grounds and b) IQ is a very strong predictor for social mobility and wealth, should make us more determined to ensure the least privileged children enjoy the same benefits as the most privileged. There’s good reason to believe that the IQ gap between different racial/ethnic groups is likely to be environmental rather than genetic in origin (and some reason to think this is exaggerated in the US where such environmental differences may be more stark) but there’s no reason for supposing the gap doesn’t exist.

From about 57 minutes into this video, Stephen Hsu talks about how IQ is a much better predictor than race as to whether children increase their socio-economic status (H/T Mario Lopez)

Simply denying that there are differences in the IQ scores of different groups won’t help anyone, acknowledging that this may be the case and seeking to alter what is to be more in line with what ought to be seems to be the fairest possible approach. I fail to understand by what twisted definition this could be considered racist. On the contrary, those who seek to deny the existence of racial differences are, whatever their intentions adding to the problem and getting in the way of the solution. As James Flynn, discoverer of the Flynn Effect, argues is this interview:

“It’s whites, not blacks, who complain,” he says. “Blacks know the score. Facts are facts.” On recorded IQ tests, he says, African Americans have persistently lagged behind [pdf] most other ethnicities in America [pdf] (including, according to some commentators, black immigrants from, for example, the Caribbean) and this cannot be explained by the Flynn effect since, as he puts it, “blacks don’t live in a time warp”.

As such, perhaps it’s more reasonable to describe the position of those who deny the existence of these issues as racist?

To summarise:

  1. At least some of the differences between different racial/ethnic groups is genetic, some of the differences are environmental in origin.
  2. Some racial/ethnic groups score lower, on average, than other groups.
  3. Our similarities are far more pronounced than our differences.
  4. IQ is one of the best predictors for social mobility.
  5. The aim of education ought to be to increase the intellectual capability of all children
  6. Some children will need more help than others.
  7. Ignoring racial/ethnic differences is only in the interest of privileged groups and acts to discriminate against the most disadvantaged.

If you’re interested, my manifesto for a fairer approach to education is outlined here.

2017-08-21T21:47:56+00:00August 16th, 2017|research|


  1. @davowillz August 16, 2017 at 12:59 pm - Reply

    I can finally see what you were trying to discuss now. Thanks David. I’m not sure how this would change anything I do as a teacher, but it’s interesting stuff nevertheless. 🙂

  2. David F August 16, 2017 at 2:02 pm - Reply

    Hi David–I’m not sure how much this helps teachers–regardless of genes, I’m going to do my damnedest to teach my students the knowledge they need to understand this crazy world.

    This isn’t my area or really my interest, but I will say that the link you provide of the signatories to that letter is research built around Charles Murray’s The Bell Curve. Just so you understand what hackles that raises in the US (where we have a nasty history of race), see this on him from the Southern Poverty Law Center (one of the most prominent anti-racist orgs in the US): https://www.splcenter.org/fighting-hate/extremist-files/individual/charles-murray

    • tomaswhiteucl August 16, 2017 at 3:55 pm - Reply

      Hi David, I should have read your comment first. I posted something similar. Scary stuff.

      • David F August 16, 2017 at 6:37 pm - Reply

        Not so much scary, as off the mark, IMHO. As knowledge-based teachers, we strive to teach knowledge–one of the cool things about The Learning Scientists page is that there are things we can do for ALL students (none of this learning styles baloney!) to help them acquire knowledge and then use that within the context of the subject at hand.

        Erasmus talked about education in terms of talent, training and practice–I would say that most trad teachers see the same—yes students differ in terms of talent, just as there are talent differences in athletes. However, training and practice help develop what nature has provided and it’s that which ought not be individualized.

        What I did find problematic is that DD seems to have embraced genetic determinism to a degree–I find this troublesome as that line of thinking assumes that we don’t have individual choices (i.e., free will). I’m not 100% sure where he was taking that argument, but in a world where we are talking 24-7 about Nazis that’s not a good place to be. Individuals still have choices (and hopefully have been taught the knowledge to make those choices)–even if their circumstances/social backgrounds influence them.

        If anything, I’d like to see us as teachers talk more about helping students develop that knowledge and to develop a sense of duty so that they may use their will wisely (and maybe I’ve read too much Kant).

    • David Didau August 16, 2017 at 7:11 pm - Reply

      David, I agree that teaching children a knowledge rich curriculum is absolutely the way forward.

      I know (of course) of the hysteria surrounding Murray & Herrnstein’s book but rejecting empirical data on the grounds that we don’t like the source is just an ad hominem, isn’t it?

      If a decent data set comes along and contradicts any of these positions I’ll be the first to acknowledge I was wrong. Promise.

