The average IQ in the UK is 100, and the reason is the ruler

0 Quick Answer

The average IQ in the UK is 100 by construction, because the British samples on which Raven's Progressive Matrices were standardized in 1979, 1992 and 2007 are the reference scale that Lynn and Vanhanen adopted for every country and that Becker's dataset still uses; the small departures from 100 in the published tables are what happens when a handful of British samples are scored against norms drawn from other British samples in other years. Becker's national IQ dataset, version 1.3.5 dated October 25, 2023 and downloadable from View on IQ, gives the United Kingdom 99.66 on its combined measure, from six samples totalling 5,794 people, and ranks it 19th of the 157 countries with a value. Lynn and Vanhanen's IQ and the Wealth of Nations set Britain at exactly 100 in 2002, as the reference point for the whole table, and Becker's record of their 2012 revision gives 99.1. The self-selected International IQ Test table that the ACIS page on average IQ by country uses shows the United Kingdom at 101.57 from 7,310 takers, on a scale whose overall mean the site fixes at 100.

The comparison with average IQ estimates for the USA depends on whether the figures share a reference scale and an appropriate sampling design.

Those three figures are three ways of saying the same thing: on the scales in use, the British population is the population the scales were built to describe, and it therefore sits at the centre of them. That is not a finding about British ability. It is a fact about the ruler, and it explains why the more interesting British evidence is not on the IQ scale at all. Scotland tested almost every eleven year old in the country in 1932 and again in 1947, the only occasions on which a whole national age cohort has sat an intelligence test anywhere. Raven's data from 1938 to 2008 show British children gaining about 14 points over 65 years and then stalling among teenagers. PISA 2025 put England's fifteen year olds at 516 in science, 497 in reading and 492 in mathematics against OECD averages of 482, 461 and 463, and the 2023 Survey of Adult Skills put adults in England above the OECD average in literacy and numeracy.

What no British dataset does is report the population's IQ on an instrument normed for the population today. The Wechsler scale that British psychologists use most, the WAIS-IV UK, carries American norms after a British validation study, according to Pearson's own page, and its 2026 successor has only just been published. The sections below trace each number to its samples, explain what the regional and city rankings that circulate actually measure, and state what the evidence supports about Britain, which is a good deal more precise than the round figure.

1 Why the Figure Is 100 by Construction

An IQ is a position within a norm sample, and the norm samples for the tests that dominate the international tables are British, so a British mean of 100 is a statement about where the scale was anchored rather than a measurement of anything. Raven's Progressive Matrices, the test used in the great majority of the samples in every national IQ dataset, was standardized in Britain in 1938, in a national sample of 1979, in a further sample of 1992 and again in 2007, and John Raven's own account of that history in Cognitive Psychology explains how each standardization was drawn and how far scores rose between them. When a sample from another country is converted to an IQ, its raw scores are placed against one of those British tables. When a British sample is converted, it is placed against a table drawn from British children of a nearby year, and it lands, unsurprisingly, near the middle.

Lynn and Vanhanen made the anchoring explicit. Their 2002 table set the United Kingdom at 100 and expressed every other country relative to it, which is why the older national IQ literature reads as a ranking of distances from Britain. Becker's dataset inherits the same structure in a subtler form. Its conversion of school assessment results into the IQ metric gives the United Kingdom exactly 100.00, which identifies Britain as the anchor of that conversion too, and its sample based estimate of 99.32 is the average of British samples scored on British norms. The page on how IQ scores are normed sets out the general principle: a norm sample defines the scale, the scale does not measure the norm sample, and a country whose samples are the norm samples will score 100 for the same reason that the reference kilogram weighs a kilogram.

The consequence is that the British figure is uninformative about Britain in exactly the way the Indian figure examined on the page about average IQ in India is uninformative about India, but for the opposite reason. India's number is low because unrepresentative Indian samples were placed against British norms of another decade. Britain's number is 100 because British samples were placed against themselves. Neither tells a reader how a representative sample of adults in either country would perform on an instrument normed for them today, because no such measurement exists for either country. What exists for Britain, and is unusually good, is a set of surveys of skills and one pair of surveys of a whole cohort, which the later sections describe.

