Book Notes: Intelligence — All That Matters

Intelligence: All That Matters by Stuart Ritchie

[Disclaimer:  This is not meant to be a book summary or book review. This is just stuff in the book that I found personally valuable or interesting at the time of reading. Most of these “notes” are actually highlights, i.e. directly copied lines from the book, but some notes are personal adaptations or added personal insights.]

NOTE: The subtitle of this book is yet another provocative title/subtitle that is obviously misleading! The title should be Intelligence: It’s Complex or Intelligence: It Generally Matters Quite a Bit

– ‘How do you even define intelligence?’

One attempt at a definition (Gottfredson, 1997):

Intelligence is 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.

– The first steps [in testing inteligence] were made by a true genius: Charles Darwin’s half-cousin, Sir Francis Galton.

Galton began to reflect on ‘eminence’. What was it that made some people rise to the top in society? Why did some families seem to have many eminent individuals while others had none?

In attempting to answer these questions, Galton weighed up the effects of socialization and heredity, and coined the phrase ‘nature and nurture’.

– The first true intelligence tests came not from studying eminent individuals with the highest ability levels, but instead those at the bottom of the distribution.

In 1904 the psychologist Alfred Binet (1857–1911) was tasked by the French government with creating a tool to identify children with what we’d now call learning disabilities: children who would need special educational attention. Binet wanted something that wasn’t subjective like psychiatrists’ reports, and something that could group children into different levels of disability.

– Hanging over all of this early research, from the very start, was the spectre of eugenics.

Francis Galton invented the term eugenics. [Eugenics: “selectively ‘breed’ humans to produce healthier, smarter future generations – either by encouraging people in higher social classes to have more children (‘positive eugenics’), or by discouraging the lower classes from reproducing (‘negative eugenics’)”]

It’s reasonable to be appalled at the results of the eugenics movement.

In any case, we must separate out the political beliefs of these early researchers from any facts they discovered about human psychology.

Facts, after all, have no necessary moral or policy implications: it is up to us to decide what to do about them once they are discovered.

– One of Francis Galton’s most enduring inventions, correlation is a way of quantifying the strength of the relation (the ‘co-relation’) between two ‘variables’ (a variable is just anything that can be measured).

We know that taller parents are very likely to have taller children, and we also know that there are some exceptions to this rule. In statistical terminology, we’d say that there’s a ‘strong positive correlation’ between parent and offspring height: ‘strong’ because of the high likelihood of finding tall parents with tall children (but not ‘perfect’, which would imply that taller parents always have tall children); ‘positive’ because as one variable (parental height) increases, the other variable (offspring height) tends to increase, too (if one variable decreased as the other increased, the correlation would be ‘negative’).

Correlations are sometimes referred to as ‘associations’, ‘links’, ‘relations’, or, if one variable precedes the other in time, ‘predictions’, but these all refer to the same thing.

– If intelligence and educational achievement are correlated, this might mean (a) that intelligence causes you to do better in school, (b) that schooling helps you do better on intelligence tests, (c) that something else, perhaps social background, causes you to do better at school and on intelligence tests, or (d) a mixture of all the above.

Nevertheless, we shouldn’t forget that a correlation does sometimes imply causation: it’s difficult to think of any instances of causation where there’s no correlation.

– Francis Galton created a statistical formula that puts numbers on these correlations, indexing the strength of the relation. The numbers range from –1.00, a perfect negative correlation, to 1.00, a perfect positive correlation. The closer to 1.00 (or –1.00 for negative correlations), the stronger is the relation. A correlation of zero indicates no relation between the variables whatsoever.

The italicized letter r is used to denote a correlation (in the form ‘r = 0.50’).

– Beware online tests that claim to provide your IQ after a ten-minute set of puzzles: they bear little relation to real intelligence tests and are usually pretty worthless.

– IQ Scores will tend to vary across time (perhaps during one test you didn’t try as hard as you possibly could, or were staring out of the window when the examiner asked you the question), but for each person multiple scores will cluster around their ‘true’ ability level. The more tests you take, the better idea you will have of what this ‘true’ ability is.