      • David F August 17, 2017 at 1:18 am - Reply

        Hi David—I’ll look more closely at your free will piece later…feel free to DM me if you’d like (I’m on Twitter as @stuckinphilly). Regarding The Bell Curve and the data, are you familiar with this?: https://www.brookings.edu/articles/does-the-bell-curve-ring-true-a-closer-look-at-a-grim-portrait-of-american-society/

        • David Didau August 17, 2017 at 7:36 am - Reply

          I have read that, yes. It’s a sensible, proportionate critique and I’m persuaded that there are plenty of errors and omissions in The Bell Curve (one reason why I chose not to cite it) but it doesn’t show that the IQ data sets and factor analyses is faulty, does it?

      • elkysmith August 18, 2017 at 3:42 am - Reply

        David, have you also read Gould’s ‘The Mismeasure of Man’? Just as criticism of Gould’s work should not be immediately dismissed simply because of the individuals who make it, there is clearly more to the Murray & Herrnstein ‘hysteria’ than ad hominems.

        • David Didau August 18, 2017 at 5:09 pm - Reply

          Yes I have. I think both Gould and H&M *both* did a great deal to poison the well. Both are guilty of selectively ignoring and misinterpreting evidence. At the same time, both have merits. Disentangling the good from the bad is made harder because I’m not sure either argue in good faith.

  3. sputniksteve August 16, 2017 at 2:30 pm - Reply

    Hi David,

    I must confess to being troubled by all this. Perhaps that’s my own cognitive dissonance.

    But I think there is a risk of the causation/correlation with regards to IQ in different groups. Like you, I’m not a geneticist. But it seems to me that we can’t be certain of a genetic determination of IQ or the differences observed in IQ between different groups unless we can somehow devise an IQ test for application in utero. And even then, I’d be sceptical of the results. The environmental and cultural influences on such constructs as IQ are likely far larger than we appreciate. And that’s assuming that we trust IQ measures, which I don’t.

    I’ve read that even within the field of genetics, the concepts of both “race” and “IQ” are considered to be controversial and that geneticists regard “race” as a social construct anyway.

    Identifying disparities between groups is one thing, but you’ve also said that parents, schools and the environment have little impact on us as adults. I’d dispute this, particularly with regard to issues around UBS rave, ethnicity, and gender.

    • sputniksteve August 16, 2017 at 3:32 pm - Reply

      I have no idea where “UBS rave” came from! Should have read “race”.

    • David Didau August 16, 2017 at 7:06 pm - Reply

      I’m troubled by it too. I’d much prefer it not to be true and if a reputable data set comes along to show it’s false I’ll be mightily relieved. But denying empirical data isn’t helpful for anyone

      • Liz August 16, 2017 at 10:04 pm - Reply

        My concern is that “evidence” based on a false premise or based on a social construct (which race is – there are no genetic markers which distinguish race, it is mostly visual, and family trees that group us) is not the best evidence, even if it is the only evidence.

        Data is interpreted to become evidence. If the scientists interpreted the data through their own views (and those views saw race as a legitimate grouping; or were actually racist, ie based on notions of superiority) that would suggest their work is flawed. Possibly even that they have fallen for some logical fallacies: correlation equals causation being the obvious one. (This is before you look at the tools used to measure intelligence: these are problematic as there is a cultural context here too)

        Isn’t it just more possible that social and environmental factors, such as nutrition, access to resources, cultural capital and ability to focus on education result in better outcomes for some groups, which show up in testing.

        Race and heritability suggest we can’t do anything. It’s an excuse. But worse in these days of Neo-nazis and hate crimes, it is irresponsible to argue ‘evidence’ as the gold standard, without checking whether the evidence IS gold standard.

        Side note: funding is problematic in research – big tobacco and the Oil industry are two case studies that show this. If your commenters have identified links to groups with clear racial bias, this does suggest a review of the evidence.

        Thanks for your time in reading this. I do enjoy reading your pieces and the thought they inspire, but in this instance I think there are serious problems, rather than just “uncomfortable” ideas.

      • sputniksteve August 16, 2017 at 11:51 pm - Reply

        I might come back to this in more detail.

        But in sum:

        – I don’t trust IQ as a measure in the first place. I certainly don’t see it as a reliable measure of anything to do with genetics.

        – I’m given to understand that the scientific consensus is that “race” is a social construct and not a genetic one.

        So we are using an unreliable measure to measure something that doesn’t exist.

        But even if we accept IQ and race as things, there’s no way of being able to say it’s genes what done it, unless you can test for IQ prior to birth. Even then, I’d be cautious.

        • David Didau August 17, 2017 at 7:23 am - Reply

          IQ is the most stable construct in psychology with the best predictive power. You can choose not to accept these facts but that doesn’t make them any less true.

