How the test is built and scored
The test has 32 original questions and a 25-minute limit. Your answers are marked on our server and placed on the usual IQ scale, and every score comes with a margin of error. The norms are provisional, and this page says exactly what that means.
In brief
| Questions | 32, one at a time, with no going back |
|---|---|
| Time limit | 25 minutes |
| Areas | Patterns (10), Series (8), Spatial (7), Logic (7) |
| Scale | Mean 100, standard deviation 15 |
| Scores shown | 70 to 145 |
| Margin of error on the overall score | About 9 points either side |
| Norms | Provisional until real results replace two assumptions |
The questions
Every question is original. None is taken from a published test or from another site.
Each one is generated from a written rule. The pattern questions combine the rule types researchers identified in matrix tests of figures: a value held constant, a progression in steps, a set of three shared out across a row, and figures added or subtracted.1
A separate program then solves every question without being told the answer. A question is kept only if exactly one answer can be defended.
Each sitting draws its 32 questions from a bank of 40 to a fixed plan, so every sitting has the same mix of easy, medium, hard and very hard questions in each area. Within an area the questions start easy and get harder.
There are no vocabulary or general knowledge questions, so the test leans on reasoning more than on schooling or first language.
How the score is worked out
Your answers are sent one at a time and marked on our server. The answer key never reaches your browser, and the clock is kept by the server too.
The number of questions you get right is placed on the usual IQ scale, where 100 is the average and the standard deviation is 15.2
To do that, a test needs to know how hard its questions are and how consistent it is. Established tests measure both on a large sample of people. This test is new, so for now it assumes them.
| Assumption | Value used now |
|---|---|
| Share of people expected to answer an easy question correctly | 85% |
| A medium question | 60% |
| A hard question | 35% |
| A very hard question | 15% |
| Reliability of the whole test | 0.87 |
From these, the expected number correct and its spread follow for the exact questions you drew, and your result is placed on the scale. Scores are shown from 70 to 145, because a test this short cannot measure precisely beyond that.
The margin of error
No test gives the same score twice. The likely range shown with your score is where about 90 in 100 repeat sittings would fall. For the overall score that is about 9 points either side.
Each area has fewer questions, so its range is wider: 14 to 16 points either side.
Professional testing standards ask that any reported score come with a statement of its precision. That is why the range is part of the free result, not an extra.3
What is provisional
Both assumptions in Table 2 will be replaced by measured values once each question has about 250 answers from real sittings. Until then, treat a score as an estimate inside its range.
A sitting answered faster than the questions can be read is marked as implausible and left out of the count shown on the home page.
What this test cannot tell you
- It is for interest. It is not a clinical assessment and cannot diagnose anything.
- It has not been checked against a professionally given test, so no score here should be read as the score you would get on one.
- It covers reasoning with figures, series, shapes and symbols. It does not measure vocabulary, memory or speed, which fuller tests do.
- It should not inform a decision about school, work or health. A psychologist who gives a test in person is the right source for that.
- It was written for adults. It is not for anyone under 13.
Sources
- Carpenter, P. A., Just, M. A., and Shell, P. (1990). What one intelligence test measures: A theoretical account of the processing in the Raven Progressive Matrices Test. Psychological Review, 97(3), 404–431. doi.org/10.1037/0033-295X.97.3.404
- Neisser, U., et al. (1996). Intelligence: Knowns and unknowns. American Psychologist, 51(2), 77–101. doi.org/10.1037/0003-066X.51.2.77
- American Educational Research Association, American Psychological Association and National Council on Measurement in Education (2014). Standards for Educational and Psychological Testing, chapter 2: Reliability/Precision and Errors of Measurement. www.testingstandards.net/open-access-files.html
Changes to this page: 12 November 2026, first version.