– Can you ‘game’ an IQ test by practising it beforehand? Nobody would deny that, if you practise something, you get better at it. But IQ tests are designed to be used with the general population at large.

Importantly, even if everyone practised the tests, there would still be differences in the speed it takes to learn them. This is what we might call intelligence.

– One of the most important findings about intelligence: the g-factor.

Contrary to many people’s intuitions, it turns out that, to a great extent, intelligence is general. In other words, people who are good at one kind of mental test tend to be good at them all.

The positive manifold of seemingly unrelated tests is one of the most well-replicated findings in psychological science.

– In the 1980s the psychologist Howard Gardner argued that, instead of a general factor of intelligence, there are in fact multiple intelligences.

He argues that people differ on this whole range of intelligences, and that the intelligences don’t necessarily correlate together.

There’s just one problem with this theory: there’s no evidence for it.

The notion of ‘multiple intelligences’ has become very popular among educators, as a kind of wishful thinking: if a child has poor (say) logical-mathematical abilities, the argument goes, they might still be good at another kind of intelligence! But denying the huge amount of evidence for general intelligence does nobody any favours.

Many of Gardner’s ‘intelligences’ can be conceptualized as skills, or even personality traits, but to describe them as ‘intelligences’ makes a mockery of the definition of the word.

– Psychologists use a statistical technique called factor analysis to look for patterns in complicated data. [Eg measures of fitness, such as ability to do more push-ups correlating with ability to run faster, jump higher, etc]

The reason the measures correlate together might be that there is something shared between all of them – something that makes them generally high in one person and generally low in another. That something is what we might call ‘fitness’.

Factor analysis allows us to extract this shared aspect from the assortment of correlated measurements. It works in exactly the same way for intelligence tests.

– The positively correlated IQ subtests … have something in common, and this is what [Charles Spearman] called the ‘general factor of intelligence’, the ‘g-factor’, or sometimes just ‘g’, for short.

Spearman thought that no single subtest was a perfect measure of g, but that when factor analysis was used to isolate only what a range of subtests had in common, this gave an insight into a person’s all-purpose intellectual power.

Averaging across the wide variety of tests, the usual result is that differences on the g-factor explain around half of people’s overall intellectual differences.

– Regardless of how you test someone’s intelligence, there’s ‘just one g’.

But what is the g-factor? Spearman proposed that g reflected something like ‘mental energy’: one single property of the brain that varied between people, reflected better or worse genetic make-up, and determined whether their test scores were generally high or low. This would be as if one single process, like muscle strength, were responsible for differences in all [common fitness measures].

Some modern psychologists have followed in Spearman’s footsteps, theorizing that processing speed might be the key to this mental energy. They argue that the brain’s efficiency is at the root of all mental abilities: faster brains get more done in a brief time frame, and can thus pack in more operations (Jensen, 2006).

Other researchers have suggested that working memory … might be the basic mechanism that underlies g.

Whatever its cause, there is a general factor of intelligence; comprehensive IQ tests measure it and … it matters.

– One of the most popular models splits general intelligence into two subgroupings, called ‘fluid’ and ‘crystallized’.

Fluid intelligence involves working things out on the basis of no previous knowledge. For example, the matrix reasoning test or the reaction time test can be given completely ‘cold’ – people don’t need to use any prior experience to solve the puzzles or to react to the light.

On the other hand, ‘crystallized’ intelligence is made up of the sorts of tests (like vocabulary and general knowledge) that rely on information gained outside the testing environment, in the real world.

[Short version for remembering fluid and crystallized intelligence:

Fluid intelligence — Intelligence as ability to efficiently process new information and problem solve

Crystallized intelligence — Intelligence as the coherent applicability of learned/acquired knowledge]

– Arguments between psychologists over the substructure of intelligence show no signs of abating.

– Crystallized abilities continue to rise for many years, as we accumulate more knowledge (this is perhaps what people mean when they talk about the ‘wisdom of age’).

From around the mid-twenties, fluid intelligence declines, and continues to do so right into old age. … Ageing makes our brains more sluggish, and we progressively lose the ability to solve problems efficiently and hold new information in our working memories.