          The ‘scientific consensus’ on race is a tad inconsistent. This thread is worth a look: https://twitter.com/AndrewSabisky/status/897945259006726168

          So, as I *do* accept that IQ and race exist, it may surprise you to know that in my post I have *not* said genes are the cause. In fact, I said, “There’s good reason to believe that the IQ gap between different racial/ethnic groups is likely to be environmental rather than genetic in origin (and some reason to think this is exaggerated in the US where such environmental differences may be more stark) but there’s no reason for supposing the gap doesn’t exist.”

          • sputniksteve August 17, 2017 at 12:26 pm

            My apologies for suggesting, then, that you might be implying a genetic link. This was the result of me conflating your recent series of blog posts about behaviour, parenting and so on.

            What’s your view on the results reported here: http://www.independent.co.uk/news/science/iq-tests-are-fundamentally-flawed-and-using-them-alone-to-measure-intelligence-is-a-fallacy-study-8425911.html ?

            I’ve found the paper that report is about and it turns out that it spurred an interesting sparring match between its author, Hampshire, and its critics Haier et al.

            The original paper can be found here: http://www.sciencedirect.com.ezproxye.bham.ac.uk/science/article/pii/S0896627312005843

            The first response can be found here: http://www.sciencedirect.com.ezproxye.bham.ac.uk/science/article/pii/S0160289614000270

            In which its critics argue several points about the paper, but do state that “g and IQ are not synonymous”. They also say, “intelligence researchers have never argued that a single construct like g can account for the “entire distribution of general intelligence” ”

            ” g may not be manifested in all brains in the same way, further complicating the concept of “neuro g”.”

            And, in their conclusion, “We are still a long way from identifying the brain characteristics of general intelligence and why some people reason better, remember more, and learn faster than other people. ”

            I think this summarises the complexity of the issue here pertaining to IQ.

          • sputniksteve August 17, 2017 at 12:29 pm

            However, I agree that any differences that can be identified between *any* groups in terms of anything that might claim to measure such a nebulous concept as “intelligence” are most likely to be environmental.

          • David Didau August 17, 2017 at 3:09 pm

            Intelligence is *not* a nebulous concept.

        • Bapz August 17, 2017 at 7:56 pm - Reply

          The Illusion of Culture-free Intelligence Testing: http://www.web.uvic.ca/~lalonde/psyc435c/cole.pdf

  4. tomaswhiteucl August 16, 2017 at 3:32 pm - Reply

    Hi David,

    I just wondered if you were aware that some of the signatories on the 1994 Gottfredson editorial, including Gottfredson herself, have been large grantees of the Pioneer Fund, which is listed as a hate group by the Southern Poverty Law Center. The Pioneer Fund was set up by eugenicist Nazi symapthisers in the 1930s and has been referred to as a ‘neo-Nazi’ organization. More information can be found here: https://en.wikipedia.org/wiki/Pioneer_Fund and here: https://www.splcenter.org/fighting-hate/extremist-files/group/pioneer-fund

    I don’t believe that your intention in using this kind of research has any racist motivation on your part, and I think that your being branded a ‘racist’ by some people is unfair. But I mention this because I think you should be careful in ascribing ‘mainstream’ status to this kind of research. I don’t know many researchers who accept large grants form neo-Nazi hate groups.

    All the best,


    • Chester Draws August 16, 2017 at 8:59 pm - Reply

      Well Tom, there will be a large group of evidence showing that Gottfreson was wrong, won’t there? Bring it on.

      That a researcher was funded by a group that was set up 60 years earlier by someone who had some WrongThought (TM) is a ridiculous slur by association.

      Am I a Nazi sympathiser because I drive a Ford?

      • tomaswhiteucl August 17, 2017 at 11:37 am - Reply

        Dear Chester Draws,

        Where is the wealth of research NOT funded by racist organizations which proves the claims that there are ‘fairly clear racial differences in IQ’? If this is the ‘mainstream’ as David claims it is, then questions must be asked. WHy is the ‘mainstream’ only being given a voice by neao-Nazis? Is there some sort of collective scientific obscurantism at play in contemporary society? This is the domain of conspiracy theory.

        I’m not sure where the number 60 was plucked from, as the Pioneer Fund has, in fact, been in existence since 1937. Its mission was to pursue “race betterment” by promoting the genetic stock of those “deemed to be descended predominantly from white persons who settled in the original thirteen states prior to the adoption of the Constitution.” In the 1930s, as a result of the close personal links between its directorate and the Third Reich, it actively promoted the dissemination of Nazi film and documentation to do with eugenics. In the decades following the war, it actively worked to thwart the civil rights movement in the USA. Aside from funding a slew of researchers into ‘racial intelligence’ who have been branded as racists, it has been engaged financing repugnant people and organisations such as white supremacist Jared Taylor, founder of the New Century Foundation and American Renaissance, a mouthpiece of the fascist right in the USA, whose thinking on ‘race’ and society can be found here: https://www.amren.com/about/issues/

        The Pioneer Fund is currently listed as a hate group by the Southern Poverty Law Center, an organization which exists to protect civil rights. The Pioneer Fund is a contemporary agent in the funding and propagation of a racist ideology that has a direct historic link to Nazism, and is currently influential in the politics of hate that we see being given a voice in Trump’s America. Given all of this, I trust that you will agree that your Ford analogy is not relevant, and is perhaps best parked elsewhere.

