Average IQ

Wechsler tests, the most widely used adult IQ tests, label scores of 90 to 109 as average. That band holds about 49% of people. Scores of 80 to 89 are low average and 110 to 119 high average, and each holds about 16%. The IQ classifications page lists every band with its percentile.

Other sources use different cut points. Some call 85 to 115 the normal range because it covers the middle 68% of people, one standard deviation either side of the mean. Both are conventions. No score is a clinical threshold on its own.

Publishers do not simply test whoever is available. They build a norm sample designed to mirror the population by age, sex, region, ethnicity and education, usually matched to census figures, and they test many people in each age band. The WAIS-5, for example, was standardised on about 2,000 people aged 16 to 90.

The raw scores of each age band are then converted so that the middle of that band sits at 100. The score table you get with a test is simply a lookup from raw score and age to a standard score. This is why two tests can give different numbers for the same person: they used different samples, in different years, with different spreads.

It is also why an online quiz that gives you "an IQ" without age bands or a norm sample is not reporting a position in a population. It is reporting a number on a scale the author chose. Our guide to how accurate online IQ tests are explains what to look for.

The earliest IQ scores were ratios. A child's mental age, the age at which the average child performs as they do, was divided by their real age and multiplied by 100. A 10-year-old performing like the average 10-year-old scored 100, and a 10-year-old performing like the average 12-year-old scored 120. The average was 100 here too, by construction.

Modern tests dropped the ratio, because mental age stops rising in adulthood and the formula breaks down. They compare each child with others of the same age instead, as adults are compared. See the history in our guide to the Stanford-Binet.

It is 100 at every age. Scores are age-normed: your result is compared with people of your own age, so a 12-year-old and a 60-year-old with the same standing among their peers get the same IQ. Pages that show "average IQ by age" as a falling line from 100 are usually mixing raw scores with IQ.

Raw performance does change with age. Children answer more items correctly each year, and abilities then peak at different times. In a study of nearly 49,000 people on standardised IQ and memory tests, Hartshorne and Germine found that processing speed and short-term memory for new information peak around the end of school. Abstract reasoning plateaus in early adulthood and declines from the 30s. Vocabulary and general knowledge keep rising into the 40s and later. Age norms hide all of that, which is the point of them.

Our IQ by age pages show how the same raw performance reads at different ages. A test that is not age-normed, including most online tests, will not behave this way.

Published national averages exist, and they should be read with care. Each country is normed against itself, so a national average of 100 on a locally normed test says nothing about other countries. Cross-country tables rely on a mix of tests, old norms, small samples and conversion steps, and the choices change the results.

Education, nutrition, health and familiarity with testing all move scores. Wicherts, Dolan and van der Maas reviewed the literature on sub-Saharan Africa in 2010. They estimated an average around 82 against British norms, and warned that the figure had to be read alongside the Flynn effect, differences in schooling and doubts about whether western tests measure the same thing in different settings. Their point is that a gap on a test is not a measure of capacity.

The statistics hub collects the country and state figures we publish, with their sources. Treat them as estimates and compare like with like.

People with more years of education score higher on average, and part of that is education itself. A 2018 meta-analysis by Ritchie and Tucker-Drob pooled 142 effect sizes from 42 datasets and more than 600,000 people. It found that an extra year of education raised cognitive test scores by about 1 to 5 IQ points, with effects across the lifespan. The other part is selection, since people who score higher stay in school longer.

So a group average, such as the figures for occupations, mixes cause and selection. It describes the group, not what any member scored.

About one person in two lands within 10 points of 100, and a score of 95 or 105 is ordinary. A single score also carries a margin of error of a few points on a full professional test, so small differences between two people, or between two attempts, are noise. The IQ test validity guide explains the margin of error.

To place any score on the curve, use the IQ percentile calculator, or read what makes a good IQ for the bands above average.