Monday 1 June 2015

Intelligence: All that matters. Stuart Ritchie. John Murray Learning, 2015



Stuart Ritchie is one of the rising stars of intelligence research. Coming from the Ian Deary hothouse, he already has a good publication list to his name. Here are some selected papers drawn from the 28 or so he has published in the last few years:  Brain white matter micro-structure and fluid intelligence in later life.  Is education associated with improvements in general cognitive ability, or in specific skills?   Enduring links between childhood mathematics and reading ability and adult socioeconomic status.  Education is associated with higher later life IQ scores, but not with faster cognitive processing speedBeyond a bigger brain: Multivariable structural brain imaging and intelligence.

Even better than his publication list, Ritchie has an open-minded and optimistic attitude to research. Better still, he has infinite patience with clever sillies, and answers their expostulations and obfuscations with gentle reference to good quality contemporary research.

These characteristics serve him in good stead when taking up the unenviable task of bringing Ian Deary’s “Intelligence: A very short introduction” up to date. Although it is a hard act to follow, Ritchie proves a very worthy successor, capable of explaining without condescending, and confident enough to develop his own prose style, which is clear, direct, uncluttered and at the service of his intended aim: to give an up to date summary of the modern science of intelligence.

The book clears up confusions briskly, with a good understanding of the essentials. The almost compulsory and off-putting walk through the graveyard of past researchers is dumped in favour of simply picking up the important themes and amplifying them. Edinburgh is making history in intelligence research, so why dwell on anything but the best ideas?

The content sticks to the brief: all that matters. The concept of intelligence is introduced, testing explained, why intelligence matters explained, the biology laid out (virtually all of this brand new research), the “boost your IQ” meme dissected (stay at school longer?) and the IQ controversies patiently listed and their varying claims teased out and answered. In no other scholarly topic have the protestors and hecklers against intelligence been given such an easy ride in the popular culture, so it is good to see them wait their turn at the end, not poison the well of research before anyone can drink from it.

I think that it is unwise to summarize a book which is itself a summary of contemporary research and debate, so I will make only a few detailed comments. You will be better off reading the book than reading any further exposition about what is already a concise book.

Quibble: I am not as convinced as Ritchie that the Norwegian “experiment” in increasing the years spent at school really boosted IQ by 3.7 points a year.  I think this claim is based on the “difference in difference” statistic which I find problematical. The matter is statistically complex, so I would not argue my view over Ritchie’s with any great confidence, but it seems an unlikely result.

Quibble: On the contentious matter of racial differences in intelligence, although I agree with Ritchie that the data are not good enough to resolve the matter beyond reasonable doubt, I think that on balance of probabilities about half of racial intelligence differences are probably due to genetics. However, I can see that it is hard to get very far with this topic in a brief introduction.

Additional thanks to Ritchie for: Chapter 6, discussing why intelligence is so controversial, and doing so in a fair and straightforward manner; giving good reasons as to why it is worth studying intelligence (health, ageing, societal impact, scientific curiosity); listing 15 intelligence books; listing 5 websites (double thanks for that); 5 commonly used IQ tests, 20 intelligence researchers working today, and their research areas; 10 review papers you must read; 5 surprising things related to intelligence; 5 historical intelligence researchers apart from Galton and Spearman;  5 anti-IQ books; 10 common myths (all debunked in the book); 5 fictional bright people; 5 fictional dull people; and 10 big questions for future research. In fact, Chapter 6 is a mini-book in itself.

In my opinion “Intelligence: All that matters” is the best available short introduction to intelligence, and word for word the most effective. It shows a keen understanding of the misconceptions which bedevil the public understanding of the subject, and the scholarly benefits of replying to these wild imaginings with cool evidence. Indeed, as many intelligence researchers know to their cost, the word “intelligence” carries so much baggage which has been heaped upon it that there is a temptation to retreat to euphemisms: cognitive abilities, learning styles, executive functions, and learning readiness. Anyone who reads this book will understand that intelligence is real, and has important consequences. This is intelligence re-claimed.

