How Is IQ Actually Measured? The Science Behind the Number
The first intelligence test wasn't built to find geniuses. It was designed in Paris in 1905 for a much humbler job: deciding which schoolchildren needed extra help to keep up in class.
A school assignment in Paris
French schooling had been compulsory since the 1880s, and teachers needed something sturdier than their own impressions to recognize the students who were falling behind. The Ministry of Public Instruction handed the problem to psychologist Alfred Binet, who worked with Théodore Simon. Their answer was the Binet-Simon scale, the first intelligence test put to practical use.
Binet was wary of his own invention. He saw it as a way to steer educational help toward the right children, not as a permanent stamp on a child's ability. That a century later its descendants would end up as shareable "certificates" on social media would have struck him as strange, to say the least.
From mental age to quotient
Binet and Simon reported results as a "mental age": the typical age at which children could solve what this child had solved. In 1912, German psychologist William Stern suggested dividing mental age by actual age and multiplying by 100. That gave us the intelligence quotient, or IQ. An 8-year-old who solves what 10-year-olds solve scores 125 (10 ÷ 8 × 100); one who performs exactly at their age scores 100. The idea was elegant because it packed into one number something that used to take paragraphs to describe, which is why it spread so fast, but it carried a problem that took decades to sort out.
Why the formula had to change
With adults, the math falls apart. Mental age doesn't keep climbing, and saying a 40-year-old performs "like a 60-year-old" means nothing. David Wechsler solved this with what's called deviation IQ, which he already used in his 1939 scale and which became standard with the WAIS, published in 1955 and still among the most widely used tests in the world. Instead of comparing ages, your score is compared with a large sample of people your own age, on a scale where the average is 100 and the standard deviation is 15.
What landing on the bell curve means
On that scale, about 68% of the population falls between 85 and 115, and roughly 95% between 70 and 130. Scores far outside that range, high or low, are uncommon by statistical definition. But the number only tells you where you stand relative to a reference group. It says nothing about what a person is worth. A 115, for example, doesn't mean "15% better" than a 100: it means sitting one standard deviation above the average, around the 84th percentile. And a 130 means outscoring roughly 98% of the reference group.
It's also worth remembering that no score is exact. Every administration has a margin of error, and professional tests acknowledge it by reporting a range rather than a single value. Retake a test short on sleep or in a hurry and your result will move; a lone number, stripped of that context, looks more precise than it is.
What a modern test measures and what it leaves out
The WAIS-IV and the Stanford-Binet aren't single tests but batteries of subtests, some of them timed, covering different abilities: verbal comprehension, visual and abstract reasoning, working memory, and processing speed. The overall IQ comes from combining all of those results, not from one magic question; each index can also be read on its own, and a very uneven profile says more than the average does. Raven's Progressive Matrices, by contrast, focus on a single kind of reasoning: completing patterns of shapes.
What's left out is plenty: creativity, emotional intelligence, social and practical skills. A high IQ tells you a lot about abstract reasoning and very little about the rest of what makes someone capable in everyday life.
The test on this site borrows that logic: questions that get harder, a time limit, several kinds of reasoning. It reports the result on the same scale (mean 100, standard deviation 15), but it hasn't been normed on a population sample and it doesn't replace an evaluation with a psychologist. Take the test, check your percentile, then come back to the bell curve section to put it in context.