        In the meantime, I ask David, again, to acknowledge the problematic context from which the research he cites as ‘mainstream’ has come. David, were you aware of these facts? If you were, why did you not reference them? If you were not, has their coming to light changed anything for you?

        I look forward to hearing from you.

        All the best,


        • Michael Pye August 17, 2017 at 2:19 pm - Reply

          Tom it is a bit of a red herring. The idea that research showing a racial IQ difference would be of interest on racist groups is pretty obvious and what we would expect. It doesn’t mean more reasonable or less contaminated research is not being carried out. You have just committed the moralistic fallacy David was talking about in a previous post.

          • tomaswhiteucl August 17, 2017 at 2:47 pm

            Hi Michael.
            My issue is that David has not used what you call “more reasonable or less contaminated” research upon which to base his claims. (where is this “more reasonable or less contaminated” body of research? – does your choice of words signal that you agree that there is a problem with the research cited above, that it is unreasonable and contaminated?). Instead, he has used research that was funded by a white supremacist organization who, by the way, have faced criticism for the lack of rigour in their grant-awarding process.
            There is also the issue that the quality and integrity of Gottfredson et al’s claims has come under criticism and does not represent a ‘mainstream’ view of the field, as David claims it does. More detail can be found here: https://www.splcenter.org/fighting-hate/extremist-files/individual/linda-gottfredson
            Furthermore, he has not pointed any of this out, nor has he acknowledged it post-fact. He owes it to his readers to highlight this glaring conflict of interest.

          • Davuthoca August 21, 2017 at 9:08 am

            What you are holding on to is the idea that the research has been done correctly. Isn’t it also possible that within today’s context certain types of researcher self select? By certain types I mean … a bit racist. This gottfredson and her cohorts are unashamedly racist. How can we be sure that there overt biases have resulted in good science?

          • David Didau August 21, 2017 at 10:46 am

            Here’s a much more recent review of the evidence and one that has not been contaminated by accusations of racism: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4270739/. Here too, is an excellent paper by Eric Turkheimer (also not a racist) http://people.virginia.edu/~ent3c/papers2/Perspectives%20on%20Psychological%20Science-2016-Turkheimer-24-8.pdf

          • SamJeeves August 21, 2017 at 11:35 am


            Thank you for providing those.
            However, those two papers are fundamentally different to the Gottfredson et al. your core conclusion rests on. They are about the heritability of intelligence; which I don’t think people were ever questioning. Very importantly, at no point do either of these studies try to relate their findings to “race”. In our extensive discussion below I have already given plenty of detail as to why this is problematic for that paper. These sources cannot be seen as a non-racist endorsement of the gottfredson paper. Turkeimer does not reference the paper at all and the Plomin paper only uses it in reference to the definition of intelligence, importantly it does not reference the conclusions. There is a really significant fundamental difference between the conclusions reached in this more contemporary science and the problematic gottfredson paper. I hope that you can see this. Is this really the best evidence which backs up your claims about race and intelligence? if so, those claims to not seem to stand on very firm ground (particularly give the concerns tom has raised above)



          • David Didau August 21, 2017 at 11:56 am

            That IQ score averages are different for different racial/ethnic groups is a brute fact and isn’t in dispute. There’s tons of evidence for this. Here are ten papers to explore:
            1. http://www.sciencedirect.com/science/article/pii/0160289688900293
            2. http://www.sciencedirect.com/science/article/pii/S0160289602000806
            3. http://www.sciencedirect.com/science/article/pii/0160289694900515
            4. http://journals.sagepub.com/doi/abs/10.1177/0095798401027002004
            5. http://www.nber.org/papers/w12066
            6. http://www.sciencedirect.com/science/article/pii/0191886995001581
            7. http://psycnet.apa.org/record/2005-00117-007
            8. http://www.sciencedirect.com/science/article/pii/0160289683900120
            9. https://search.proquest.com/openview/fe0092108bd40bacbcd331c8bb00921e/1?pq-origsite=gscholar&cbl=33429
            10. https://www2.le.ac.uk/departments/npb/people/amc/articles-pdfs/racediff

            There are very many more. They all reach different conclusions and investigate different questions but they rely on the existence of the *fact* that there are differences as their starting point.