I hope the fact that Stuart Ritchie knows a great deal about his subject will not stop his book from being sold in very large quantities. So many other books about intelligence are effusions of babbling, evasion and misapprehension that a work of education imparted by a real researcher deserves a very wide readership. As Ritchie concludes: the intelligent way forward is that which helps us uncover the science of what makes us differ in this most human of attributes.

http://www.amazon.co.uk/Intelligence-That-Matters-Stuart-Ritchie/dp/1444791877/ref=sr_1_1?ie=UTF8&qid=1433196561&sr=8-1&keywords=Intelligence+all+that+matters




22 comments:

  1. Good that a book like this is out there.

    Now I'm going to level my criticisms, for the reason that it's important to get corrections to the message out there quickly with the message itself.

    On the title, no, intelligence is not "all there is":

    Predictions on the Worldwide Distribution of Personality | JayMan's Blog

    start here | hbd chick

    But sure, compared to the usual drivel (schools, good neighborhoods, computers in the classroom, good teachers, etc), yes, intelligence matters and those things do not.

    "Quibble: On the contentious matter of racial differences in intelligence, although I agree with Ritchie that the data are not good enough to resolve the matter beyond reasonable doubt"

    Come on now, that's complete nonsense. What about brain size? What about the failure of interventions? What about behavioral genetic studies, like transracial adoption studies (I mean the newer ones that look at international children adopted in various European countries)? What about genetic ancestry-educational attainment studies like those over at Human Varieties? And most of all, what about Occam's Razor?

    But I guess Ritchie wants to choose his battles. Fair enough.

    "I think that on balance of probabilities about half of racial intelligence differences are probably due to genetics."

    Maybe for sub-Saharan Africans in Africa. Probably not so much for anyone else.

    ReplyDelete
    Replies
    1. On the title, no, intelligence is not "all there is":

      "All That Matters" is the name of the book series.

      Delete
    2. And a clever name too; not that many people will notice.

      Delete
    3. Like most people, you haven't defined what you mean by race. If you are simply talking about taking salient physical features (skin colour) and extrapolating that to some underlying genetic variation that ought to affect performance on IQ tests, then I think that is a leap to far. i.e. there is more genetic variance within 'races' than between.

      Also, in my own country (NZ) these salient physical differences have been found to moderate outcomes like you describe. e.g. Maori who look more Pakeha, get jobs/mortgages with greater ease than those who look more Maori. These are just a few of the problems with the psuedo-science that underlies some of your (by definition) racist views.

      Delete
    4. Simon, if you are evidence based you will wish consider views in addition to those proposed by Lewontin before modern genomic analysis became available. You repeat his claim "there is more genetic variance within races than between" without looking at modern research on ancestry. Even the NHS makes corrections for ancestry in the analysis of blood tests. If you just want to have a stale argument and be insulting then that is up to you. If you want to look at contemporary research then there are many sources of information. Razib Khan is a very good guy on these matters. Have a look at recent results before making unpleasant accusations.

      Delete
    5. In fact, this would get you started on current arguments
      https://westhunt.wordpress.com/2016/04/19/such-a-thing/

      Delete
  2. Hi JayMan,

    You raise many topics, each of which would probably be worth an entire post on its own, but I think that if we look at just one, we can get an idea of why someone might want to remain agnostic about some of the classic HBD claims. So let's look at your first question: "What about brain size?"

    On its face, it does seem that brain size would need to have some functional relationship to intelligence. Though the brain is a costly organ, made costlier with greater size. But after millions of years of evolution, apparently in the direction of superior tool use, social skills, and problem solving, Homo s. sapiens has a much higher encephalization quotient than other mammals.

    Yet anatomically modern brains are a step down from Neanderthal and Cro Magnon, even while intelligence appears to have increased. That modern humans are more intelligent than either of these earlier groups may not be clear, but it does raise the question of whether "there's no replacement for displacement."

    Looking at more modern data, we also find a more direct example of groups with strong brain size disparities, but intelligence gaps near zero: men and women. Brain size differs quite significantly between the sexes - by around 11% - yet even Richard Lynn doesn't try to argue for more than a 1/3 sigma intelligence gap between the sexes.

    In fact, one could even make a case that, within humans, brain size variation has no important functional relationship with intelligence: See Schonemann's 2000 piece, "Brain size does not predict general cognitive ability within families," an article which is very much what's on the tin. Let me stress that I don't think that any such case is right; one surprising study with 36 sibling pairs isn't much to go on. But clearly, such studies do exist.

