What is the average IQ, and why is it always 100?

100 is a decision, not a measurement

When a test publisher builds an intelligence test, it gives the test to a large, carefully sampled group of people and records their raw scores. Those raw scores are then transformed so that the group's middle becomes 100 and its spread becomes a standard deviation of 15. Every later test-taker is compared with that group. So the average is 100 in the same way that water boils at 100 on the Celsius scale: the number marks a reference point that was chosen, and the scale is built around it.

This is why the phrase "the average IQ is rising" needs care. Raw performance on the same old items has risen over decades in many countries, the pattern usually called the Flynn effect, but each new edition of a test is re-normed so that the average is set back to 100. A score of 100 on a 1980 edition and a score of 100 on a current edition both mean "the middle of the group that took this edition", not the same amount of solved items.

What the 85–115 range means

Because the scale is built on a normal distribution with a standard deviation of 15, the proportions are fixed. Roughly 68% of a norm sample scores between 85 and 115, about 95% between 70 and 130, and about 2% above 130. A score of 115 is at about the 84th percentile; a score of 130 is at about the 98th; a score of 85 is at about the 16th. Test publishers attach descriptive labels to these bands, and the labels differ between publishers, so the percentile is the more portable number.

A published score also carries a confidence band. Test manuals report a standard error of measurement of a few points, which means two people whose scores differ by three or four points cannot be told apart with confidence. Reading "103" as better than "100" reads noise as signal.

Why tables of average IQ by country mislead

Pages listing an "average IQ" for every country combine studies that used different tests, different years, different age groups and very different sample sizes, then average them as if they were the same measurement. Some entries rest on a few hundred schoolchildren tested decades ago; others are estimates borrowed from neighbouring countries. Because every test is re-normed to 100 inside its own reference group, comparing raw performance across groups also requires the same items, the same administration and the same era, which those tables do not have.

The honest version of the question is narrower: on one specific test edition, normed on one specific population, how did one specific group score? Anything broader is a chart with the appearance of precision.

What "average" means for one person

For an individual, an average score means the test found nothing unusual in the sampled abilities on that day. It does not describe learning speed at work, creativity, judgement, or how the person performs after rest, practice or in another language. Educational and clinical assessments use IQ scores together with interviews, history and other measures, because the number on its own answers a narrow question.

What a browser test can tell you instead

MyTestAtlas does not publish an IQ, because none of its tests has a norm sample or a controlled administration. What a browser test can do honestly is show how many items you solved in each rule family and at each difficulty tier, and let you review every answer. That is useful for practice and for understanding the item types that standardized tests use; it is not a substitute for a professionally administered test when a real score matters.