            There’s even some very positive research on gains made which also rely on the statistical differences: http://journals.sagepub.com/doi/abs/10.1111/j.1467-9280.2006.01802.x
            Does this seem like firm ground?

            As I’ve explained, repeatedly, I do not believe such differences are genetic in origin. There are plenty of people who explain away these differences saying the test are biased on that IQ is meaningless or other postmodern denialist tactics, but pretending there are no differences doesn’t help anyone.

            Can we move on and discuss how best to make education more equitable?

          • SamJeeves August 21, 2017 at 1:41 pm

            Thank you for taking the time to provide those sources David.
            You may remember in my first comment below (on the 17th), that it was not my intention to question the data, but to have a discussion about the use of race as a useful causal indicator. I was just surprised to see you provide those two sources as a defence against the concerns Tom raised above, hence my questioning its validity. (Those concerns are important, if true, and should be recognised.)

            Looking through those papers does bring about additional questions which tie in well with your concerns with wanting to discuss making education more equitable and with my questions about using “race” as a factor and its use in applying the conclusions outside of the population:

            – The data seem to be largely from 1980-2000 in the United states:

            How is this useful in a UK education perspective? You recognised in your piece that the differences are likely to be exaggerated in the US. Is there any data on UK populations IQ distributions? If there isn’t then surely we know nothing which could help us – it would be very dangerous to assume that we could try to map that data onto our own population. (In fact, the only way this could be justified is if you think it has a genetic causal connection which is well linked to “race” which we’ve long established is not the case) This is what I was questioning in my comment below on August 17, 2017 at 11:58 pm

          • David Didau August 21, 2017 at 8:49 pm

            We don’t do IQ tests at scale in the UK so there isn’t any comparable data. Also we tend to collect data differently – by ethnicity rather than race – and so even by looking at math scores (a good proxy) we can’t directly make any comparisons.

            You’re right to highlight this as a potential weakness – maybe it’s true that even the environmental conditions in the US lead to such asymmetrical outcomes, over here in the UK we have a much more equitable system and there’s nothing to worry about. But to arrive at such a conclusion would be to ignore several important data points. We could look at GCSE grades, 11+, gap between degree classifications etc. Here’s some interesting data on CAT scores to chew over: https://lesacreduprintemps19.files.wordpress.com/2012/02/glassessment.pdf

            So, in answer, I don’t think it would be “very dangerous’ to assume that the experiences of different racial/ethnic groups in the UK and US are not broadly analogous, but obviously it won’t be the same. The only reason for ignoring this is if you believe children of difference ethnicities have equitable environments in the UK. I think that would be, at best, naive.

  5. Jenn August 17, 2017 at 12:50 am - Reply

    Hi David – isn’t there a basic question about how we determine IQ as, from what I understand, there are inherent cultural biases in the tests e.g. there are arguments that some of the Wechsler Adult Intelligence Scale (originally developed in the US) subscales are not culturally fair and lean towards the understandings inherent in the culture in which they were developed. Additionally, how are we actually defining ‘intelligence’. While there is considerable data out there about so called ‘racial’ differences (which is an illogical term because there is only one race i.e. human) these seem to be based on the assumption that the tests can be devised and applied without any cultural bias and that we are all working on a common definition of ‘intelligence’.

    • David Didau August 17, 2017 at 7:30 am - Reply

      No, I don’t think so. There is a function in America for people to raise concerns about any question in the test they think has got past and has cultural bias in it. The mechanism they use is Differential Item Functioning. It controls for parent education and looks how different groups score on then they can then see if it’s biased against minority groups (similar white kids doing better) findings generally show that white/Asians do better on questions considered easy, Black kids do better on questions considered ‘harder’ but questions have small differences.

      Here’s a link to explain Different Item Functioning: link to DIF: https://en.m.wikipedia.org/wiki/Differential_item_functioning

      Here’s my favourite definition of intelligence: “A very general mental capability that, among other things, involves the ability to reason, plan, solve problems, think abstractly, comprehend complex ideas, learn quickly and learn from experience. It is not merely book learning, a narrow academic skill, or test-taking smarts. Rather, it reflects a broader and deeper capability for comprehending our surroundings—”catching on,” “making sense” of things, or “figuring out” what to do.”

      If we’re going to say that there’s only one race, that’s fine, but we then we need some sort of term to describe the groups who do better and worse on IQ tests. In the post I’ve used “racial/ethnic groups”.

  6. education86466 August 17, 2017 at 5:53 am - Reply

    I have trouble accepting your interpretation (or that of your sources) that the different IQ bell curves for different races represent a genetically controlled feature. We know education has an effect on IQ. We know that poverty affects educational outcomes. We know that environmental effects in the first 3 years of life can affect cognitive development. Therefore in my view the differing bell curves for IQ could be accounted for by factors other than genetics. The fact that the average 17 year old African-American male has a reading age of the average 13 year old white male could have far more to do with the socio-economic status than genetics. It could be I haven’t quite understood your arguments or evidence, please correct me if this is the case.

