Whatever the question is, intelligence is the answer

//Whatever the question is, intelligence is the answer

Here are the slides I used in the talk I gave at this year’s Education Festival:

The antipathy of very many otherwise sensible people to the concept of intelligence is really quite remarkable. This aversion seems only to be increased by bringing up the subject of IQ tests. The idea that IQ tests are only useful for showing how good some people are at taking IQ tests is a deeply ignorant view based upon a breathtaking piece of intellectual dishonesty. It’s difficult to believe that people like Professor Guy Claxton who really should know better, are willing to repeat utter garbage like, Angela Duckworth’s construct ‘grit’ is a better predictor of academic success than IQ. This is entirely false. The correlation between ‘grit’ and academic acheivement is only 0.18 whereas an IQ test taken at the age of 11 produces a massive correlation (0.81) with GCSE results. (See here)

In fact, IQ scores seems to correlate positively with pretty much everything we find desirable. Whether you want children to be healthier, happier, safer, more conscientious, more creative, or to live longer, raising intelligence is probably the best route to achieving any of these ends. It even turns out that higher IQ is likely to lead to better mental health – a fact which is routinely ignored by those who believe, in the absence of any evidence that high intelligence is associated with mental illness. One of the only negative correlations that’s been found is that a higher IQ increases your likelihood of needing to wear glasses.

My argument is that increasing IQ is, relatively speaking, not that hard. Although at least 50% of intelligence is heritable, the remaining 50% or so  is a product of our environment. It’s a reasonably safe bet that the bits of intelligence that are heritable are what’s known as fluid intelligence. Crystallised intelligence – our ability to apply what we know to solve problems is most likely a product of our environment. It turns out, much to most people’s surprise, that the aspect of the environment that has the least impact on how we turn out is parenting. What seems to make the most difference is peer culture.

Schools are in a unique position to both mould peer culture and provide the raw material of crystalised intelligence: knowledge. Education has one of the strongest relationships with IQ. Not only does higher IQ lead to better educational outcomes, but more schooling leads to higher IQ. Genetic differences do not determine how children’s lives unfold, but they do have a strong effect. We make choices depending on what we’re good at and children who experience early success in school tend to have a very different academic experience. But our environment can also prevent our genetic potential being expressed. If you have a high fluid intelligence but suffer with undiagnosed glue ear during the first few years of schooling, you may not learn to read with sufficient fluency to comprehend complex texts. An education system can try to flatten the environmental landscape to ensure all children are able to develop to their full genetic potential. In a truly equitable system, all differences between children would be heritable because their experiences would be sufficiently similar so as to ensure all have equal access to a great education.

Not only can education help move the entire bell curve further to the right, individual schools and teachers can make dramatic differences to individual children’s lives.

These thoughts are the basis for a new book I’m writing. Working title: Cleverer.

2017-06-26T07:55:06+00:00June 25th, 2017|Featured|


  1. thomas burkard June 25, 2017 at 3:18 am - Reply

    The great irony here is that the ferocious assaults on Herrnstein, Murray and Eysenck are due to race, yet their orthodoxy that we’re all equal at birth effectively allows educators to blame underachievement of blacks on social factors. You’d think the utter failure of Head Start and other social interventions to produce lasting improvements in educational outcomes after over half a century (and vast amounts of cash) would alert them to the need for providing knowledge and hard skills. Instead, when interventions like Distar showed striking gains in the Abt Associates evaluation of Follow Through, the Ford Foundation supplied grants for them to discredit Abt.

  2. monkrob June 25, 2017 at 7:33 am - Reply

    Am I miss reading this slide or is there a 0.7 correlation between intelligence and schizophrenia? Or is it a 0.7 correlation between intelligence and positive mental health?
    Interested to read your critiques of Claxton. He seems to be a master of the sweeping generalization. Have you written previously about his work? I’d like to read more. He is held up as an “unquestionable” in many Australian education forums.

    • David Didau June 25, 2017 at 9:46 am - Reply

      Sorry – poor phrasing: the higher your IQ, the lower the probability of suffering with mental health issues

  3. chrismwparsons June 25, 2017 at 11:03 am - Reply

    I’m sorry David – having sniffed around this matter in your previous posts, I’m still struggling with a couple of things which don’t quite map-onto each other for me:

    If what we say really matters with intelligence is “The ability to act intelligently in the world”, then I have absolutely no qualm whatsoever with your ‘fluid+crystallised intelligence’ equation, and I fully agree with the value of developing intelligence through learning.

    However, measures of IQ (to my knowledge) are still focused on NVQ tests which are meant to be independent of learning, and are intended to measure the underlying ‘fluid intelligence’. Now, I know that these tests have always been criticised as not being as culture-independent as they were supposed to be, but, they have always been specifically intended to be about some kind of ‘raw’ ability.

