What Is the Flynn Effect? Why IQ Scores Rose—and Why Some Have Fallen

The Flynn effect is the long-term rise in population performance on intelligence tests when comparable scores are examined across generations. Average gains during much of the twentieth century were commonly estimated at roughly two to three IQ-score points per decade. But the trend is neither universal nor endless: results vary by country, cognitive domain, historical period and testing method.

The key distinction: higher average test scores do not establish that every new generation is smarter in every meaningful sense. Recent studies in Norway, Germany and the United States have found plateaus, reversals or different trends across abilities. Research published in 2026 highlights why education, socioeconomic conditions and test design must be considered before interpreting a generational change.

What is the Flynn effect?

Named after James R. Flynn, who documented major historical score gains in the 1980s, the effect refers to differences between population cohorts, not an automatic increase in an individual’s IQ over time. When a newer generation and an older one are compared on an appropriate common scale, the later generation has often performed better on certain test questions.

The American Psychological Association Dictionary of Psychology describes it as a gradual rise in raw general-intelligence test scores across cultures, with unequal gains across domains. Abstract and fluid-reasoning tasks have often shown stronger increases than some accumulated-knowledge tasks.

How can IQ scores rise if the average IQ is always 100?

Most widely used modern IQ scales are periodically standardized on fresh population samples, with the reference average set around 100 and a standard deviation of about 15. Scores against older norms can therefore appear higher even though a newly standardized test resets its reference mean.

Illustrative example, not empirical data: imagine an adult group averaging 100 under norms established in 1990. If a comparable group in 2020 would average 108 against those older norms, a modern re-standardization can again place the population mean near 100. The eight-point difference in this hypothetical example demonstrates why the same performance must be placed on a consistent historical scale.

The phenomenon should not be reconstructed by casually comparing scores from unrelated online tests, different age bands or instruments with incompatible scoring procedures.

How large were historical intelligence-test gains?

A 2014 meta-analysis covering 285 comparisons and 14,031 participants estimated an average increase of 2.31 IQ points per decade. Another major meta-analysis published in 2015 synthesized 271 independent samples encompassing nearly four million people across 31 countries and found substantial international and domain-specific variation.

Therefore, “three points per decade” is a useful historical shorthand, not a rule that applies to every society, age group, ability or calendar period.

Why did IQ test performance increase?

Expanding education and practice with abstract concepts

Longer schooling, changing curricula and greater familiarity with diagrams, classifications and symbolic problems may improve performance on tasks used in cognitive assessments. A large 2026 Norwegian study strongly illustrates the importance of educational expansion, but it does not identify schooling as a complete explanation for every historical gain.

Health, nutrition and childhood development

Improved childhood nutrition, healthcare and reductions in environmental hazards may support cognitive development. Their relative contributions are likely to differ across locations and decades, and the available evidence does not justify assigning all historical gains to one factor.

A world demanding more symbolic and analytical skills

Modern environments often require people to process text, graphs, menus, categories, interfaces and unfamiliar systems. These experiences could strengthen skills resembling those tested, although it is difficult to isolate their effects from education, economic development and other social changes.

Test versions, participation and measurement

Different generations may take different test versions or enter samples through different routes. Researchers ask whether a scale measures the same underlying construct in the same way across time—known as measurement invariance. Without that check, score differences can sometimes exaggerate what can be inferred about broader ability.

What is the reverse Flynn effect?

The reverse Flynn effect describes a slowdown, plateau or decline in average test scores over time within a particular setting. It does not establish a global decline in human intelligence. Nor must all the cognitive domains examined move in the same direction.

An American study published in 2023 analyzed data from 394,378 adults who participated in online cognitive testing between 2006 and 2018. Some composite and reasoning scores declined over time, while scores on three-dimensional rotation generally improved over a shorter comparison period. Because respondents took part through an online project, its size alone does not make the sample representative of every American adult.

German research published in 2024 found declines in figural-reasoning performance among 19,474 secondary-school students in samples covering 2012–2022. The finding concerns specific tasks and student populations rather than an all-purpose statement about a country’s intelligence.

What do the most important studies actually show?

The July 2026 Norwegian study: why group averages can hide a reversal

A paper published in PNAS on July 6, 2026 linked cognitive tests collected during military conscription to education and household records for 579,379 Norwegian men from 25 birth cohorts spanning 1967–1991.

The researchers reported that declines began earlier among sons of high-income fathers. At the same time, expanding upper-secondary education helped raise performance among lower-income groups, temporarily masking the decline in the overall population average. As educational attainment became more similar across groups, declines became visible more broadly. The research highlights the importance of social stratification and educational opportunities, not a single established explanation for every change.

Do falling test scores prove that general intelligence is falling?

Not necessarily. A 2025 re-analysis of Norwegian military test scores raised questions about whether some observed changes reflected the same underlying latent ability over time, because scalar measurement invariance was not established. Conversely, an August 2026 study of 1,187 children and adolescents aged 7–15 found clear generational gains, but their magnitude and direction relative to test versions differed by subtest and analytic approach.

This distinction separates observed scores from the theoretical construct often labeled general cognitive ability, or g. Researchers continue to investigate how strongly changes in particular test questions track changes in that broader construct.

Are smartphones, social media or the internet causing reverse Flynn effects?

Current results do not establish this causal claim. Changes in screen exposure, reading habits or attention are plausible research questions, but the simultaneous rise of digital media and decline of some scores is not enough to prove an explanation. Careful studies must account for education, economic background, participant selection and changes in the tests themselves.

Does the evidence tell us IQ is rising or falling in Egypt?

The original studies reviewed here do not provide a sufficiently comparable, representative Egyptian time series to establish a recent national Flynn effect or reversal. Internet country rankings and self-selected online IQ quizzes are not substitutes for longitudinally comparable, locally standardized data.

Five checks to make before sharing an IQ trend headline

1. Population: Is the claim about one country, age group, school system or the world?
2. Sample: Was the sample representative, or selected through online volunteering or a particular institution?
3. Measure: Were test versions and scoring norms made comparable over time?
4. Domain: Is the change in a composite score or only vocabulary, matrices, arithmetic or spatial tasks?
5. Cause: Does the study identify a cause, or only report a correlation or historical pattern?

What does the Flynn effect mean for an individual?

The findings demonstrate that cognitive test performance is shaped by historical context and measurement choices. They do not make any person’s IQ score a complete description of their potential, judgment or worth. Good education, adequate sleep, nutrition and healthcare can support learning and wellbeing, but no specific increase in IQ should be promised on that basis alone.

Frequently asked questions

Who was James Flynn?

James R. Flynn was a scholar who documented long-term shifts in intelligence test scores and helped draw attention to their historical and social implications. The effect is named after him.

Did IQ really rise by three points every decade?

That is an approximate historical rule of thumb. A major 2014 meta-analysis estimated 2.31 IQ-score points per decade on average, with considerable differences across tests and settings.

Why does the average IQ still equal 100?

Standardization sets the reference mean for many modern IQ tests near 100. Historical changes emerge when cohorts are assessed on appropriately comparable norms instead of treating each new standardization as an absolute scale across time.

Are younger generations less intelligent now?

There is no scientifically justified universal verdict. Some recent populations and domains show declining scores, while others exhibit improvement or stability. Study design and measurement comparability matter.

Can practising IQ questions prove that general intelligence has improved?

Practice may increase scores on similar tasks without demonstrating comparable improvements across every cognitive domain or in latent general intelligence. Valid interpretation depends on standardized assessment and the purposes of the test.

The bottom line

The Flynn effect is a well-documented historical pattern in cognitive test results, not a guarantee that later generations are uniformly more intelligent. Research through 2026 paints a more complex picture: rising and falling scores coexist, social and educational differences matter, and the relationship between changing test performance and underlying cognitive abilities remains an active research question.

Primary research and authoritative sources

• American Psychological Association Dictionary of Psychology — definition.
• Trahan et al. (2014), Psychological Bulletin — meta-analysis.
• Pietschnig & Voracek (2015) — global IQ trends, 1909–2013.
• Bratsberg & Rogeberg (2018), PNAS — within-family evidence.
• Dworak et al. (2023) — US adult cognitive test trends.
• Reevaluating the Flynn effect (2025) — measurement invariance in Norway.
• Bratsberg et al. (2026), PNAS — educational expansion and hidden reversals.
• Giangrande et al. (2026) — childhood and adolescent test domains.