    • education86466 August 18, 2017 at 9:29 am - Reply

      Oops I understand now you were not suggesting the IQ curves represent genetically controlled features, apologies.

  7. Tom Sherrington August 17, 2017 at 10:16 am - Reply

    Hi David

    I’m not normally keen to wade into online exchanges but I think this is an important area so here’s my contribution. I have huge respect for you and your work but my sense of this is that people are right to be agitated. Comments such as ‘there are fairly clear racial differences in IQ’ are highly problematic in terms of the science but also dangerous when isolated from an immediate exploration of the details. Until reading this follow-up post in detail, I had the (false) impression that you were asserting that this was a genetic factor (because you are currently exploring hereditary factors) – when in fact you are suggesting variations in IQ profiles might be wholly environmental – and therefore a social wrong that can be corrected by appropriate action. That needs to be said up front, loud and clear so as not to give any fuel to the genetic-supremacist racism that is out there, finding new strength.

    I worry about making bold claims that suggest you’re presenting the ‘truth’ – that others simply need to accept, even it unpalatable. For example, the Gottfredson paper you cite from 1994 presents a big gap in IQ averages – 85 vs 100. That seems implausibly large and my immediate question is around the source and validitiy of that data, and whether it has been reproduced in other contexts. It’s such an incendiary ‘fact’ that I wouldn’t publish that figure alone. One figure is never the truth. I don’t know any educational data that is rigorous enough to be presented as one single data point and, given the issue at hand, ultra data skepticism is warranted here.

    Also, the paper predates the explosion of genetic science since genome sequencing became routine and it seems hugely problematic as a source – even leaving aside the debate about the dubious funding and the disputed claims of representing mainstream science at the time. There seems to be a suggestion that because IQ has some genetic origin (which is pretty much beyond dispute – accepting caveats about the limitations and biases of the IQ measure itself) and different racial groups have different (albeit largely overlapping) IQ distributions, that this therefore means that the IQ differences between groups must also have a genetic origin. Is there more to it than that?

    I don’t think there is any genome-based evidence for this assertion – especially given the genetic complexity of IQ-factors vs surface ethnic identifiers. The notion of ‘race’ is basically false; there is only one race with a continuous spectrum of ethnic variations along which people self-identify at various points. The timescale of human evolution, population growth and global migration doesn’t really allow for significant genetic changes at the level of IQ between geographically separated groups. Essentially, any population IQ changes and differences between so-called racial groups must be entirely driven by environmental factors and that should be asserted strongly. If the distributions of IQ have significant differences sustained over time, that simply indicates how entrenched the social/environmental inequalities are.

    Given that, ultimately, you are trying to make a strong social justice case for better educational fairness based on a deeper understanding of what constitutes learning, intelligence and so on – why not start off with that instead of getting mired in this kind of argument? Your blog is so influential that you can’t even make a comment that won’t be read widely. It should be possible to conduct this discussion a million miles away from accusations of racism but that requires an acknowledgement of the legitimate concerns and sensitivities that people have.


    • David Didau August 21, 2017 at 9:28 pm - Reply

      I can’t tell you how bitterly disappointed I am to see you implicitly endorse the bullying and abuse I’ve experienced over this.

      By all means disagree with the views I’ve presented – you’re welcome to dismiss me as a fool and a charlatan. And by all means critique the study I cited in this post (which is not in any way a “single data point” but a review of huge data sets) there are very more out there – these difference are an established fact – what causes the gap is a matter of debate, and I’ve made my position on this clear – there is NO genetic cause for these differences. It’s hard not to suspect that you didn’t actually read the post as you appear to making the very same arguments I made myself.

      I too have huge respect for me but please don’t lecture me on my responsibility to take what I consider pretty toxic postmodernist mumbo jumbo into account when writing.

      • Tom Sherrington August 28, 2017 at 11:45 am - Reply

        OK David. I read the whole piece – of course. I’ve seen criticism and concerns but if there’s been bullying I’m sorry to hear that. I’ll email at some point. Take care. Tom

  8. SamJeeves August 17, 2017 at 2:04 pm - Reply

    Hi David,

    I am not going to dispute any data which has been referenced or analysed above. However I am going to challenge your use of race a a scientific descriptor which has any useful causal connection. You can group people into many arbitrary groups based on various characteristics and find differences in the distribution of another variable for those different groups (in your case race and IQ). However, unless the initial grouping is demonstrably causal to the second measure and has accurate predictive power then it is not useful and does not have explanatory power; at that point the “pattern” becomes relatively meaningless and the selection of the original grouping (race) against which you choose to measure your other variable (IQ) does have inherent bias as you could have chosen another way to group people. for example:

    Eye colour (a phenotype where one type is often much more prominent in some “races”) may well have also shown a, probably less pronounced, difference in IQ based on this data set, but a difference non the less. If such a difference was shown I could have plenty of bluster about the fact that “people with blue eyes have on average a higher IQ than brown” the numbers don’t lie and there is no point in hiding from the “truth”! But why did I select “eye colour” in the first place? What is my aim here? Does it give me any tools as an educator to help an individual based on their eye colour?

