What Is IQ? A Plain-Language Explanation

"IQ" gets used as a synonym for intelligence, but the two are not the same thing. Intelligence is a broad idea that psychologists still argue about. IQ is something much narrower and much more specific: a number describing how your performance on a particular test compares to other people's.

The number is a comparison, not a measurement

This is the single most useful thing to understand about IQ, and the thing most often missed. An IQ score is not a quantity of intelligence in the way that 70 kg is a quantity of mass. Nothing is being counted. What is being reported is a position in a distribution.

When you take a test, you answer some number of questions correctly. That raw count is meaningless on its own — a raw score of 14 tells you nothing without knowing how hard the questions were and how everyone else did. So the raw score is converted by comparing it against a norm group: a large, carefully selected sample of people who took the same test under the same conditions.

The conversion is defined so that the norm group's average comes out at 100, and so that scores spread out with a standard deviation of 15. An IQ of 115 means "one standard deviation above the average of the comparison group." An IQ of 85 means one standard deviation below. The number describes where you sit relative to others, and nothing else.

A direct consequence: an IQ score is meaningless without knowing who it was compared against. A score normed on Japanese adults and a score normed on Brazilian teenagers are not interchangeable, even if the number printed on both reports is 112.

Where the idea came from

In 1904 the French government asked the psychologist Alfred Binet to find a way to identify children who would struggle in ordinary classrooms and needed additional support. With Théodore Simon he produced the Binet–Simon scale in 1905 — a set of graded tasks of increasing difficulty. A child's performance was expressed as a "mental age": the age at which a typical child solved the same tasks.

The quotient in "intelligence quotient" arrived shortly after, from the German psychologist William Stern, who proposed dividing mental age by chronological age and multiplying by 100. A ten-year-old performing like a typical twelve-year-old scored 120. That formula is where the number 100 as "average" originally comes from.

The ratio formula was abandoned decades ago, because it breaks down in adulthood — mental age stops climbing with chronological age, so every adult's score would fall over time for no meaningful reason. It was replaced by the deviation IQ used today, which compares you to others in your own age band rather than to a mental age. The scale was kept at mean 100 purely for continuity with the older system.

Binet himself was cautious about his own instrument. He warned against treating the scale as a fixed measure of an unchangeable quality, and intended it as a tool for identifying children who needed help — not for ranking them. That caution is worth carrying forward.

What "general intelligence" means

In 1904 Charles Spearman noticed something that has held up remarkably well: people who do well on one kind of cognitive task tend to do well on others. Vocabulary performance correlates with spatial reasoning performance, which correlates with arithmetic performance, and so on. The correlations are far from perfect, but they are consistently positive.

Spearman proposed that a common factor underlies all of them, which he called g, for general intelligence. Modern IQ tests are built to estimate g as efficiently as possible — which is why a good test mixes several different kinds of problem rather than testing one skill deeply.

Later work refined the picture considerably. The widely used Cattell–Horn–Carroll (CHC) framework describes a hierarchy: g at the top, with broad abilities beneath it. Two of these come up constantly:

• Fluid reasoning (Gf) — solving novel problems without relying on prior knowledge. Spotting the pattern in an unfamiliar sequence of shapes uses fluid reasoning.
• Crystallised knowledge (Gc) — accumulated knowledge and learned skill. Vocabulary and general knowledge questions tap crystallised ability.

The two behave differently over a lifetime. Fluid reasoning tends to peak in early adulthood and decline gradually afterward; crystallised knowledge typically keeps growing into middle age and beyond. Non-verbal tests — the kind used in most online assessments, including this one — lean heavily on fluid reasoning, which is one reason they travel across languages and cultures more easily than vocabulary-based tests.

What IQ scores do and do not predict

IQ scores are among the better-studied predictors in psychology, and they do carry real predictive weight. Scores correlate moderately with academic achievement and with job performance, with the relationship tending to be stronger for cognitively demanding roles. Population-level associations with health and longevity outcomes have also been reported.

But "correlates moderately" needs to be read carefully, because it is routinely overstated. A moderate correlation across a large population means the average outcome differs between score bands. It does not mean you can predict an individual's outcome from their score. The variation within any score band dwarfs the difference between bands. Plenty of people with high scores achieve little; plenty with ordinary scores achieve a great deal.

An IQ score also says nothing at all about a long list of things that matter enormously: persistence, curiosity, judgement, creativity, emotional insight, integrity, physical skill, or the ability to work with other people. These are not measured by any IQ test, and their absence from the score is a limitation of the instrument, not evidence that they are unimportant.

How stable is a score?

Reasonably stable in adulthood, and less so in childhood. Professionally administered tests typically report high test–retest reliability over short intervals. Over decades, adult scores are considerably more stable than most people expect, though not fixed.

In children the picture is looser — scores measured in early childhood predict adult scores only weakly, and individual trajectories can shift substantially during development.

Heritability estimates for intelligence are substantial and, counter-intuitively, tend to increase with age: genetic influence appears larger in adults than in young children. Heritability is a population statistic, though. It describes how much of the variation within a particular population under particular conditions is associated with genetic differences. It does not tell you the cause of any individual's score, and it does not mean a score cannot change.

Reading your own result sensibly

A few habits make an IQ result more useful and less misleading:

• Read it as a band, not a point. Every test has measurement error. A score of 118 is better understood as "somewhere in the low-to-mid 110s to low 120s" than as exactly 118.
• Check the standard deviation. A score of 130 on an SD-15 scale is not the same as 130 on an SD-16 or SD-24 scale. Always look for the scale before comparing two numbers.
• Weigh the conditions. Fatigue, distraction, illness, unfamiliarity with the format and simple nervousness all depress scores. An unsupervised test taken at midnight is not a controlled measurement.
• Do not treat it as a verdict. A number describing your position on one afternoon's set of puzzles is not a statement about your worth or your ceiling.

In short

An IQ score is a standardised comparison of your performance on a specific test against a specific reference group, expressed on a scale where 100 is average and 15 points is one standard deviation. It captures something real and reasonably stable about reasoning ability, it predicts group-level outcomes with moderate accuracy, and it leaves out a great deal of what makes a mind effective. Held that way, it is genuinely informative. Held as a verdict on a person, it is simply wrong.