What IQ scores mean
An IQ score is a way of saying how your performance on a set of reasoning tasks compares with other people's. That's all it is: not a measure of your worth, your potential or your knowledge, but a comparison on a particular set of problems. Understood that way, it can be genuinely interesting and useful. This guide explains how the scale works, what the numbers mean, why a good result always comes with a range, and where online tests fit in.
In short: the average score is 100. About two in three people score between 85 and 115. A score of 115 means you did better than about 84% of people, and 130 better than about 98%. Every score has some uncertainty, so a result like "116, likely range 110 to 122" is more honest than a single number.
The scale: why 100 is average
Modern IQ scores aren't calculated from a "mental age" any more, as the earliest tests were. Instead, a test is first given to a large group of people chosen to represent the population, called the norm group. Their results are then rescaled so that the average is exactly 100 and the spread is fixed, using a measure called the standard deviation, which is set at 15 points on almost all modern tests.
Scores in a large population fall in a bell-shaped curve: most people are near the middle, and the further from 100 you go, in either direction, the fewer people there are. Because the shape of this curve is known, every score corresponds to a share of the population.
| Score | Usual description | Share of people who score lower (roughly) |
|---|---|---|
| 130 and above | Very high | 98% |
| 120 to 129 | High | 91% to 98% |
| 110 to 119 | High average | 75% to 91% |
| 90 to 109 | Average | 25% to 75% |
| 80 to 89 | Low average | 9% to 25% |
| Below 80 | Well below average | under 9% |
The names vary a little between tests and publishers, but the numbers underneath them are the same, because they come straight from the shape of the curve.
Percentiles: the more useful number
A percentile tells you the share of people who scored lower than you. If you're at the 84th percentile, you did better than 84 out of every 100 people. Many people find this easier to understand than the score itself, and it makes the scale's shape obvious: going from 100 to 115 moves you from the 50th to about the 84th percentile, a jump of 34 places, but going from 130 to 145 moves you only from about the 98th to the 99.9th.
That's also why differences at the extremes are so hard to measure. Very few people score above 140, so a test needs many very hard puzzles to tell 140 from 145 reliably. Most tests, online or not, are least precise at the very top and the very bottom of the scale.
Why every score should come with a range
No test measures perfectly. If you took two different versions of the same test a week apart, you'd get two slightly different scores, because of which puzzles happened to come up, how well you slept, and plain luck with guesses. Test makers measure this wobble and report it as the standard error of measurement. Professional reports usually give a confidence interval: a range that your "true" score very likely falls in.
A short test has more wobble than a long one, which is why a 20-puzzle test should show a wider range than a two-hour clinical assessment. A site that gives you a single exact number with no range is claiming more precision than any test of that length can have. The IQ Centre report shows your score with the range it most likely falls in, worked out from which puzzles you got right and how informative each one was.
Online tests and clinical tests
The best-known clinical tests, such as the Wechsler Adult Intelligence Scale (WAIS) and the Stanford-Binet, are given one to one by a trained psychologist and take an hour or more. They measure several abilities separately, use norms from carefully chosen samples of the population, and are used to support decisions such as diagnosing a learning disability. Only a qualified professional can give and interpret them.
An online test can't replicate that, and you should be wary of any that claims to. What a well-built online test can do is use the same kinds of puzzles, score them sensibly, and be honest about its uncertainty. Its main limitations are:
- Conditions vary. A small phone screen, a noisy room or interruptions all lower scores.
- Who takes it isn't random. People who choose to take an IQ test online aren't a representative sample of the population, so any comparison is approximate.
- It's short. A shorter test means a wider range around your score.
Treat an online result as a good estimate and an interesting snapshot, not a diagnosis. If you need a score for an official purpose, such as a clinical assessment or an application to a high-IQ society, you'll need a supervised test.
What IQ scores do and don't tell you
IQ is one of the most studied measurements in psychology. Across large groups of people, scores are linked with how well people do at school and at work, and they're fairly stable across adult life. But the link for any one person is loose. Plenty of people with average scores do extraordinary things, and plenty with very high scores don't, because motivation, knowledge, skills, health, opportunity and personality all matter enormously and an IQ test measures none of them.
It also isn't fixed. Scores in many countries rose by roughly three points per decade through much of the twentieth century, an effect named after the researcher James Flynn, which is why tests have to be re-normed regularly. Education raises scores too. And on any single test, your result can move by several points from one day to the next. Our guide to improving your IQ score covers what does and doesn't make a difference.
Common misunderstandings
"A score of 100 is mediocre." 100 is exactly in the middle of everyone. Half of people score lower.
"A few points' difference matters." Two people scoring 112 and 116 on the same test can't really be told apart; their ranges overlap almost entirely.
"Scores from different tests are directly comparable." Mostly, but not exactly. Tests use different puzzles and different norm groups, and some use a spread of 16 or 24 points rather than 15, which changes what a given number means. A score of 132 on a scale with a spread of 16 corresponds to about 130 on the usual scale.
"IQ measures how much you know." Good reasoning tests are designed to depend as little as possible on knowledge. That's why they use abstract shapes, simple numbers and made-up words.
See where you sit
The IQ Centre test takes about 15 minutes: twenty puzzles of the types described in IQ test questions explained, plus a short memory game.