    In the example above I have in fact picked a grouping which is more precise than “race” perhaps I should have gone for something more similarly arbitrary like “premier league football team supported” . At least the number of football teams is a well defined number. How many races are there David and what are they? I certainly could’t tell you! If one person could tell you that there are four races, another would say there are six and others could argue that there are 20 how on earth could that category be meaningfully causal? My predicted IQ could change based on your arbitrary choosing of the number of groups I could fall into. I don’t think that people believe the “patterns” you have described are “racist” – it’s just data, but the decision to use that as the way to group people even though it is a hugely ambiguous term does show a bias in what you want to show.

    To further the conversation a little let us try to use this pattern you have in a meaningful way:
    If I was to show you a photo of a random child:
    – could you accurately identify their “race”? (how many races are you going to use?) – we’re already into hot water here re: bias !
    – Can you make an accurate prediction of the IQ of this individual based on your determination of their “race”?
    – Does this help you in any meaningful way to teach this child?

    In my opinion, if the answer to the final point here is that it doesn’t help you then the pattern you have found in the data is meaningless and unhelpful to anyone other than someone who wanted to show that there are differences either side of a line they chose to draw in the sand. As that grouping has no demonstrable underlying causality then you have chosen to draw that line in the sand, you didn’t need to!

    Hopefully you can see where I am coming from here.

    I look forward to continuing this discussion



    • David Didau August 17, 2017 at 4:31 pm - Reply

      I used the phrase “racial/ethnic groups” in the post to try to capture the fact that data analyses of IQ tests often asks participants to identify their racial/ethnic group, just like equality opportunity forms in job interview packs routinely do. I’m not claiming there any particular scientific validity to the term ‘race’ beyond the fact that this is how participants in studies choose to identify themselves. The brute fact is that those who identify as Jewish or Asian do better than those who identify as white. I don’t think that this result has got anything to do with genetics per se *but* the differences between individuals is highly heritable. There is no incompatibility with these two positions. Most researchers seem to accept that the gap between these groups is almost certainly environmental in origin. Does that all make sense?

      • SamJeeves August 17, 2017 at 6:51 pm - Reply

        Hi David, thanks for your reply.

        Yes, that all makes sense and I think your most recent post has made your position much clearer.
        I think that this is an unfortunate and out dated way do divide people into groups. It is only really a useful piece of information to collect if you might want to work out what a person might look like without a photo. I’m sure that there are plenty of people who genuinely struggle to fill in that part of a form and wonder why they are constricted to one box! I suppose what we really need is the same IQ data collected against environmental factors which might actually affect the performance of an individual in those tests! Unfortunately that does not seem to have been the focus of the researchers you’ve referenced. That’s a shame as that’s the only way we can do anything to remove differences that shouldn’t exist.



        • David Didau August 17, 2017 at 8:02 pm - Reply

          I don’t think that’s true. Surely we NEED to know if sections of the population are losing out to this degree? Ignoring a problem doesn’t make it go away.

          • SamJeeves August 17, 2017 at 8:27 pm

            Sections of what population? The whole world? Surely those racial categories smooth over too much to be of any use there if the affects are environmental in origin? Perhaps triangulating ethnic identity and say other measures like income bracket might allow people to identify vulnerable /underperforming communities IN THEIR SITUATION, but thinking that you can look at the conclusions reached in that data and then immediately apply it to any situation in any country, in any city, based on some sort of best fit to their ‘race/ethnic identity’ is surely false? This would certainly then seem to be implying that ‘race’ and the associated genes were the contributing factor to performance! (Which is exactly not what we’re saying) In research like the one you are referencing surely the aim is to reach some sort of universal conslusion that could be applied in any/ many human contexts? If that is your aim, then ‘race’ is not a useful piece of information. (Unless you think that races are well defined things which are deterministic of performance measures – which you don’t)

          • David Didau August 17, 2017 at 10:18 pm

            Of the population who took the test. It’s s statistical term.

