How our IQ test works, and what it cannot tell you
We would rather you understand your result than just trust it. Here is exactly how the test is built, scored and checked, including the parts that are still provisional.
What the test measures
The test has 24 questions in four sections of six: pattern reasoning (grids of shapes), number reasoning (sequences), verbal reasoning (analogies, odd-one-out, vocabulary) and logical reasoning (deductions from stated facts). Most items test fluid reasoning, the ability to work out new problems, with a smaller verbal component. Every question was written for this site. Within each section, questions run from easy to hard. For background, see what IQ is.
How the score is calculated
- Difficulty weighting. Each question is worth 1 to 5 points depending on how hard it is, so a hard question counts more than an easy one.
- Guess correction. A wrong answer subtracts a small fraction of the question’s points, set so that random guessing earns zero on average. An unanswered question scores zero. This means guessing when unsure never costs you in expectation.
- Conversion to the IQ scale. The total is converted to the standard IQ scale, where the population average is 100 and the standard deviation is 15. Scores are limited to a range of 70 to 145, because a 24-question test cannot distinguish reliably beyond that.
- Percentile. Your percentile is calculated from your IQ using the normal distribution. See the IQ score chart.
What is still provisional
This is the main reason we show a likely range instead of a single exact number, and why you should not treat the result as a precise measurement. The range we show is about 11 points either way (a 90% interval). That figure assumes a reliability of roughly 0.80 for a test of this length, which is an assumption we have not yet measured.
Quality checks on your result
Your report flags results that are likely to be unreliable, so you are not misled by a score that does not reflect your ability. We check for: many unanswered questions, very fast answering (a typical sign of guessing), and answers clustered in the same on-screen position. When a flag appears, treat the score as provisional and consider retaking the test calmly.
Limits of any online IQ test
- It is an estimate, not a diagnosis. It cannot identify learning differences, giftedness or disability. Those need an assessment by a qualified psychologist.
- It is not age-adjusted. Professional tests compare you with people your own age. This test does not, so it is designed for ages 16 and over and is not suitable for children.
- Conditions matter. Noise, tiredness, stress, a small screen and unfamiliarity with puzzle formats all affect the score.
- Practice effects. Retaking the same questions mostly measures your memory of them. Wait at least four weeks.
- Language and culture. The verbal section is in English and assumes familiarity with common English words.
- A ceiling. Very high ability cannot be measured accurately by a short test.
Average IQ by age
On professional tests the average is 100 at every age, because each person is compared with others of the same age. Raw problem-solving speed changes across a lifetime, but the IQ scale adjusts for that. Since this test is not age-adjusted yet, we do not publish age-specific averages for it.
When to see a professional
If you need an IQ score for educational placement, clinical decisions, workplace requirements or high-IQ society membership, use a supervised test from a qualified psychologist or the society itself. See online vs professional IQ tests for the full comparison.