IIF Certified Assessment
Start IQ Test
IIF Certified Assessment
On the scale most modern tests use, 120 sits around the 91st percentile. Change the scale and the same number moves. Add the measurement error every test carries and it stops being a point at all.
Someone tells you their IQ is 120. Before that means anything at all, you need three further pieces of information: which scale, measured against whom, and with how much error. Without them the number is a decoration. With them it is a reasonably precise statement — and a narrower one than most people assume.
Modern IQ scores are what psychometricians call deviation scores. They are not a ratio of anything. The test is given to a large reference sample chosen to represent a population, the raw results are arranged so the average is set to 100, and the spread is set so that one standard deviation equals a fixed number of points. Standard deviation is simply a measure of how spread out the scores are: on a scale where it is 15, roughly two in three people fall between 85 and 115, and roughly 19 in 20 fall between 70 and 130.
On that mean-100, standard-deviation-15 scale, a score of 120 is one and one-third standard deviations above the average, which places it at approximately the 91st percentile of the reference sample. Not the 95th, not “top 5%” — approximately the 91st, meaning roughly nine people in a hundred in that sample scored higher.
This is why a printed number without its scale is close to meaningless, and why comparing two people’s scores from two different tests is a conversion problem, not a subtraction problem.
A score is a position in a specific reference sample on a specific scale — not a quantity of anything you carry around.
No test measures without error. Test manuals publish a figure called the standard error of measurement, which describes how much a person’s observed score is expected to bounce around their true standing across repeated sittings. That is why a responsible report gives a confidence interval — a range within which the true score most likely lies — rather than a bare number.
The practical consequence is blunt: two scores a few points apart are usually indistinguishable. A 118 and a 122 are the same result. Someone whose score moved from 116 to 121 between two sittings has not improved; they have produced two estimates from the same band, which is what the band is for.
A number worth keeping comes with the three things above attached: the named scale it was measured on, the percentile against a stated reference sample, and the confidence range around it. If a result gives you a single figure and nothing else, the missing information is not a detail that was left out for simplicity — it is the part that would have let you check the claim.
Find your IQ score now! →None of this makes 120 uninteresting. On the standard modern scale it is a solidly above-average result, comfortably above the middle of the distribution and a long way from either extreme. It is simply a narrower statement than the internet treats it as: on this scale, against this sample, on this day, within this range.
On the scale most modern tests use — mean 100, standard deviation 15 — it is clearly above average, at approximately the 91st percentile of the reference sample. It is not in the range usually described as very superior, and on a scale with a larger standard deviation, such as the Cattell scale used by some societies, the same printed number sits considerably lower.
Three reasons, all normal. They may use different scales. They may be normed on different reference samples. And every test carries measurement error, so repeated sittings scatter around your true standing. Differences of a few points are expected rather than surprising.
Roughly the 91st on a mean-100, standard-deviation-15 scale — about nine people in a hundred in the reference sample scored higher. On a standard-deviation-16 scale it is slightly lower, and on the standard-deviation-24 Cattell scale it is far lower. The percentile follows the scale, not the number.
Corrections: spotted an error? Email corrections@iqmetrics.org and we will update this story and note the change here.
Average scores climbed for decades, then flattened and in places slipped. None of it shows on a score report, because every test is reset so the average is 100 again — which is why a 1990 score is not a 2026 score.
Two people can perform identically on two different online tests and walk away with scores twenty points apart. The scale, the comparison group and the margin of error are what separate an assessment from a quiz.
Our IIF-certified assessment reports your score with its scale, percentile and confidence range — and a breakdown of the cognitive domains behind it.
Start IQ Test →