(Ritchie, Figure 2.4; Kindle Loc 404):
‘Fluid’ skills (dashed line) decline far more steeply than ‘crystallized’ ones (solid line). Tests of speed (dotted line) show a particularly pronounced decline.

– The decline in intelligence matters.

It happens to people whether or not they go on to develop dementia. In later life, it’s been linked to a drop in the ability to make important decisions and perform everyday tasks like reading labels, managing finances, and remembering to take medicines (Tucker-Drob, 2011).

It seems that encouraging older people to be more physically active can preserve their fluid intelligence (Lövdén et al., 2013).

But the biggest predictor of being smarter in old age is rather obvious, and perhaps somewhat disheartening: being smarter in childhood.

(Ritchie, Figure 2.5; Kindle Loc 425)
The evidence on the stability of intelligence tells us two things:

First, whereas the correlation across the lifespan is quite strong, it is far from r = 1.00. Many of the reasons some people age better than others are not to do with their prior levels of intelligence: we just need to find out what they are. Second, we might want to look at the period of life before that age-11 intelligence test, to see if we can boost intelligence in childhood.

There’s no single way to assess someone’s intelligence.

The most important fact is that, despite the sheer diversity of the tests, people doing better at one test usually do better at them all.

This ‘positive manifold’ of test correlations gives rise to the general ‘g-factor’ of intelligence, which has a number of different interpretations –it might reflect the general speed of brain in performing all its tasks, or a more complicated interaction of many different basic abilities. To get a good measure of it, you need to give a wide variety of different cognitive tests.

Some – but not all – of the functions measured by intelligence tests show a precipitous decline as we get older.

Smarter people live longer. In study after study, it’s been found that people with higher IQ scores –and, thus, higher general intelligence –tend, on average, to outlive their less intelligent peers.

In some studies, IQ score is about as predictive of the risk of death as is smoking.

– IQ test scores are strongly and unambiguously correlated with two different, though related, measures of education:

achievement (that is, your exam results) and duration (that is, how long you stay at school, and whether you go on to college or university).

– Intelligence is a major predictor of a whole host of outcomes. But it’s by no means the only one.

Take educational achievement. Being bright is critically important, but so are other characteristics like conscientiousness, motivation and self-control, as well as various social skills.

It’s been found that, for instance, motivation levels and intelligence both have their own separate predictive effects on school performance (Ritchie and Bates, 2013).

The existence of these studies should instantly refute the silly but commonly heard criticism of IQ research that states: ‘You can’t sum up a person in a single number!’ Nobody with any sense ever claimed you could.

– We’ve already begun to answer … why higher intelligence predicts a later death. Not only will better-educated people be more aware of how to stay healthy, but more qualifications allow entry into better jobs, which bring all the health benefits that more income provides.

– A huge number of scientific studies now attest to the fact that intelligence tests are predictive of performance in the workplace.

– Smarter people don’t always do better, and there are always other non-intellectual factors that contribute to good performance. But if employers were to select workers using intelligence tests – as many already do – they’d be more likely to get a more productive workforce.

There simply isn’t another tool that gives such a good general prediction of how people will do at work across such a variety of jobs.

– However you want to measure it – and the best measures of class will take into account many different factors – your score on an intelligence test will correlate with your social class: smarter people tend to end up in higher social classes.

As with job performance, the relation of class to IQ is nowhere near as strong as that for education: the correlations are usually in the region of r = 0.30 to 0.50 (Strenze, 2007).

– [Argument against intelligence researchers:]

But don’t children who start life in lower social classes tend to have poorer life chances because of their poorer environment?

To put it another way, maybe it’s not that IQ causes better job performance but, instead, higher social class causes both better job performance and higher IQ.


Social class can indeed have important effects on IQ … nobody has ever claimed that intelligence is the only thing that matters in the world of work: numerous other factors make a crucial difference.

Brighter people tend to do healthier things: they exercise more, eat better and are less likely to smoke (Gottfredson, 2004).

