What an IQ score actually means

An IQ score is a rank, not a quantity. Modern tests convert your raw answers to a scale where the average is set at 100 and most results cluster tightly around it. A score of 115 does not mean you have 15 more of something than average — it means you performed better than roughly five people in six who took the same test.

Why is the average always 100?

Because it is placed there. Raw scores — how many questions you got right — are converted onto a standard scale that is anchored so its mean sits at 100.

That is worth sitting with for a moment. The average is not a result anyone discovered. If the whole population improved next year, the scale would be re-anchored and the average would be 100 again. The number tells you where you stand relative to other people, and nothing about an absolute amount.

How are scores spread out?

Scores follow a bell curve with a standard deviation of 15 on the most widely used scales. Applying the standard rule for a normal distribution:

Two implications follow, and both get lost in the way scores are usually talked about.

Average is a wide band. Two thirds of everyone falls in a thirty-point range. Someone at 92 and someone at 110 are both squarely ordinary.

The extremes are genuinely rare. By the same arithmetic, a score above 130 puts you in roughly the top 2%.

What does a percentile mean here?

A percentile converts the score into the thing people actually want to know: how many people scored at or below you. A score of 100 is the 50th percentile — dead centre. The further from 100 you go, the faster the percentile moves, because the curve is crowded in the middle and thin at the edges.

That crowding is why small differences near the average mean very little. A handful of points between two people in the middle of the distribution can come down to a good night's sleep.

Why do two tests give different numbers?

Several reasons, none of which requires either test to be broken.

• Different reference groups. A score is relative to whoever the test was normed against.
• Different content. Verbal, numerical and spatial reasoning do not measure the same thing, and tests weight them differently.
• Practice with the format. Taking any test a second time usually produces a better result, which says more about familiarity than about you.
• Conditions. Tiredness, distraction and time pressure move scores.

Is an online test the same as a clinical one?

No, and it is worth being plain about it. A clinical assessment is administered in person by a qualified professional under controlled conditions, as part of a wider picture. A test you take on a phone is none of those things.

That does not make an online test worthless — it can be a genuinely interesting way to see how you handle a kind of reasoning, and repeated practice makes the format familiar. It does mean the number it gives you should not be treated as a diagnosis or a credential.

One thing the research is clear about: getting better at a test is not the same as getting better at thinking. The largest review of the evidence, by Simons and colleagues in Psychological Science in the Public Interest, found strong evidence that practice improves performance on the trained task, much weaker evidence that it transfers to closely related tasks, and little that it transfers to everyday cognitive ability. We cover that in do brain training apps work.

So what is the number good for?

As a rough position on a curve, taken under known conditions, it is meaningful. As a summary of a person, it is not — it describes performance on one test on one day, and the scale itself is built to guarantee that most people land near the middle.

If you want to see how you handle this kind of reasoning, iQ Asist has tests in three lengths and reports where the result sits on the distribution rather than just handing you a number.

More on this in tests and training.