Frequently Asked Questions

Understanding your result

Is the "estimated IQ" on my result a real IQ score?

It is an estimate, and we label it as one. We work out where you stand among everyone who has taken the same form, then convert that position to the familiar mean-100 scale so it is easy to read, and we always show it with its range. A clinical IQ is different: it means something because the test was standardised against a representative normative sample, which is slow, expensive, professionally recruited work that we have not done. So read our number as "where I stand among this test's takers", not as a population IQ. The methodology page sets out the whole chain.

What does a percentile actually mean?

It is your position in a queue, not a mark out of 100. The 70th percentile means you performed better than about 70% of the people who took this test — it does not mean you got 70% of the questions right, and the two numbers are usually nowhere near each other. This trips people up constantly, in both directions: a 95th-percentile result is not an A, and a 40th-percentile result is not a fail. (Your overall figure is a weighted average of your six domain percentiles, so it is close to, but not exactly, such a rank.)

Why is my result shown as a range?

Because a single number would overstate what twelve questions can establish. The band is a deliberately conservative allowance for error: the short form places you roughly, the 60-item form places you much more precisely because it has five times the evidence. The practical rule is that if two bands overlap, treat the difference between them as uncertain. Read the band, not its midpoint.

What is a 'normal' IQ?

On the clinical mean-100 scale, 100 is average by construction and roughly two thirds of people fall between 85 and 115. About one person in 44 scores 130 or above, and the same proportion 70 or below. Those boundaries come from how the scale was built rather than from anything discovered about people. On this test, an estimated 100 — the middle of our sample — plays the same role.

What if I scored lower than I expected?

First, check the conditions. Tiredness, interruptions, a phone screen, doing it late at night, or treating the first few items as a warm-up all cost real points, and none of them say anything about your reasoning. Second, look at the six domains rather than the overall figure — an uneven profile is the normal case, and a single low domain pulls the composite down by its weight. Third, remember what this is: one short test, taken once, under one set of conditions. It is a snapshot, not a verdict, and it is not a measure of your worth or your potential.

How cognitive testing works

Why do my six domain scores differ so much from each other?

Because that is the usual pattern. The domains are correlated but genuinely distinct, so most people are visibly stronger in some than others — a tall verbal bar beside a shorter spatial one is a profile, not a defect. On the short form, with two questions per domain, most of the difference between domains is measurement noise; the full form puts ten questions behind each one and gives a far more trustworthy profile.

What is the difference between fluid and crystallised intelligence?

Fluid reasoning is working out the rule behind something you have never seen before; crystallised reasoning is bringing knowledge and vocabulary you already have to bear on a problem. The distinction matters because they behave differently over a lifetime: fluid reasoning tends to peak in early adulthood and decline slowly, while crystallised reasoning rises or holds steady into the 60s. This is why "does IQ decline with age" has no single answer — it depends entirely on which kind you mean.

Are IQ tests culturally biased?

Every cognitive test carries some cultural and linguistic loading, and one that claims otherwise is overselling itself. Verbal items depend on fluency in the language you take the test in; pattern and numerical items depend on it far less, which is why they are often described as more culture-fair — though even "fair" is relative, since familiarity with test formats is itself learned. What a score describes is performance on this battery, in this language, against this sample — not an innate quantity.

Is the test timed?

No. There is no clock on either form, so you can think each item through. The one domain this affects is processing speed, which clinical tests measure with a stopwatch: here its items are scored for accuracy only, and it carries the smallest weight in your overall result. Answers given implausibly fast are flagged as low-confidence.

Your data

What happens to my answers?

We store which items you saw, whether each was answered correctly, how long each took, and the language you took it in. That lets us review how difficult each question really is, and the growing pool of completed sittings is what makes the percentiles more stable over time. We do not sell it, we do not enrich it with anything bought from elsewhere, and we do not train AI models on individual responses. If you take the test without an account there is no identity attached to it at all; if you later sign in, the finished session attaches to your account and you can delete it.