How the Cogniva test works out your IQ range
Cogniva's reasoning test has 20 original visual puzzles. Your number of correct answers is placed on the IQ scale, where the average is 100 and the standard deviation is 15, by this test's norm. We report a range of about ±8 to 10 points because 20 questions cannot support a single exact score.
Your estimated IQ range places you on this test's own scale. It is not a comparison with the general population.
This is not a clinical or medical test.
Don't use this result for hiring, admissions, or any other decision about a person.
Live figures are temporarily unavailable. Please reload in a few minutes.
What the test measures
The test measures visual and logical reasoning: finding the rule in a pattern and applying it. There are four pattern types, with five questions each: Progression, Rotation, Combination and Distribution. Every question was written for Cogniva. None is copied from a published or clinical test.
How your score becomes a range
- We count your correct answers out of 20.
- We place that score on the IQ scale with the current norm.
- We convert that position to the IQ scale, with an average of 100 and a standard deviation of 15.
- We add the margin. A score from 20 questions is less precise than one from a 90-minute clinical test, so your report shows a range, for example 114 to 130, and never one number.
What the bands mean
Above average means your whole range sits above 100. Below average means your whole range sits below 100. Average means your range includes 100.
Who you're compared with
Most people who take this test find it through our ads and website, so their average may differ from the general population's.When we have enough data, we weight results by country and age group, and we say so here when we start.
How the norm changes
Each norm version is recorded with its sample size and date, and your report names the version that scored you. The cut-off for 'above average' comes from the published norm and is never adjusted to sell more reports. If more than 60% of test takers land above average for a sustained period, we revise the norm upward.
We recalibrate at 300 and 1,000 completed tests, and we list every version below.
What we don't claim
This test measures how you solve visual patterns. It doesn't measure processing speed, memory, emotional intelligence, or job aptitude.We don't give a population percentile or an age-based percentile.
How the written report is produced
Your score, range and pattern results are calculated by code before any AI is involved. An AI model then writes the explanations from those calculated results, using a fixed template with automatic checks for accuracy and tone. We review a sample of reports by hand, including low scores and very fast tests. If the AI step fails, you get a structured report written from the same data.
You can read a complete example in the sample report.
How the questions were developed
Each question has an explicit rule and exactly one answer that can be justified. Before launch, volunteers took the test so we could remove ambiguous questions and fix ones that were too easy or too hard. Questions are versioned, and the test version appears on every report.
Limitations
- Twenty questions give a wide margin.
- You take the test on your own device, with whatever distractions you have.
- If you have practiced similar puzzles, your score may be higher than on a first attempt.
- The questions are visual, so screen size and eyesight can affect results.
This test is visual. If you use a screen reader, it can't measure your reasoning fairly. If you bought the report anyway, email us and we'll refund it in full.
Norm history
| Version | Status | Sample size | Date |
|---|
Questions about the method? Email support@cognivainstitute.org. More answers are in the FAQ.