Forty questions. Four perspectives.
The exercise covers spatial, numerical, logical and applied reasoning, with ten questions in each area and a 25-minute total limit. It uses English instructions and basic arithmetic, so it is not language-free or culture-free.
Correct answers and weighted scores
A domain percentage is simply the number correct out of ten. Nine correct answers means 90%. The draft weighting assigns one point to each of three introductory items, two to each of four intermediate items, and three to each of three advanced items per domain.
| Item tier | Per domain | Points each | Domain total |
|---|---|---|---|
| Introductory | 3 | 1 | 3 |
| Intermediate | 4 | 2 | 8 |
| Advanced | 3 | 3 | 9 |
| Total | 10 | n/a | 20 |
The whole exercise therefore has 40 questions and 80 possible weighted points. The experimental mapping is round(55 + 90 × weighted points ÷ 80), version linear-provisional-v1. Zero points maps to 55, forty to 100, and eighty to 145. This is a transparent provisional scale, not a formula fitted to population data. Neither its midpoint nor its weighting establishes an average IQ or a standardized test. Stored results retain their original scoring version when the model changes.
What the curve actually means
The reference curve uses a normal distribution with mean 100 and standard deviation 15. For a given score, the theoretical percentile is Φ((score − 100) / 15). It is not a distribution measured from IQInsight users, and it does not establish your actual rank in the population.
For example, 130 corresponds to approximately the 97.7th percentile on that mathematical scale. Explore the relationship with our percentile calculator.
Mathematical reference: NIST/SEMATECH e-Handbook, normal distribution. It explains the distribution and its cumulative function; it is not an endorsement or validation of this exercise.
What has not been validated
We have not established population norms, test–retest reliability, measurement error, age-specific interpretation or equivalence with a professionally administered IQ test. A result should not be used for diagnosis, selection, employment, educational placement or legal decisions.
How to read the four areas
A higher correct-answer rate is an observation about these particular questions on one occasion. The areas have not been equated for difficulty. Differences do not establish fixed strengths, weaknesses, learning styles or career suitability.
Samples, practice and transparency
We do not currently publish participant statistics. If aggregate data is introduced, we will state its source, version, inclusion rules and sample size. Small groups will not be presented as reliable population comparisons.
Some questions are public samples. Repeating them can improve performance through familiarity. Reordering questions does not make an attempt an independent measurement.