    So… assuming that what you are saying is correct, should we all – once age-standardised factors have run-out – expect to see an inexorable increase in our IQ scores throughout our 20s and 30s and beyond, simply because we are learning more through whatever kind of experience/ongoing learning we are doing…?

    • Michael pye June 25, 2017 at 6:30 pm - Reply

      Isn’t there an age adjustment built in?

      • David Didau June 25, 2017 at 7:06 pm - Reply

        To IQ tests? Yes

      • chrismwparsons June 25, 2017 at 7:10 pm - Reply

        What seriously? After the age of 18? I’m not aware of such adjustments. Don’t all of us commenting on these blogs think of ourselves as more educated than we were when we left school? Would we do better on IQ tests?

        • David Didau June 25, 2017 at 8:26 pm - Reply

          Have a look at the graph on slide 9 of my presentation

    • David Didau June 25, 2017 at 8:32 pm - Reply

      No, the king of increasingly specialised domain-specific expertise we acquire in adulthood is unlikely to show up on IQ tests. The relationship with schooling and IQ is pretty robust though. It’s estimated that a year of school can result in 2-3 IQ points, as compared with not being at school. I suspect there’s something about the very general nature of what we acquire in school that seems to make a difference.

      • chrismwparsons June 25, 2017 at 10:06 pm - Reply

        Thank you David – this is a truly interesting perspective.

        To summarise my understanding: Whilst ‘expertise’ should be seen as domain-specific, and there is little to be gained from teaching ‘generic’ skills (other than the slim advantage of meta-cognitive strategies), a broad and intellectual education does – overall – nudge-up general, and domain-independent, intellectual functioning.

  4. Michael Rosen June 25, 2017 at 12:18 pm - Reply

    If what you (DD) are saying is 100% true, do you think there are any steps which schools/education system should take re setting and/or streaming and/or selecting at 4, 5, 7, 11, 14 years of age? Do you think that children and school students should be IQ tested at regular intervals e.g. at those ages, so that schools can check of the pupils are/are not getting cleverer?

  5. Grumpywearymathsteacher June 25, 2017 at 1:00 pm - Reply

    Um, in your slideshow it says, correctly, that correlation does not imply causation….but then in the blog post you say “IQ scores seem to correlate positively with pretty much everything we find desirable…raising intelligence is probably the best route to achieving any of those ends”. What have I missed? Is there some evidence that there _is_ causation here, or is it just that the environments which produce happy, healthy children also produce intelligent ones?

    • David Didau June 25, 2017 at 8:36 pm - Reply

      I’m not sure what you’ve missed, but correlation clearly *does* imply causation, it’s just not the same thing. If IQ positively correlates with everything good then the causal link becomes increasingly hard to ignore. We know that environment only accounts for about 50% of IQ and that almost none of that comes from parenting. It’s school and peer culture that appears to make the difference.

      • Grumpywearymathsteacher June 25, 2017 at 9:51 pm - Reply

        Wooah, hang on, I think you need to be extremely careful making assumptions about what causes what. Also, the variables here appear to be all tangled up and interrelated…
        And I still don’t agree that parenting makes little difference. Just look at the really dysfunctional kids whose parents have done the full Philip Larkin…
        I’m not disagreeing that making people more educated is really important, just very suspicious of ‘statistics’.

      • peter4logo (@peter4logo) July 1, 2017 at 6:56 pm - Reply

        “correlation clearly *does* imply causation” – I do not understand how you mean this, as there are those famous examples of correlations, which are not at all implying causation, taught in every statistics course?

        • David Didau July 2, 2017 at 8:19 pm - Reply

          Erm, ye-es. Pirates & global warming etc. That is why I used the word ‘imply’.

  6. seonaid June 25, 2017 at 4:10 pm - Reply

    Apologies if I have misunderstood, but I’d always thought that IQ tests measured (or at least attempted to measure) fluid intelligence and so it’s fluid intelligence that is correlated with the various positive outcomes, rather than a mixture of fluid and crystallised intelligence.

    • David Didau June 25, 2017 at 7:06 pm - Reply

      No, IQ tests measure both. IQ is an aggregate of Gf & Gc. For instance, tests of vocab & general knowledge are obviously Gc whereas working memory and processing speed are clearly Gf.

      • chrismwparsons June 25, 2017 at 7:15 pm - Reply

        But surely all normative tests of IQ are based on NVI? Aren’t they – by default – designed to try to avoid socialisation, and just measure ‘underlying ability’ – whatever that might be? I admit that my experience in this area is out of date – can you point to more up-to-date ‘IQ’ tests and the basis upon which they are standardised…?

          • chrismwparsons June 25, 2017 at 9:04 pm

            Ok … So, apart from small amounts of numerical and alphabetical processing, the vast majority of IQ tests are about our best attempts at abstract pattern recognition.