    Ultimately, agnosticism is always more difficult and less exciting than the alternatives. Some people are never satisfied until they've taken a side. But for those who have the temperament to wait in the middle, reasonable doubt can be entirely reasonable.

    ReplyDelete
    Replies
    1. >Yet anatomically modern brains are a step down from Neanderthal and Cro Magnon, even while intelligence appears to have increased. That modern humans are more intelligent than either of these earlier groups may not be clear, but it does raise the question of whether "there's no replacement for displacement."

      I wouldn't make that argument. http://rspb.royalsocietypublishing.org/content/280/1758/20130168.full

      >Looking at more modern data, we also find a more direct example of groups with strong brain size disparities, but intelligence gaps near zero: men and women. Brain size differs quite significantly between the sexes - by around 11% - yet even Richard Lynn doesn't try to argue for more than a 1/3 sigma intelligence gap between the sexes.

      It isn't because he is holding back, it is because statistics. You see if brain size x IQ is r = .33 (controlled for body size, which is not always done), then a brain size difference of about 1 SD means that the estimated difference in IQ is 1 * .33 = .33 d, which is 15 * .33 = 5 IQ points. Lynn is arguing for a difference around 4.5 IQ, so...

      >In fact, one could even make a case that, within humans, brain size variation has no important functional relationship with intelligence: See Schonemann's 2000 piece, "Brain size does not predict general cognitive ability within families," an article which is very much what's on the tin. Let me stress that I don't think that any such case is right; one surprising study with 36 sibling pairs isn't much to go on. But clearly, such studies do exist.

      Why are you bothering with a sophist when we have huge meta-analysis? http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2512128 Note: these are not controlled for body size. Median brain size x IQ is

      >Ultimately, agnosticism is always more difficult and less exciting than the alternatives. Some people are never satisfied until they've taken a side. But for those who have the temperament to wait in the middle, reasonable doubt can be entirely reasonable.

      A wise man proportions his belief to the evidence. (Hume). Sometimes this means holding back when others believe. Other times this means believing when others hold back.

      Delete
    2. Forgot to add the number. Median brain size X IQ is .28, weighted mean .24. These numbers are from healthy participants only and only adults using data from the large meta-analysis. The numbers are not corrected for body size, so they are somewhat lower. You also want to correct for measurement error. Assuming very little error for MRI (altho they don't frequently report the test-retest reliabilities), and perhaps .90 for these fullscale IQ tests, then we get something like .25 or .30 depending on which value was chosen.

      Delete
    3. Mark,

      The silliness of your response (no offense to you intended) kinda makes my point for me. Emil's excellent refutation aside, on this point:

      "Yet anatomically modern brains are a step down from Neanderthal and Cro Magnon, even while intelligence appears to have increased. That modern humans are more intelligent than either of these earlier groups may not be clear, but it does raise the question of whether "there's no replacement for displacement."

      Because human brains have apparently gotten more efficient over that time. You're not seriously going to argue that the smaller-brained races of Earth today also happen to have more efficient brains, leaving their cognitive abilities identical to the larger-brained races – even though the former also test lower as well?

      See Greg Cochran on it:

      How do we do it? Volume! | West Hunter

      But putting aside all that, what about all the other key facts?

      Like I said, Occam's Razor.

      Delete
    4. Forgot to add the number. Median brain size X IQ is .28, weighted mean .24.

      I don't think that meta analysis corrected for range restriction. Had they done so the correlation would likely have been at least 0.1 higher. There was a massive study on head circumference in children (thousands of participants) that found a 0.23 correlation between IQ and HC, and given that HC is just a crude proxy for brain size, I am sure a direct measure of brain size correlates close to 0.4 with IQ, just as Jensen claimed.

      Delete
  3. Emil: "You see if brain size x IQ is r = .33 (controlled for body size, which is not always done), then a brain size difference of about 1 SD means that the estimated difference in IQ is 1 * .33 = .33 d, which is 15 * .33 = 5 IQ points. Lynn is arguing for a difference around 4.5 IQ, so..."