2 The Six Samples Behind 99.66

Becker's British figure rests on six rows, and four of them are the standardization samples of the Raven's tests themselves, which is why the rows cluster around 100 and why the two that do not are informative about method rather than about Britons. The national table lists 5,794 people across the six samples, with a mean age of about 15 and a quality rating that is among the higher in the dataset. The largest row is the 1979 British standardization of the Standard Progressive Matrices, 3,176 children aged 5 to 15, which scores 99.84 against its own norms, 0.16 from 100 for reasons of rounding in the conversion. The 1992 Advanced Progressive Matrices sample of 603 adults aged 18 to 70 scores 99.39. The 2007 Coloured Progressive Matrices standardization of 846 children aged 4 to 12 scores 101.72. A 1982 Coloured Progressive Matrices sample of 598 children, scored against the 2007 norms, comes in at 90.87 before correction and 96.12 after, the gap being the Flynn effect between the year of testing and the year of the norm. A 1990 sample of 239 nine and ten year olds from a reaction time study by Lynn and Holmshaw scores 99.62 after correction. And a 2014 comparison of children in Britain and Oman by Al Said, 332 children aged 8 to 10 on the Coloured form, scores 92.39.

Weighting the six rows by their listed sample sizes, our arithmetic on the dataset file gives a mean of 99.25 after correction, a fraction below Becker's quality weighted 99.32, and the three standardization samples that are scored against their own norms average 100.1 by the same arithmetic. The only rows that depart from 100 by more than a point are the two that were scored against norms from a different year or drawn for a different purpose, and both departures are in the direction the Flynn effect predicts. That is the whole content of the British figure: a scale centred on British samples returns British samples to its centre, and the noise around the centre is norm drift.

The comparison countries in the same table put the figure in context. Japan stands at 106.39, Singapore at 106.17, Hong Kong at 105.35, China at 102.84 and South Korea at 102.64; Estonia at 101.11, the Netherlands at 100.76, Finland at 100.66 and Germany at 100.61; Canada at 99.62, Australia at 99.20, the United States at 97.50 and Ireland at 95.21, by our reading of the same file. Ireland's value rests on two samples of 391 children with a mean age of seven and a half, which is a reminder that the precision implied by two decimal places is not present in the rows. The page on the standard deviation of 15 explains why a difference of two or three points between countries is smaller than the difference between two administrations of the same test to the same person.

Comparing Britain with Germany's reported average IQ requires checking the reference scale and edition instead of comparing two country labels alone.

3 The Instrument British Psychologists Use, and Whose Norms It Carries

The Wechsler scale that clinicians in Britain have used since 2010 reports a British adult's IQ against American norms, which means that even the individual British score most people would call official is a position within a population of another country. Pearson's product page for the WAIS-IV UK, read on September 19, 2026, gives a publication date of 2010, an age range of 16 years to 90 years 11 months and a qualification level of C. Its FAQ answers the norming question directly: data were collected in the United Kingdom between September 2008 and April 2009 to see whether the British data validated against the American norms; confirmatory factor analysis showed that the subtests and indexes replicated the structure found in the American standardization; and the validation study was taken as sufficient evidence that the British data closely reflected the American, allowing the test to be used in Britain. In plain terms, the British edition adapted the materials and confirmed the structure, and it kept the American norm tables.

The contrast with Australian and New Zealand norms shows why an edition's language adaptation and its norm reference must be checked separately.

The page for the new WAIS-5 UK, read the same day, gives a publication date of 2026, the same age range, a completion time of 45 minutes for the seven subtest Full Scale IQ or 60 minutes for the ten primary index subtests, and describes updated norms, without stating on the page whether the norms are British or American. The page on what the WAIS-5 is sets out the structure of the American edition, and the page comparing the WAIS-4 and WAIS-5 explains what changed between editions. A British reader who receives a Full Scale IQ from a clinician should ask which edition and which norm tables produced it, because the answer determines what the 100 in the report refers to.

None of this makes the clinical score less useful for its purpose. A Wechsler score is used to answer a referral question about one person, and for that purpose a large, well drawn norm sample from a similar population serves well; Pearson's validation study is the evidence that the American tables serve British adults adequately. What it does mean is that the sentence "the average IQ in the UK is 100" cannot be supported by pointing to Wechsler results, because Wechsler results in Britain are not scaled to Britain. The page on reliability and validity explains what a norm sample has to support for each use of a score, and the national average is not among the uses any of these instruments was designed for.

For another national norming decision, see average IQ estimates in Canada, including the effect of scoring the same participants against a different reference.

4 The Only Time a Whole Nation Was Tested: Scotland, 1932 and 1947

Britain holds the two best national cohort measurements of intelligence ever made, and they are the standard against which every convenience sample in every international table should be read. On June 1, 1932, the Scottish Council for Research in Education tested 87,498 children born in 1921, almost every eleven year old attending school in Scotland, on the Moray House Test No. 12, a group test of verbal and numerical reasoning. On June 4, 1947, it tested 70,805 children born in 1936 on the same test. Deary and colleagues' account in the Journal of Personality and Social Psychology of the two surveys and of the follow up studies that grew from them describes both the design and what they later made possible. The 1947 survey had been commissioned partly to test the expectation that national intelligence was falling because larger families had lower scores; it found the opposite, a rise of a little over two points on a 76 point test between the two cohorts, which in retrospect was an early sighting of the Flynn effect.

The surveys are important to the question on this page for two reasons. The first is that they are what a national figure actually looks like: a defined birth cohort, a single instrument, a single day, and coverage close to complete. Nothing in Becker's British rows or in any online table approaches that standard, and nothing in the rows for any other country does either. The second is that the scores were never expressed as an IQ against a foreign norm. They were raw scores on the Moray House Test, later converted to an IQ scale within the Scottish cohort itself, which is the only conversion that a national sample can support without importing another population's distribution.

The follow up work, much of it by Deary and colleagues on the survivors of the 1921 and 1936 cohorts, showed that a score at age eleven predicted survival, health and cognitive ageing decades later, and the page on IQ and longevity reviews that literature. What the surveys could not show, and never claimed, was a Scottish or British mean relative to anyone else, because the instrument had no norms outside Scotland. A reader who wants to know what Britons scored on a fair national test has the answer in those two surveys, and the answer is a distribution on a Scottish scale, not a number on a ranking.

5 The Flynn Effect in Britain: 14 Points, Then a Stall

British Raven's data are the longest continuous record of rising test scores in any country, and they show both the size of the rise and its end among teenagers, which together explain why a British score depends on the year of the norm as much as on the person. Flynn's analysis of Raven's gains in Britain from 1938 to 2008, published in Economics and Human Biology in 2009, found that over about 65 years British schoolchildren aged 5 to 15 taken together gained 14 IQ points, a rate of 0.216 points per year, and that since 1979 the gains had declined with age, so that between ages 12 to 13 and 14 to 15 small gains turned into small losses. Shayer and Ginsburg's study of Piagetian tests of formal operations in the British Journal of Educational Psychology found a large decline among 13 and 14 year olds between norms collected in 1976 and 2006 to 2007, on a different kind of test, which Flynn cited as consistent with the Raven's picture.

The general size of the effect elsewhere is well established. Trahan and colleagues' meta-analysis of 285 studies put the average gain at 2.31 points per decade, and Pietschnig and Voracek's century of global IQ gains across 271 samples and 31 countries put full scale gains at 0.28 points per year and fluid gains, which is what Raven's measures, at 0.41 per year, with a slowing in recent decades. Britain is one of the countries in which the slowing is best documented, and the British case is the reason a careful reader asks the year of the norm before reading any score.

