The Flynn Effect Explained

The Flynn effect is the large, long-running rise in IQ test scores recorded across much of the world during the 20th century, at roughly three points per decade in the United States. Each new generation scored well above 100 on tests that had been normed on older generations. Most researchers think this reflects environmental change, such as schooling and more abstract habits of thought, rather than a jump in raw intelligence.

What is the Flynn effect?

To understand the Flynn effect, start with how IQ tests are built. When a test is first standardised, a sample of people takes it and the average result is fixed at 100, with a standard deviation of 15 (sometimes 16) points. That means a score of 100 simply says you performed at the middle of the reference group, which is why average IQ is always 100 by design.

Tests are revised from time to time and re-standardised on a new sample, usually people born more recently. The new average is again set to 100. The striking finding is what happens when the newer group sits the older version of the test: in almost every case their average lands clearly above 100. In other words, the same raw performance that earned an average score in one era earns a below-average score a generation later.

That steady upward drift in raw performance is the Flynn effect. Researchers also use the more neutral phrase secular rise in IQ scores, which names no individual.

Who discovered it and how did it get its name?

The effect is named after James Robert Flynn (1934–2020), an American-born moral philosopher and intelligence researcher who emigrated to New Zealand in 1963 and taught political studies at the University of Otago. In 1978, while writing a rebuttal of claims that group differences in IQ were genetic, he studied historical test data and noticed that raw scores kept climbing even though tests were always recalibrated to an average of 100.

He published Race, IQ and Jensen in 1980, then a 1984 paper reporting that Americans' average scores on Stanford-Binet and Wechsler tests had risen by 13.8 points between 1932 and 1978. In 1987 he followed with a study of 14 nations showing gains of between five and 25 points.

Flynn did not coin the label himself. Richard Herrnstein and Charles Murray called it the Flynn effect in their 1994 book The Bell Curve, and textbooks followed. Flynn was also clear that he was not first. Earlier writers, including Read D. Tuddenham in 1948, Daniel C. Calhoun in 1973, Robert L. Thorndike in 1975 and Richard Lynn, who recorded a rise in Japan in 1982, had noted rising scores. Flynn said that if asked, he would have named it after Tuddenham, whom he credited with the first convincing nationwide evidence.

How large were the gains?

The numbers are big enough to sound implausible at first, which is part of why the finding caused such a stir.

• About three points per decade in the United States, as measured by Wechsler tests. Flynn's 1984 figure of 13.8 points over 46 years works out to roughly 0.3 points a year.
• A 2014 meta-analysis by Trahan and colleagues found an average of 2.31 points per decade across 285 studies since 1951, and 2.93 points per decade for modern Stanford-Binet and Wechsler tests normed since 1972.
• British children's average scores on Raven's Progressive Matrices rose by 14 points between 1942 and 2008.
• Dutch military conscripts gained 21 points in the 30 years from 1952 to 1982, about seven points per decade.
• A study of Spanish children tested 30 years apart found a rise of 9.7 points.

Projected backwards, the results look odd. Ulric Neisser estimated that, scored against 1997 norms, the 1932 American average on the first Stanford-Binet test standardisation sample would be about 80. Flynn calculated that the average American of 1900 would score around 67 on a test of 2000. Nobody seriously believes most people a century ago were intellectually impaired, and that tension sits at the heart of the debate about what the effect means.

Gains appeared across Western Europe, Japan, South Korea and many other places where testing was common, though not always at the American rate. Similar rises have been reported for attention and for semantic and episodic memory.

Which abilities rose and which did not?

If longer schooling simply made people better at school subjects, you would expect the biggest gains in vocabulary, arithmetic and general knowledge. The opposite happened. Those areas showed relatively small gains, and occasionally declines.

The largest rises came on tasks involving classification, analogies and abstract reasoning. On the Wechsler children's scale, the similarities subtest, which asks how two things are alike, rose sharply while arithmetic barely moved. Non-verbal pattern tests such as Raven's matrices, designed to measure on-the-spot problem solving with little reliance on learned content, also showed large gains. That is why Flynn turned to them when critics argued education alone explained the rise.

A 2015 meta-analysis by Pietschnig and Voracek, covering nearly 4 million participants, put annual gains at:

• 0.41 points for fluid IQ
• 0.30 points for spatial IQ
• 0.28 points for full-scale IQ
• 0.21 points for crystallised IQ

They also found the effect was stronger in adults than in children. Several studies, including the Spanish one and work by Teasdale and Owen in 1989, found that the gains were concentrated in the lower half of the distribution, mostly reducing the number of low scores rather than adding very high ones.

What causes the Flynn effect?

There is no single agreed cause. The rise was far too fast to be genetic, so the explanations on the table are environmental. The main candidates are these.

• Schooling. People stay in education much longer than they used to. A 2018 review led by Stuart Ritchie of nearly 30 studies and about 600,000 people concluded that each extra year of formal schooling raised IQ by between 1 and 5 points. Starker evidence comes from Virginia in the 1960s, where some counties closed public schools; African-American children who received no formal education saw scores fall by about six points a year. The catch is that when older and younger Americans with similar education are compared, the gains barely shrink, so schooling cannot be the whole story.
• Abstract thinking. Flynn's own view was that modern life trains people to take hypotheticals seriously, sort things into abstract categories and work with symbols. Asked what dogs and rabbits have in common, someone in 1900 might say a dog hunts rabbits; someone in the 1990s would more likely say both are mammals. Only the second answer scores well on an IQ test.
• Test familiarity. Retaking the same test typically adds five or six points, which suggests practice matters. But that sets a ceiling on the explanation, and in the US the groups most familiar with testing showed smaller gains.
• Health and nutrition. Better diets, fewer infectious diseases and reduced exposure to hazards such as lead paint are often proposed.
• Literacy and stimulation. David Marks pointed to a close link between rising literacy and IQ gains, and others cite a more complex, demanding everyday environment.

Flynn and William Dickens proposed a model in which causes feed each other: a challenging environment raises IQ, and higher IQ draws people into still more challenging environments.

Are people actually getting smarter?

This is the question the Flynn effect forces, and the honest answer is that it is still argued over. As Neisser put it, test scores are certainly going up, but whether intelligence itself has risen remains controversial.

In 1987 Flynn took a sceptical line, suggesting the tests measured only a narrow kind of abstract problem-solving with limited practical significance. He reasoned that if intelligence had truly surged, we should have seen something like a cultural renaissance, and we had not. By 2012 he had moved away from that view and refined his account of what the rising scores meant.

Today a common position is that the gains reflect a shift in how people think, towards the abstract reasoning that IQ tests reward, plus some test-specific skill, and possibly some real improvement in ability. Studies have found that the gains do not line up fully with the underlying general factor that IQ tests are meant to capture, which is a technical way of saying the scores rose for partly different reasons than individual differences in IQ arise. The safest reading is that IQ scores are sensitive to environment and era, which is itself useful to know when you interpret any score.

Is there a reverse Flynn effect?

Yes, in some places and some groups, though how widespread it is remains disputed. Researchers have reported a mild decline in scores in several developed countries from around the 1990s, sometimes called the reverse or negative Flynn effect. Declines have been identified in Scandinavia and elsewhere.

The evidence points in more than one direction:

• Pietschnig and Voracek's 2015 meta-analysis found the Flynn effect had weakened in recent decades.
• Trahan and colleagues' 2014 meta-analysis found no sign that it was diminishing on modern US tests.
• As of the mid-2010s, meta-analyses suggested the effect was continuing overall, either at the same pace or more slowly in developed countries.
• A 2019 study of over 10,000 US adolescents by Platt and colleagues found no significant overall effect on a fluid reasoning measure, but large differences beneath that average: scores rose 2.3 points per decade at age 13 and fell 1.6 points at age 18, and fell 4.9 points among those scoring 70 or below while rising 3.5 points among those scoring 130 or above.

Some apparent declines may come from cultural change making parts of older tests out of date, rather than from any real loss of ability. Some research also suggests the effect may have ended in a few developed nations while continuing elsewhere, which could narrow gaps between countries over time.

Why does the Flynn effect matter for IQ scores today?

The practical consequence is norm obsolescence. Because raw performance tends to rise, a test scored against old norms will usually flatter you. Trahan and colleagues estimate that a test overestimates IQ by about 0.3 points for each year between when it was normed and when it is taken.

That has real effects. Their example: in 2014, a qualifying score of 132 on an older Stanford-Binet edition would correspond to roughly 126 on the newer edition, because the older norms were about 20 years staler. Bear that in mind when comparing results against something like the Mensa qualifying score, which requires a result in the top 2%.

The stakes are much higher in clinical and legal settings. In the United States, a score around 70 is commonly used as one indicator of intellectual disability, and the Flynn effect is argued over in death penalty cases, special education eligibility and disability benefits. Anyone facing an assessment of that kind should rely on a qualified psychologist using current, properly normed tests, not on an online result.

For everyday purposes, the lesson is to ask what any score is compared against and how recent that comparison group is. Knowing how a test is scored tells you more than the number alone, and it is worth understanding how accurate online IQ tests are before reading much into one. If you are curious where you stand, you can take our free IQ test with those caveats in mind.

Frequently asked questions

Sources

1. Flynn effect (Wikipedia)
2. James Flynn (academic) (Wikipedia)
3. The Flynn Effect: A Meta-analysis - PMC
4. Flynn effect | Education | Research Starters | EBSCOhost
5. The Flynn effect for fluid IQ may not generalize to all ages or ability levels: A population-based study of 10,000 US adolescents | Stress & Development Lab
6. What Is The Flynn Effect In Psychology?