– Cognitive epidemiology (the study of the links between intellectual abilities and health and disease) …

studies in this relatively new field have repeatedly found correlations between health and intelligence:

smarter people are somewhat less likely to have medical conditions, like heart disease, obesity or hypertension, that decrease life expectancy. This is found for mental as well as physical health: large-scale studies have shown that those with lower intelligence test scores are more likely to be hospitalized for psychiatric conditions (Gale et al., 2010).

– It’s no surprise, then, that there’s such an impressive link between intelligence and mortality.

[Swedish study of almost a million men] People in the lowest of the nine IQ categories were over three times more likely to die in the 20 years after their testing session than those with the highest IQ scores.

A review study, taking into account all the evidence, reported its results in terms of IQ points, rather than categories: they concluded that being 15 IQ points higher in childhood led to a 24% lower risk of death in the subsequent years (Calvin et al., 2010).

– As I said at the outset, the connection between intelligence and mortality is complex. All of the factors we’ve discussed above –education, jobs, social class, health, genetics, risk perception, and more –are likely to play a role in a complicated chain of cause and effect.

– We shouldn’t ignore the essential role that the g-factor plays in creativity.

It seems that higher spatial ability – measured by, among other things, the ‘mental rotation’ task … seems to be particularly helpful (Wai et al., 2009). The kind of abstract visualization that’s required to manipulate shapes in one’s mind seems to be essential for thinking through new scientific questions.

– IQ tests are indeed related to political and religious beliefs.

Research shows that higher intelligence scorers tend to be less racist, less sexist, and less accepting of authoritarian parenting and policing: that is, more socially liberal.

– Left-leaning readers shouldn’t feel too smug upon reading this, though: the evidence seems to show that higher-IQ people are economically as well as socially liberal, tending to be in favour of freer markets (Carl, 2014).

– Only just over 2 per cent of people get scores above 130 on a properly ‘normed’ IQ test

– There is, however, one negative trait that’s consistently linked to higher IQ: short-sightedness (Teasdale et al., 1988). It really does seem that smarter people are more likely to need glasses.

– We can assume that, to a great extent, [human] intelligence is an adaptation. In order to survive and successfully reproduce in Darwin’s ‘struggle for life’, our ancestors would have had to adapt to particular challenges and pressures found in their environment.

– We can know whether there are genetic effects on intelligence (from quantitative genetics) without knowing any of the specific genes involved (that is, without knowing the molecular genetics).

50 per cent of the differences in intelligence are due to genetic differences.

In behavioural genetic terminology, intelligence is 50 per cent ‘heritable’ (of course, behaviour genetics uses many more complex methods than just this one formula: for the details, see Plomin et al., 2013).

That is, half of the reasons why people vary on intelligence test scores are genetic.

Intriguingly, it’s been found that the genetic effect on intelligence is stronger in adults (heritabilities of up to 80 per cent) than it is in children (around 20 per cent), suggesting that our biology becomes more important for our intelligence as we age (Plomin and Deary, 2014).

– ‘Heritable’ doesn’t mean the same thing as ‘hereditary’. These words are often confused: a ‘hereditary’ trait is simply anything that’s passed on from parents to offspring, whereas ‘heritability’ is about the genetic variation in that trait.

– ‘Nurture’ implies parenting, but non-genetic effects on intelligence are far broader than that.

– What’s surprising to many is that, taking into account all the data from twin studies, the shared environment appears to have a very small effect on intelligence.

If you measure intelligence in adulthood, almost all of the variance is explained by a combination of genes and the non-shared environment. Outside of cases of abuse or neglect, the things that parents do don’t seem to have a strong effect on their children’s intelligence, in the long run (Harris, 2009).

To put this another way, the main reason smart parents tend to have smart children is because of the genes they pass on, not because of their parenting decisions.

– There is no single gene, or even handful of genes, for intelligence (though see the box below for some exceptions to this rule). Instead, the intelligence-linked genes turn out to be many thousands of tiny needles in the massive haystack of the human genome.

– A common mistake is to come away with the impression that, since intelligence is related to biology, it must be immutable. Nothing in the genetic studies (which never show 100-per-cent heritability) or the neuroimaging (which shows only neural correlates of intelligence) leads to that conclusion.