The consequence for the average IQ in the UK is arithmetical. A British adult scored on the 1979 Standard Progressive Matrices norms in 2026 would, if the historical rate had continued, score several points above 100, because the norm group of 1979 performed less well than the population of today; scored on the 2007 norms, the same adult would sit closer to 100; scored on norms drawn today, at exactly 100 by construction. The page on the Flynn effect works through the correction, and the page on matrix reasoning explains why the fluid tests on which the gains were largest are also the ones on which the stall among British teenagers is most visible. There is no sense in which the national mean rose from 100 to 114 over the century, because the mean is fixed at 100 by each new norm; what rose was the raw performance, and what a single figure hides is that the ruler was redrawn every time.

6 The Self-Selected Online Tables and the Regional Rankings

Every table of the average IQ by UK region, city or nation of the union comes from people who chose to sit an online test, and the differences between London and Leeds in such a table are differences between the takers in each place, not between the places. The International IQ Test country table, read on September 19, 2026 through the ACIS country page, shows the United Kingdom at 101.57 from 7,310 takers, within a point and a half of the United States at 101.04 and of India at 98.44, and well below China at 106.48 and Japan at 106.30, whose takers number in the tens and hundreds of thousands. The ACIS page uses that table because it is one dataset with one consistent selection, and it prints the warning beside the tool: the dataset mean is 100 by the site's own choice, and sample sizes must be inspected before any conclusion is drawn.

The regional and city rankings that circulate in British newspapers each year come from the same kind of source, usually a single test provider's own takers broken down by postcode or self-reported location. No instrument has regional British norms, no probability sample of any region has been tested, and the takers in each region are self-selected by interest, device, time and confidence. The page on free versus validated IQ tests explains why a self-selected mean is not a norm, and the review of the International IQ Test examines that particular site on its own terms. A region whose takers score three points above another region's takers has told a reader something about who takes tests in each place and nothing about the residents.

The right use of the online tables is narrow and real. They show that British takers of a given online test score about where American and Indian takers of the same test score, which is a fact about the selection process operating similarly in each country, and they show that within one dataset the British sample is large enough that its mean is stable. What they cannot support is the sentence most often built on them, that one part of Britain is cleverer than another. The page on whether online IQ tests are accurate treats the general problem, and the page on average IQ by state shows how the same selection problem plays out in the American tables.

7 What Representative British Data Actually Measure

Britain has representative measurements of cognitive skills at fifteen and in adulthood, and they are more informative than any IQ table because they say who was tested, how they were sampled and what the score means. PISA 2025, whose national report for England the Department for Education published in September 2026, tested 5,619 fifteen year olds from 183 schools between February and May 2025. England scored 516 in science against an OECD average of 482, with seven countries significantly above it; 497 in reading against 461, with three above it; and 492 in mathematics against 463, with nine above it. The science score was significantly above the 503 of 2022, the reading score was unchanged across cycles, and the mathematics score was significantly below the 504 of 2018. In all three subjects pupils in England scored significantly above those in Northern Ireland, Scotland and Wales. The report is careful about its own sample, noting that non-responding pupils within schools were slightly more likely to speak English as a first language, to have been eligible for free school meals and to receive special educational needs support, and that trends and international comparisons should be read cautiously.

The 2023 Survey of Adult Skills, the OECD's PIAAC, measured adults aged 16 to 65 and is summarized in the OECD country note for England: a mean of 272 in literacy against an OECD average of 260, 268 in numeracy against 263, and 259 in adaptive problem solving against 251, with 18 percent of adults at or below the lowest literacy level against 26 percent across the OECD. Those are the closest things Britain has to a national adult cognitive measurement, and they are on scales of their own, defined across the participating countries rather than against a British norm.

At the individual level, the link between these skills and what intelligence tests measure is close. Deary, Strand, Smith and Fernandes' study of intelligence and educational achievement followed more than 70,000 English pupils from a cognitive ability test at age eleven to their GCSE results at sixteen and found a latent correlation of 0.81 between general ability and achievement. Rindermann's analysis of international comparisons found national means on student assessments and on intelligence tests correlating at about 0.85, which is why Becker folds school assessments into his figure and why a PISA score is, in effect, the second half of every national IQ. The page on IQ and academic achievement explains the individual level relationship, and the page on average IQ by education shows the gradient within a single country.

8 UK Biobank: Half a Million Adults and No IQ

The largest cognitive dataset ever collected in Britain contains no IQ scores, and the reason is instructive: a well designed study measures a construct on a scale it can defend rather than borrowing a scale it cannot. UK Biobank recruited about half a million people aged 40 to 69 between 2006 and 2010 and gave a subset of them, at baseline, a brief battery that included a two minute, thirteen item verbal and numerical reasoning test that the resource labels fluid intelligence. Fawns-Ritchie and Deary's study of the reliability and validity of the UK Biobank cognitive tests in PLOS ONE reported test retest reliabilities and correlations with a reference battery for each of the tests, and concluded that the brief measures were adequate for large scale research despite their length. The scores are reported as the number of items correct. Nobody has converted them to an IQ, because the participants are volunteers who are healthier and better educated than the population, and an IQ scale drawn on them would describe them alone.

That decision is the one every national IQ table fails to make. Becker's rows, the online tables and the regional rankings all convert a convenience sample's scores to a scale drawn on another sample and then report the result as if it described a population. UK Biobank, with a sample a hundred times larger than the whole British contribution to Becker's dataset, declines to do so, and its investigators publish the selection of the cohort as a limitation. The page on what IQ measures explains why a score is a position within a defined reference group and nothing more, and the Biobank test is a clean example of a cognitive measurement that is useful precisely because it makes no claim beyond that.

The British birth cohorts of 1946, 1958 and 1970 followed the same principle a generation earlier. Each administered ability tests in childhood to a national sample and reported the results on the scale of the test used, so that within cohort comparisons are exact and between cohort comparisons require care. The page on whether IQ changes with age draws on that kind of longitudinal data, and the page on average IQ by age explains why age norms exist at all. None of these studies produced an average IQ for Britain, and all of them produced something more useful.

9 Within the UK: Nation, Class and Schooling

The differences that are documented within Britain are differences by nation of the union, by socioeconomic status and by schooling, and each is measured on a stated scale by a representative survey, which is more than can be said for any regional IQ table. PISA 2025 found pupils in England significantly above those in Northern Ireland, Scotland and Wales in all three subjects, with a stable science trend in each nation since 2015. Within England, the gap in science between the most and least disadvantaged quarters of pupils was 81 score points, not significantly different from the OECD gap of 85; in reading it was 66 points, significantly narrower than the OECD's 78; and in mathematics it was 89 points against 83. Pupils eligible for free school meals scored significantly below those who were not in every subject. Girls scored 13 points below boys in science, a gap that did not exist in 2015, 2018 or 2022, and 26 points below in mathematics, a gap that has widened from 12 points since 2015, while girls led in reading by 11 points, down from 23.

Those gaps are large on the PISA scale, where the OECD has in past reports described a difference of roughly 30 to 40 points as comparable to a year of schooling, and they are the kind of thing an average hides. They are also the kind of thing schooling moves. Ritchie and Tucker-Drob's meta-analysis of the effect of education on intelligence, across 142 effect sizes and more than 600,000 people, found gains of about one to five IQ points per additional year of education, persisting across the lifespan, and the British school leaving age reforms of 1947 and 1972 are the kind of natural experiment that literature draws on. A country whose leaving age rose from 14 to 15 to 16 within living memory has moved its own mean by that mechanism, and the Flynn effect data above are partly the record of it.

What the within country evidence does not contain is any IQ difference by region. The Scottish surveys measured Scotland alone; the English cohort studies measured England; PISA reports the four nations separately on its own scale. A reader who wants to know whether the north or the south of England has the higher average IQ is asking a question for which no representative measurement exists, and the tables that answer it are answering a different question about who sits online tests. The page on IQ and income reviews the individual level relationship that regional tables are usually trying to gesture at, and the page on IQ and success sets out what the correlations do and do not support.

10 The Evidence by Design

Read by design rather than by headline, the British evidence sorts into standardization samples that define the scale, two national cohort surveys that predate it, representative surveys of skills on scales of their own, and self-selected online tables, and the third column of the table below is the one to keep.

The pattern is the mirror image of the Indian one. Where India's intelligence samples cannot speak for India, Britain's define the scale and therefore cannot say anything about Britain except that it is the centre. Where India's representative surveys measure deprivation, Britain's measure skills and place England above the OECD average on each. And in both countries the only honest population statement is on a scale the survey itself defined.

11 What the Evidence Supports, Stated Narrowly

Stated as narrowly as the designs allow, the evidence supports five sentences about the average IQ in the UK, and the first of them is the one every ranking page omits. First, the British mean on the scales used in the international tables is 100 by construction, because the British standardization samples of Raven's matrices are the reference for those scales, and the departures from 100 in Becker's rows are norm drift rather than measurement. Second, Britain holds two near complete national cohort measurements, the Scottish Mental Surveys of 1932 and 1947, which showed a rise of a little over two points between cohorts on a Scottish scale and which no later dataset in any country has matched for coverage. Third, British raw performance on Raven's matrices rose by about 14 points among children between about 1942 and 2008, with the gains stalling and turning to small losses among 14 and 15 year olds after 1979, and a British score therefore depends on the year of the norm table it is read against. Fourth, representative surveys place England's fifteen year olds and adults above the OECD average in science, reading, mathematics, literacy and numeracy, with large and measured gaps by socioeconomic status and, in 2025, by gender, on scales those surveys define. Fifth, the online tables and the regional rankings built on them describe self-selected takers, and no regional British norm or probability sample exists.

What the evidence does not support is the reading of 100, or of 99.66, or of 101.57, as a measurement of British ability relative to anyone else, because each number is either the anchor of its own scale or the mean of a self-selected group. It does not support the claim that the British average is falling, which rests on the teenage Raven's stall and the Piagetian decline and not on any adult measurement, nor the claim that it is rising, which rests on the Flynn effect that the same data show slowing. It does not support any ranking of British regions or cities. And it does not support the inference from any of these figures to a person, which the next section takes up.

The narrow statement is more useful than the round number because it tells a reader where to look. For the population, the Scottish surveys, PISA and the Survey of Adult Skills are the evidence, and they are on their own scales. For a person, the evidence is a score they sat, on a stated scale, with a stated error. The page on average IQ explains what 100 means for the population that defines it, and the page on what a good IQ is explains why the individual question has a different answer.

12 What This Does Not Say About You

A British reader is a member of the population that defines the scale, and that membership carries exactly no information about where within the distribution they sit. On a scale with a mean of 100 and a standard deviation of 15, two thirds of the British population lies between 85 and 115 and about 95 percent between 70 and 130, and a national mean of 100 says only that the reader is somewhere in that distribution, which they knew already. A score of 130 sits at the 98th percentile of the reference population and is the threshold British Mensa sets for membership; on the Cattell scale that a common Mensa supervised test in Britain reports, the same standing is written as 148, because that scale uses a standard deviation of 24, as the page on the Mensa IQ test explains. The page on the IQ score chart shows the whole distribution, and the percentile calculator converts any score on any stated scale.

A national summary cannot supply a personal result; the UK IQ testing guide explains the routes available for an individual assessment.

The direction of inference matters as much as the arithmetic. A regional table built from online takers does not describe a reader who did not take that test. A Scottish cohort mean from 1947 does not describe anyone born after 1936. A PISA score describes fifteen year olds in a sampling frame, and a PIAAC score describes adults in a household survey, and neither is a number any individual holds. The page on score versus percentile explains why the same standing looks different on different scales, and the page on fluid versus crystallized intelligence explains why a single matrices mean, which is what most of the British rows are, touches one domain among several.