            Am I right then that the point you’re making is that…. knowledge based education develops a ‘general’ level of intelligence – abstracted from the domains in which the knowledge originated –
            and which can be applied usefully in all other areas of intellectual thought?

          • David Didau June 25, 2017 at 9:11 pm

            I think that’s a reasonable explanation for the Flynn effect

        • Michael pye June 26, 2017 at 9:50 am - Reply

          What is NVI?

          • chrismwparsons June 26, 2017 at 3:26 pm

            Non-verbal intelligence. I meant to put NVR (non-verbal reasoning).

          • David Didau June 28, 2017 at 9:35 am

            There’s always been a sizeable ‘verbal intelligence’ component in IQ testing.

          • chrismwparsons June 28, 2017 at 10:09 am

            I guess that there is now way we could ever measure fluid intelligence apart from crystallized intelligence, and I know that early intelligence tests which tried to show that indigenous peoples were less intelligent than white Europeans were criticised as being too culturally dependent.

            Nevertheless – David – hasn’t there been a continued attempt to try to focus IQ tests on culture-independent underlying ability? I’m not trying to prove you wrong – I genuinely would like to know! I notice your recommended book on intelligence is over 15 years old – is it as current as it gets? (I want to do some reading)

          • David Didau June 30, 2017 at 7:50 pm

            My first recommendation is this one by Stuart Ritchie https://www.amazon.co.uk/Intelligence-That-Matters-Stuart-Ritchie/dp/1444791877/ref=sr_1_1?s=books&ie=UTF8&qid=1498848597&sr=1-1&keywords=stuart+ritchie but Deary’s book is definitely worth reading also

  7. monkrob June 25, 2017 at 9:29 pm - Reply

    I would like to read more about this statement, “We know that environment only accounts for about 50% of IQ and that almost none of that comes from parenting. It’s school and peer culture that appears to make the difference.”
    As a parent and an educator whose children attended my school I feel the education I gave them at home while growing up was more influential than what they picked up in the 25 hours a week in classes at school. I have no data on this obviously but gut feel makes me question that “almost none” of the 50% environmental factor contributing to IQ comes from parenting.
    Reference please.

    There is nothing like an article on IQ to get the punters questioning. I’m enjoying reading the discourse.

  8. chestnut488 June 26, 2017 at 6:01 pm - Reply

    With reference to your point about people in the past being considered mentally inadequate today – So grade inflation happens because pupils get cleverer year on year ! ( if I knew how to do a wink emoticon I would do one here).

    • David Didau June 28, 2017 at 9:43 am - Reply

      1. It’s entirely obvious that the proportion of those considered mentally inadequate is unlikely to be greater in the past than today.
      2. Grade inflation is, by definition, unrelated to intelligence or ability.
      3. There’s no ‘grade inflation’ in IQ testing because they are renormed every 10 years or so.
      4. What has increased over the past 100 years or so is our ability to think hypothetically and abstractly. This is probably because we have to do much more abstract and hypothetical thinking to be successful in the modern world.
      5. Increases in these kinds of thinking when not tied to increases in general knowledge are of very limited practical value.
      6. A wink emoji is achieved by using a semi colon ; followed by a close bracket ). Like so 😉

  9. peter4logo (@peter4logo) July 1, 2017 at 7:12 pm - Reply

    “Although at least 50% of intelligence is heritable…” – you use another formulation on your slides, saying that 50% of difference in intelligence is heritable. This may be the correct way to express the evidence from studies, as correlation studies in fact can only give this information using R squared.

  10. Matt July 8, 2017 at 1:22 pm - Reply

    What do you think re the affect of family environment on IQ, David? If you have a family that does not view education as important, or there are stressful dynamics at home for whatever reason, would this reduce the impact of education on IQ?

    • David Didau July 8, 2017 at 4:01 pm - Reply

      Family environment has a negligible effect on IQ – washes out to almost nothing by adulthood.

      • chrismwparsons July 8, 2017 at 4:57 pm - Reply

        Still struggling with this!

        Presumably that disproves one of two commonly held assumptions:

        1) That being an avid reader improves intelligence more than just watching TV and kicking a ball around.
        2) That being a child who is surrounded by books, and who has parents who read with you, makes you more likely to become an avid reader than a child whose parents only own some ‘trophy’ books, and prefer to watch TV and kick a ball around with you.

        One of those two must clearly be wrong…

        • David Didau July 8, 2017 at 9:16 pm - Reply

          Yes, these sorts of observations are confounded by heritability. The reason we’re like our parents is because we share their genes.

      • Matt July 10, 2017 at 11:57 am - Reply

        What do you think of studies like these that show children’s vocabulary development to be strongly associated to the quantity and quality of their parents’ vocabulary? Do you think this is the effect of parents and children sharing genotypes that constrain vocab development rather than the characteristics of the vocabulary children are exposed to constraining their development? Seems like both to me?


Constructive feedback is always appreciated

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

%d bloggers like this: