How the test is built and scored
This page sets out exactly what the test measures, how a score is calculated and where its limits are. Last updated 11 October 2026.
What the test contains
The test has 36 multiple-choice questions and a 25 minute time limit. Every question was written for this site. They fall into four areas.
| Area | Questions | What you do |
|---|---|---|
| Spatial | 12 | Find the rule in a three by three grid of shapes and choose the missing tile from six options. |
| Numerical | 9 | Continue a number series. |
| Logical | 9 | Decide what must follow from a set of statements. |
| Applied | 6 | Solve a short word problem using everyday arithmetic. |
Questions appear from easiest to hardest and rotate through the four areas. Each question is shown once. You can skip a question but you cannot go back to it.
How points are awarded
Each question is rated easy, medium or hard and is worth 1, 2 or 3 points. A wrong or skipped answer scores nothing, and there is no penalty for guessing. Time is not part of the score, other than the overall limit.
| Difficulty | Questions | Points each | Points available |
|---|---|---|---|
| Easy | 11 | 1 | 11 |
| Medium | 15 | 2 | 30 |
| Hard | 10 | 3 | 30 |
| Total | 36 | 71 |
How points become an IQ estimate
IQ scales are defined so that the average is 100 and one standard deviation is 15 points. To place a raw score on that scale we need to know the average raw score and how widely raw scores are spread.
At the moment we use an average of 33 points and a spread of 12 points. The formula is:
estimate = 100 + 15 × (your points − 33) ÷ 12, rounded to a whole number and kept between 60 and 145.
So 33 points gives 100, 45 points gives 115 and 57 points gives 130. The percentile comes from the standard bell curve for a scale with mean 100 and standard deviation 15.
These two numbers are a starting assumption, not a measurement. They come from how hard each question was designed to be: we expect most people to get most easy questions right, about half of the medium ones and about a quarter of the hard ones. The test has not yet been taken by a large group of people, so the scoring has not been checked against real results. When it has, we will update the numbers and say so here.
The likely range
Your result page shows a range of 8 points either side of your score. A short test taken once, at home, cannot pin a score down more tightly than that. Tiredness, distractions, a lucky guess or a careless slip can each move the result by a few points.
What the test does not do
- It does not measure memory, processing speed, vocabulary or general knowledge, which full clinical tests include.
- It is not supervised, so it cannot know whether a calculator or outside help was used.
- The same questions are used each time, so repeat attempts tend to score higher.
- It has not been standardised on a representative sample, and no reliability or validity figures exist for it yet.
- It is not a clinical instrument. Do not use it for medical, educational or employment decisions.
How it differs from a clinical test
Tests such as the Wechsler Adult Intelligence Scale are given one to one by a trained psychologist, take over an hour, cover many more abilities and are scored against large samples chosen to represent the population. If you need a score that carries weight, that is the route to take.