What Is the Average IQ in the USA?

What is the average IQ in the USA?

On any properly normed American IQ test — the Wechsler Adult Intelligence Scale or the Stanford–Binet — the average score is 100 by construction.

Here is the part most articles skip. Test publishers run a process called norming. They administer the test to a large, demographically representative sample of Americans, look at the raw scores, and then build a scoring scale that forces that sample's average to equal 100. The spread is set so that one standard deviation equals 15 points.

This has a consequence that surprises people. The average American IQ cannot rise or fall. If Americans collectively became sharper or duller tomorrow, the test would be re-normed and the average would return to exactly 100. The number 100 tells you nothing about how intelligent Americans are. It tells you where the middle of the American distribution sits — because that is where it was placed.

What can change is raw performance: how many questions people actually answer correctly. That distinction is the key to everything else on this page. An IQ of 100 is the 50th percentile — you scored higher than half of Americans. To convert any score, see our IQ percentile calculator, or read what counts as a good IQ score.

Why do some websites say 100 while others say 97.43?

Both figures are real. They come from different scales with different reference points. Call it the anchor problem: an IQ score is only meaningful relative to the population the test was calibrated against. Change the anchor, change the number — without anyone's intelligence changing at all.

The 100 figure is anchored to Americans

US-normed tests set the American sample's average to 100. The reference population is the United States. By definition, the US average is 100.

The 97.43 figure is anchored to the British

That number comes from a cross-national dataset — Richard Lynn and David Becker's The Intelligence of Nations (2019). It does not norm each country against itself. It places all countries on a single shared scale anchored to the United Kingdom, whose mean is fixed at 100 with a standard deviation of 15. Lynn named this the “Greenwich IQ” metric, after the London suburb through which the prime meridian runs.

So “the US average IQ is 97.43” means, precisely: on a scale where Britain is defined as 100, the United States estimates to 97.43. It is a comparative statistic, not a score any American ever received on a test.

The 98 figure is usually unsourced

The widely repeated “98” typically appears with no citation at all. It is generally a rounded or earlier-vintage national-IQ estimate. If a page gives you 98 without naming its source, treat it as unverified.

How reliable is the cross-national data?

The national-IQ dataset behind the 97.43 figure is contested within the scientific literature on methodological grounds. Sample selection was never systematically documented, and critics describe the method as unreplicable. Many national estimates rest on small convenience samples. Some estimates are unstable in ways no population-level trait permits — one country's national IQ reportedly fell more than 15 points across seven years. Several papers built on the dataset have since been retracted.

What is the average IQ by US state?

No US state has ever been IQ-tested. There is no study in which a representative sample of Massachusetts residents sat the WAIS. Every state “IQ” figure in circulation is an estimate derived from achievement tests, converted onto an IQ-like scale.

The lineage is short and traceable. McDaniel (2006) estimated state IQ from NAEP fourth- and eighth-grade reading and mathematics scores. Pesta (2022) updated those estimates using NAEP together with PIAAC adult data, because McDaniel's figures were by then more than sixteen years old. Pesta is explicit that these are statistical constructs derived from educational assessment data, not intelligence testing.

You may see a different ranking elsewhere, with Massachusetts first and Mississippi last. Those are an older vintage of the same estimation method, closer to McDaniel's original NAEP-only figures. The two tables are not contradictory data. They are the same method, run on different assessment years — and most sites publishing them do not say which.

Treat state IQ figures as a rough proxy for educational achievement, heavily shaped by school funding, poverty and test participation. They are not a measure of anyone's intelligence. One caveat on arithmetic: averaging these fifty numbers does not give you the US average, because states differ enormously in population.

Does the average IQ in America change with age?

No — the average IQ is 100 at every age. IQ tests are age-normed: your score compares you to other people your own age, not to the population at large. A 70-year-old scoring 100 and a 20-year-old scoring 100 have each landed at the middle of their own age band. By construction, the average never moves.

So “does IQ decline with age?” is not really an IQ question. It is a question about raw cognitive performance, and there the answer is more interesting. Fluid abilities — reasoning about novel problems, processing speed, working memory — tend to peak relatively early in adulthood and decline gradually. Crystallized abilities — vocabulary, accumulated knowledge, verbal comprehension — keep improving well into middle age and often beyond.

Different cognitive abilities peak at different times of life, which is precisely why a single IQ number conceals more than it reveals. Our guide to working memory covers one of those abilities in depth.

Is there a difference in average IQ between American men and women?

No meaningful difference in average general intelligence has been established. Some pages report that American men average 99 and women 97. That claim is not supported by the research literature and you should disregard it.

This holds both for standardised IQ tests — which are deliberately constructed with balanced items so as not to favour either sex — and for cognitive assessments that were not built to that criterion. Male and female distributions overlap almost entirely.

What the evidence does support is a difference in variability, not average. Johnson, Carothers and Deary (2008) examined the greater male variability hypothesis and found men's scores more spread out, with male standard deviations running roughly 5–15% larger than female.

This finding is not unanimous; some studies find equal variability, and some find greater female representation at the upper end of particular abilities. What no credible body of evidence supports is a two-point average gap between American men and women.

How does the USA compare with other countries?

If national IQ data is unreliable, how can you compare countries at all? Use assessments built for exactly that purpose. The OECD's Survey of Adult Skills (PIAAC) tests representative samples of adults aged 16–65 across participating countries using a common instrument and a transparent, published methodology. It is not an IQ test — it measures literacy, numeracy and problem-solving — but it is a genuine, current, methodologically defensible international comparison, which is more than the national-IQ tables can claim.

The distribution matters as much as the average. 34% of US adults scored at or below Level 1 in numeracy — the lowest proficiency band — against an OECD average of 25%. In literacy, 28% scored at or below Level 1, against an OECD average of 26%. Between the 2017 and 2023 rounds, average US scores fell in both literacy and numeracy.

This is worth holding alongside the Flynn effect — the long twentieth-century rise in raw IQ test performance of roughly three points per decade, which re-norming continually absorbed. Recent skills data from several wealthy countries has prompted debate about whether that rise has stalled or reversed. The honest answer is that the evidence is mixed and still being argued.

Common myths about the average American IQ