For a British adult who wants a number that describes them, the position this page has applied to every dataset applies again: ask whose norms, whose sample and what the standard error is. A clinical administration of the WAIS-IV UK reports a position within American norms after a British validation; the WAIS-5 UK reports a position within whatever norms its manual states. An ACIS report scores a person against an English speaking adult reference frame of 3,243 records aged 16 to 90, described in the technical manual, and prints the standard error of every index; that frame is not a British national sample and ACIS does not claim it is. Any of those is a measurement of the person. The national figure, from any source, is not, and a British reader has the unusual privilege of knowing exactly why their country's number is 100.

13 Sources Behind This Page

Every figure above is traceable to one of the following, and each is linked at the point where it is used. Dataset values for the United Kingdom and the comparison countries were read from the NIQ dataset file downloaded from View on IQ on September 19, 2026, and the calculations described as ours were performed on that file. Product details were read on Pearson's British pages on the same day and will change.

• Becker D. The NIQ dataset, version 1.3.5, released October 25, 2023. viewoniq.org, read September 19, 2026.
• Lynn R and Vanhanen T. IQ and the Wealth of Nations. Praeger, 2002. doi.org.
• International IQ Test. IQ by country table, as used by the ACIS country ranking page. international-iq-test.com, read September 19, 2026.
• Raven J. The Raven's Progressive Matrices: Change and stability over culture and time. Cognitive Psychology, 2000, volume 41, issue 1, pages 1 to 48.
• Pearson Clinical Assessment UK. Wechsler Adult Intelligence Scale, Fourth UK Edition (WAIS-IV UK), product page and FAQ on the validation study and norms. pearsonclinical.co.uk, read September 19, 2026.
• Pearson Clinical Assessment UK. Wechsler Adult Intelligence Scale, Fifth UK Edition (WAIS-5 UK), product page. pearsonclinical.co.uk, read September 19, 2026.
• Deary I J, Whiteman M C, Starr J M, Whalley L J and Fox H C. The impact of childhood intelligence on later life: Following up the Scottish Mental Surveys of 1932 and 1947. Journal of Personality and Social Psychology, 2004, volume 86, issue 1, pages 130 to 147.
• Flynn J R. Requiem for nutrition as the cause of IQ gains: Raven's gains in Britain 1938 to 2008. Economics and Human Biology, 2009, volume 7, issue 1, pages 18 to 27.
• Shayer M and Ginsburg D. Thirty years on, a large anti-Flynn effect? (II): 13 and 14 year olds. Piagetian tests of formal operations norms 1976 to 2006/7. British Journal of Educational Psychology, 2009, volume 79, issue 3, pages 409 to 418.
• Oxford University Centre for Educational Assessment and UCL Institute of Education for the Department for Education. PISA 2025: national report for England, volume 1. gov.uk, September 2026.
• OECD. Survey of Adult Skills 2023: country note for England (United Kingdom). oecd.org, December 2024.
• Deary I J, Strand S, Smith P and Fernandes C. Intelligence and educational achievement. Intelligence, 2007, volume 35, issue 1, pages 13 to 21.
• Fawns-Ritchie C and Deary I J. Reliability and validity of the UK Biobank cognitive tests. PLOS ONE, 2020, volume 15, issue 4, e0231627.
• Ritchie S J and Tucker-Drob E M. How much does education improve intelligence? A meta-analysis. Psychological Science, 2018, volume 29, issue 8, pages 1358 to 1369.
• Pietschnig J and Voracek M. One century of global IQ gains: A formal meta-analysis of the Flynn effect (1909 to 2013). Perspectives on Psychological Science, 2015, volume 10, issue 3, pages 282 to 306.
• Rindermann H. The g-factor of international cognitive ability comparisons: the homogeneity of results in PISA, TIMSS, PIRLS and IQ tests across nations. European Journal of Personality, 2007, volume 21, issue 5, pages 667 to 706.

14 Frequently Asked Questions