    So, if brain size is the key to understanding group differences in intelligence, one might well expect that any ethnic disparities in intelligence should be of a similar order to the sex differences in intelligence. Yet the sex difference in brain size is greater than ethnic differences in brain size, even when the sex difference in IQ is much smaller. (See for example Jensen 1998, p 439, where height adjusted brain weights are given white males at 1392g, vs 1290g for black males and 1252g for white females). Your point about Neanderthal brains is quite apt, here - raw brain size, even when controlling for body size, is not necessarily what accounts for any differences. Given such complexities in the subject, it's not at all hard to imagine why someone might withhold taking sides on the issue.


    JayMan: "The silliness of your response (no offense to you intended) kinda makes my point for me."

    Let me try again.

    One may, without endorsing a given stance, still regard that stance as reasonable or understandable, rather than decrying it as nonsense and demanding that a long list of objections be addressed.

    Since you should be well aware by now of the way in which your own views are caricatured, vilified, or dismissed, by basically everyone in the mainstream, why in the world would you want to behave in the same way? Moreover, why do this to someone who like Richie who takes the issue seriously and even admit that your position might be correct? You're so primed to divide the world into people who agree with you and people who don't that you even approach a third party discussion (with me, whose views on the subject of group disparities you clearly don't know) with your rhetorical guns blazing.

    But congratulations on my making your point for you. :p

    ReplyDelete
  4. Nice polite response, Mark.
    The brain size excess for men is 11% by the numbers above, but body weight is 20% for the US, 22% UK, 22% Korea. So shouldn't women be the smarter? Of course, it would probably be a concave relationship since many things controlled by the brain don't rise with body size. Anybody know if men and women have the same size cerebellums? (cerebelli sounds pretentious)

    ReplyDelete
  5. The paradox about sex differences in brain being greater than for intelligence has, to my imperfect recollection, been attempted to be resolved in three main ways. 1) Men are in fact brighter than women, though girls are brighter than boys, because men mature much later (Richard Lynn's work) 2) Men's greater brain size is due to better visuo-spatial skills, which ought to be given greater weight in intelligence testing (JP Rushton) 3) sex differences items have been dropped from intelligence tests, thus blurring real sex differences in intelligence, which Nick Mackintosh mentioned to me, and probably published on somewhere)

    ReplyDelete
    Replies
    1. I have yet to see a study that sufficiently controlled for body size. I've seen studies that controlled for height and body surface area, but even men and women of the same height and body surface area will differ in lean body mass.

      To me the differences in brain size between men and women are somewhat analogous to the brain size differences between big brained robust Neanderthals and smaller brained gracile modern humans.

      Delete
  6. Remember the ancient Chinese proverb: it is not necessarily the car with the biggest engine that goes fastest.

    ReplyDelete
  7. your article is very nice i really like this article very much.online supply chain management degree.

    ReplyDelete
  8. Thanks so much for the inclusion. Honored to work with many of the great bloggers in this group!

    Custom Essay Writing Services

    ReplyDelete
  9. This comment has been removed by a blog administrator.

    ReplyDelete
  10. I'm not sure that I agree with your assessment of this book. Within the first two sentences and even on the back blurb Stuart either a) subtly changes the topic of the book from 'intelligence' to 'IQ' or b) severely conflates these terms (like your review). It is important to note that whilst intelligence and IQ are related (assuming you adhere to the idea that intelligence (aka 'g') exists) they are not the same thing. i.e. intelligence underlies performance across a number of cognitive domains, which (provided we have collected data across a great number of domains) we can then use to make an inference about IQ. Note that IQ relies not only on intelligence but many other things such as effort, familiarity with tests etc.

    Also, a common assumption that people such as Richie seem to make is that IQ testing actually tests cognition directly. It doesn't. IQ testing (actually i prefer the term cognitive testing: its more accurate) involves making inferences about cognitive function on the basis of observable behaviours. That is, cognitive tests do not directly measure cognition: they measure
    behavior from which we make inferences about cognition.

    When you frame the dialogue in this way (which I think is the correct way, and certainly the way I was trained in terms of neuropsychology), it makes it much harder for Stuart to make the kind of grandiose claims that he does.

    I don't mean to sound adversarial, but there is a context here that is sorely lacking.

    ReplyDelete
    Replies
    1. Which claims do you find grandiose?

      Delete
    2. Any claims that he makes about 'intelligence' on the basis of a handful of cognitive tests.

      Delete