  9. […] the time of writing, there’s a group of people on Twitter calling me a racist for writing this blog post. This feels pretty awful. The violence of such a term is hard to quantify, but it’s left me […]

    • SamJeeves August 17, 2017 at 11:58 pm - Reply

      Hi David,

      Thank you for your clarification.
      I still feel that there is something we need to clear up here. please bear with me.
      Based on your clarification, are you saying that the conclusions of the data do not apply outside of the sample of people who took the test? (the population) This certainly isn’t the implication of the original piece. If that were the case, there certainly isn’t much we can learn from the data.

      However, if this is not what you are claiming, and instead you think that you can extrapolate this fairly reliably (as would be the norm where variables selected have precise scientific definitions) to other people who would identify as each of these races (or be identified as) then we are saying that “race” (and the associated IQ profile) is something that can be applied to people as a hard and fast rule pretty much regardless of environment. This seems to be a contradiction considering that it has also been made quite clear that you believe the effect to have environmental cause and it is not linked to a ubiquitous genetic link to “race/ethnic identity”.

      As I see it, we have two fairly clear options here

      “race/ethnic identity” is a clearly defined thing which can be reliably applied to people outside of the population tested;

      therefore the test tells us something important about the likely success of other individuals who identify as this “race/ethnic identity” in an IQ test. As the differences between the “races/ethnic identities” is pretty marked we should probably do something about this in all communities. (This conclusion would seem to point to environmental factors not being the cause)


      “race/ethnic identity” is not clearly defined – it is an approximate measure which often correlates with other important environmental factors within a certain population;

      In the population tested there was a difference in the performance for people with different “races/ethnic identities”. Environmental factors are seemingly the cause of the difference in performance. As such this conclusion cannot be applied to other people who might identify with this “race/ethnic identity” outside of this population as the environmental factors may be different. Subsequently, collecting data about “race/ethnic identity” against IQ performance cannot reliably tell us anything about IQ performance outside of the population tested. The results are practically meaningless as “race/ethnic identity” is not something which can be reliably defined and applied outside of the bounds of the population.

      I’m struggling to see the middle ground between these two options and their conclusions.
      It currently seems to be the we are agreeing with the second option but the first conclusion. That is not logically consistent.
      Can you help me out here ?

      I’ve tried not to repeat myself too much from earlier arguments made in this conversation but they do still apply in this line of reasoning. Ultimately the conclusion I am trying to draw towards is that choosing “race/ethnic identity” as a metric against which one would measure anything else does not produce any reliable conclusions. It tells us far less than on the surface it may seem to. The differences observed are between two (or more) arbitrary groups which have no independent affect on the effect being measured. We should be measuring the environmental affects in this type of research instead. “race/ethnic identity” can be used to identify vulnerable groups in a case by case basis, but it has no place as a scientific metric from which you can extrapolate meaningful conclusions.

      • David Didau August 18, 2017 at 5:01 pm - Reply

        It might help you to read this piece from Freddie deBoer: https://fredrikdeboer.com/2017/04/10/disentangling-race-from-intelligence-and-genetics/

        • SamJeeves August 18, 2017 at 5:52 pm - Reply

          Hi David, thank you for that. It was an interesting read and a well made argument. Unfortunately it does not address the concerns raised in my last question at all.

          I am not disputing the data collected . I am not disputing th causes you have hypothesised for the differences in the population tested. My concern – and initial challenge to your article – was in the value (or lack of) using race as a descriptor against which to measure performance (in this case IQ). When I say this, I mean in investigations such as the Gottfredson et.al you were referencing, which are supposed to reveal some sort of useful information which can be applied elsewhere. The conclusions reached from that investigation have ultimately inspired this article in the first place (which is a worthwhile discussion), as they give rise to ‘uncomfortable truths’ which we ‘shouldn’t ignore’.
          Geneticists such as Richard Dawkins and Adam Rutherford have long been saying that “race” or “ethnic identity” are terrible markers with which to do meaningful science.

          I am maintaining that the use of “race/ethnic identity” makes the conclusions drawn in that investigation unreliable at best, but close to meaningless outside of the specific population tested. Many scientists (such as Adam Rutherford) would refuse to even discuss metrics like intelligence in those terms (I.e. vs race). I have given this point much more detail in my previous comment above.



          • David Didau August 18, 2017 at 7:24 pm

            OK. I think you’re wrong. I’ve explained why. The deBoer article made the same points about race so I’d hoped you’d see that is not a point of disagreement. I’m not sure what else there is to say.

  10. sputniksteve August 17, 2017 at 8:46 pm - Reply

    Hello again,

    I’ve finally got around to watching the video from 57 minutes as suggested.

    Is Down’s Syndrome a “disease”?

  11. […] the time of writing, there’s a group of people on Twitter calling me a racist for writing this blog post. This feels pretty awful. The violence of such a term is hard to quantify, but it’s left me […]

Constructive feedback is always appreciated

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: