Average IQ by Country: What the Numbers Can and Can't Tell You

There is no trustworthy table of average IQ by country. The lists that circulate online trace back to a few compilations, mainly by Richard Lynn and colleagues, that stitch together old, small and often unrepresentative samples taken with different tests in different decades. The average IQ in the United States is about 100, but that is true by design: American IQ tests are scaled so that a representative sample of Americans averages 100.

That answer disappoints people who wanted a number for their country. The rest of this page explains where the viral numbers come from, why researchers who study them do not trust them, and what better evidence exists about how countries differ.

Where the "IQ by country" lists come from

Almost every map or ranking you will find descends from a small set of books. Richard Lynn and Tatu Vanhanen published IQ and the Wealth of Nations in 2002 and updated the dataset in later books. Lynn and David Becker released The Intelligence of Nations in 2019, which much of the internet now copies, often without saying so.

The method is straightforward to describe. The authors search the literature for any study that gave an intelligence test to people in a given country, convert the results to a common scale anchored on British norms (mean 100, standard deviation 15), apply some adjustments, and average them into a single national figure. Where a country has no usable study, earlier editions estimated a value from neighboring countries.

The compilations are widely cited because they are the only dataset of their kind, not because the field regards them as accurate. Each step in the method introduces error, and the errors stack.

Why national IQ estimates are unreliable

The samples rarely represent the country

A national average is only meaningful if the people tested resemble the nation. Many of the underlying studies were never designed for that. They tested a single school, one town, a group of job applicants, patients at a clinic, or children in a particular program, sometimes with fewer than a hundred participants. A figure for a country of tens of millions can rest on one or two such studies.

Different tests, different decades

The source studies use whatever test the original researchers chose: Raven's Progressive Matrices in one place, a Wechsler scale or a locally built test in another. Some date from the 1950s, others from the 2000s. Converting all of them to one scale assumes the tests measure the same thing in the same way across languages and cultures, an assumption that is rarely checked.

Dates matter because scores on many tests rose by roughly three points per decade through much of the twentieth century, the Flynn effect. Comparing a 1960s sample with a 2000s sample without a careful correction can produce a gap of ten points that says nothing about the two countries.

Missing countries filled in by estimate

For countries with no data, the early compilations assigned a value based on nearby countries. Those imputed figures then appear on maps with the same visual weight as measured ones, and later users rarely distinguish the two.

Selective inclusion

The most detailed critique is a systematic review by Jelte Wicherts, Conor Dolan and Han van der Maas (2010). They examined the literature on one large region that the compilations score very low, applied explicit inclusion rules decided in advance, and arrived at a substantially higher average than Lynn's figures. They documented that the compilations had left out many studies that met reasonable standards while keeping some that did not, and that many retained samples were unrepresentative. A related paper by Wicherts, Borsboom and Dolan (2010) argued that national IQ figures cannot carry the evolutionary arguments built on them.

Lynn and colleagues published replies, and the exchange continues. The mainstream position among psychometricians is that the national figures are too noisy and too selectively assembled to rank countries, and that they should not be used to make claims about groups. Treat any ranking table built from them as contested at best.

What "average IQ in the US" really means

IQ is not a raw quantity like height. It is a position relative to a reference group. When Pearson builds the norms for the WAIS-5 (adults) or the WISC-V (children), it tests a large sample selected to match the US population on age, sex, education, region and race/ethnicity, then sets the scale so that sample averages 100 with a standard deviation of 15. You can see how scores spread around that center in our guide to the bell curve.

So the US average is about 100 by construction. It does not mean Americans are average compared with the world. It means the typical American scores at the middle of a scale built from Americans. A country that norms its own edition of a test will also find its average is about 100 on that edition.

Two caveats apply. First, norms age: because of the Flynn effect, a population tested years after a test was normed tends to score above 100 on the old norms, which is why publishers renorm every decade or two. Second, averages for subgroups or states are not published as official IQ figures. Lists of "average IQ by state" usually come from converting school test scores, not from IQ testing.

Credible cross-country comparisons: PISA and TIMSS

If you want to compare countries with representative data, the best sources are international school assessments. They do not measure IQ, and they do not claim to. They measure what students can do with reading, math and science at a given age or grade. Their strength is method: nationally representative samples, the same test translated and checked across countries, and documented sampling rules.

In PISA 2022, which covered 81 countries and economies, US students scored above the OECD average in reading and science and below it in math. Results like these are useful for asking what school systems do well, but they reflect schooling, language, curriculum and student motivation as much as underlying reasoning ability, so they should not be relabeled as "IQ."

Some later IQ compilations folded converted school-assessment scores into their national estimates. That blurs the line further: a number presented as national IQ may partly be a school test score under another name.

What drives real differences between countries

Average test performance does differ between populations, and researchers can point to causes they are able to demonstrate. Every one of them is environmental, and several, such as schooling and health, have changed within single countries over a few decades.

• Schooling. A meta-analysis by Ritchie and Tucker-Drob (2018) found that each additional year of education raises IQ scores by roughly 1 to 5 points. Countries differ greatly in how many years children spend in school and how good that schooling is. See how education affects IQ.
• Health and nutrition. Iodine deficiency, iron-deficiency anemia, malnutrition in early childhood, lead exposure and parasitic and infectious disease all lower cognitive development, and all are far more common in poorer countries.
• Test familiarity. Timed multiple-choice tests with abstract shapes are routine for children in some countries and completely new in others. Unfamiliar formats depress scores for reasons unrelated to reasoning.
• Language and translation. Verbal items rarely survive translation intact, and even nonverbal tests come with verbal instructions.
• Generational gains. Pietschnig and Voracek's (2015) meta-analysis of the Flynn effect found gains across the world through the twentieth century, with rates that varied between regions and slowed in recent decades. Gains of that size over a few decades cannot be genetic. In some wealthy countries the rise has slowed or reversed, and a Norwegian study of brothers by Bratsberg and Rogeberg (2018) traced the reversal to environmental causes as well.

Heritability estimates are sometimes brought into this debate, but heritability is a within-population statistic. It describes how much of the variation among people in a particular population, in a particular environment, tracks genetic differences. It says nothing about why two populations living in different conditions differ on average. We cover that distinction in nature vs. nurture in intelligence.

IQ testing around the world

Individually administered IQ tests are clinical instruments, and their use differs by country. Major tests such as the Wechsler scales are adapted and separately normed in many countries, including the UK, Canada, Australia, France, Germany, Spain and Japan, each against its own population. A score of 115 on a German edition means "above most Germans," not "above most Americans."

Many countries have no locally normed version of the major tests. Clinicians there often use norms from another country, which can misclassify people, especially children being evaluated for learning or intellectual disability. Building local norms is expensive, and it is one of the practical gaps between rich and poor countries in psychological services.

Online tests, including ours, sit outside this system. Our result describes your position among the people who have taken our test — shown as an estimated IQ with a range — not a national or international IQ, and it says nothing about the country you live in. If you are curious where you stand among our respondents, you can take the free test; the methodology page explains how the percentile is calculated.

How to read a viral IQ map

When you meet a country ranking, a few questions sort out how much weight it deserves:

1. What is the source? If it traces back to Lynn and Vanhanen or Lynn and Becker, the numbers carry the problems described above.
2. How many people were tested, and who? A national figure built on one school sample is not a national figure.
3. When were they tested? Decades-old data on modern norms will be off by several points from the Flynn effect alone.
4. Which values were measured and which were estimated? Imputed values are often unlabeled.
5. Is it really IQ? Many recent maps convert school test scores and call them IQ.

The honest summary is short. Countries differ in schooling, health and test experience, and those differences show up in test scores. A single "IQ" number per country, ranked to the decimal point, claims a precision that the data cannot support.