There may be biological limits on what we can expect from some people: although intelligence is not immutable, it is unlikely to be infinitely malleable.

– All sorts of claims are made about what parents can do to raise their child’s IQ: having books in the home, reading those books to their child, using complex vocabulary in conversation, et cetera. The problem is that all these things tend to be done by cleverer parents, who have already passed on the genes that will make their child more intelligent.

Most studies that find a link can’t tell whether the parenting or the genetics caused the higher IQ. This is not to say that these parenting activities definitely don’t have any effect on intelligence: it’s just that we can only figure out whether they do if we take genetics into account.

– Longer schooling, as well as factors like improved nutrition, are beneficial for intelligence. We also know that, for the last century or so, educational opportunities and diets have been improving generation upon generation, at least in the Western world, for almost everyone. Given these points, shouldn’t people be brighter now than they were, say, at the start of the twentieth century, just as the average person now is taller than their great-grandparents?

In fact, this is exactly what the evidence shows. Intelligence-test performance has been getting better by around 3 points per decade since IQ testing began, in a phenomenon known as the ‘Flynn Effect’.

It’s named after the political scientist James Flynn, who has written about it most extensively.

As a consequence, tests now have to be continually ‘re-normed’, with their score of 100 being set to the average of a sample of contemporary people.

– Flynn argues that improved nutrition and education are only part of the explanation of his eponymous effect: a more important reason is because the way we think, as a society, has undergone a revolution across the past 100 years.

So, at least in Flynn’s conception, the average person in the past just wasn’t accustomed to the abstract thinking we do so effortlessly in our modern, Western mindset.

Add this shift in thinking style to improved health, nutrition and education, and we begin to explain the long-term rise in IQ, although we might still need to look at some other factors to fully understand it.

– There is no clear evidence that the [Flynn] effect is stopping, even in very rich countries like the US (Trahan et al., 2014). Whatever it is that’s pushing up IQ scores – be it computer use, improvements to education, a cleaner environment, or something we haven’t thought of yet – it seems it’s still having an effect.

– The glimmers of hope for our ability to raise intelligence might also help put to bed one of the most consistent misunderstandings of intelligence: that IQ is immutable.

The jury is in. The scientific evidence shows that general intelligence can be measured, that it matters in life, that its variation is related to differences in our brains and in our genes, and that we should be at least optimistic that it can be improved.

– What intelligence research (and differential psychology more generally) tells us is that people’s minds are different from one another in measurable ways, that these differences are important in their lives, and that they might be influenced in part by biology.

Genes don’t explain 100 per cent of IQ differences, and the environment has a substantial role. Moreover, there is nothing in principle to stop us from improving intelligence, at least to a degree.

Indeed, there’s compelling evidence that IQs have already been rising across the twentieth century and into the twenty-first, as part of the Flynn Effect. And we know that a host of other factors besides intelligence, like personality, effort, motivation and a bit of luck, are vital for a successful life.

– ‘Why do we need to know any of this stuff? What’s the point of researching differences in people’s intelligence?’

There are many answers to this, and a number of them have popped up throughout the book. The simplest one is: ‘because intelligence is real, and if we don’t research it, we’re missing out on an important fact about psychology’.

– Intelligence research can help us to identify the most talented people who’ll foster the artistic and scientific innovation that drives society forwards. At the other end of the spectrum, a deeper knowledge about the nature of general intelligence, and how to measure it, can only be beneficial for diagnosing and helping people with intellectual disabilities.

Intelligence is one of the foremost parts of what makes us human: knowing exactly what it is, how it’s instantiated in the brain and DNA, and what effects it has on people’s lives is part of understanding the story of our species.

– Intelligence shouldn’t be so controversial. … Most of the debate surrounding intelligence stems from some basic misunderstandings of the research. The facts about IQ don’t justify unfairness, and nor do they mean that everyone is stuck for ever with the same ability to think and learn.

Ritchie, Stuart. Intelligence: All That Matters. Hodder & Stoughton, 2015. Kindle Edition.