Your result is not a subjective impression; it is the output of a transparent calculation. This page walks through every step, from counting your answers to placing that count on the IQ scale, including the maths and an honest account of the margin of error around any score.
From answers to a result
Scoring happens in two stages, both on your own device. First, your answers become a raw score: a count of what you got right, weighted toward the items that carry the most signal. Second, that raw score is placed on a standard scale so the number means something relative to other people rather than in isolation. A raw count of, say, 24 tells you nothing on its own; what matters is where 24 sits in a reference distribution. The same procedure runs for everyone, paying user or not, because there is no paying tier, and nothing about the calculation changes from one person to the next.
Calculating the raw score
The test runs 33 items across four domains: abstract (matrix) reasoning, verbal reasoning, numerical and letter series, and three-dimensional rotation. Each item has exactly one correct answer, so your raw score begins as a simple count of correct responses. We weight the matrix items slightly more heavily, because matrix problems are the cleanest single marker of fluid reasoning and carry the most information about general ability. There is no penalty for a wrong answer and no countdown, so the raw score reflects what you worked out, not how fast you clicked. You also receive a separate result for each of the four domains, which is where you see whether your verbal or your spatial side is doing the heavier lifting.
Turning the raw score into a standard score
A raw count only becomes meaningful once it is compared with how others perform, so we convert it onto familiar standard scales.
- Percentile. Your percentile is the share of people in the reference sample who scored below you. A percentile of 84 means you scored higher than roughly 84% of that sample. Percentiles are the most intuitive way to read a result, because they answer the plain question "where do I sit?".
- z-score. A z-score expresses how many standard deviations your result sits from the average, where z = 0 is exactly average. It is the bridge between your raw count and the IQ scale.
- Deviation IQ. We use the standard metric introduced by David Wechsler, with the population mean fixed at 100 and one standard deviation set at 15 points. On that scale, about 68% of people fall between 85 and 115, and about 95% between 70 and 130.
The conversion itself is one short formula: standard score = 100 + 15 × z, where z = (your raw score − sample mean) / sample standard deviation. A worked example: if your raw score lands one standard deviation above the sample average (z = +1), your scaled result is about 115, which corresponds to roughly the 84th percentile. The four domain results are reported the same way, so each one is readable as a position relative to others rather than a bare tally.
Standard error and the confidence band
No test returns a perfectly exact result, because every test carries a standard error of measurement (SEM). The practical consequence is that your "true" score lies within a band around the number you see, not precisely on it. A result reported as roughly "115, give or take" means the true value most likely sits somewhere in the low-to-high 110s, not exactly at 115. The more reliable a test is, the smaller its SEM and the tighter that band; a short, unproctored screen like ours has a wider band than a two-hour supervised battery, and we would rather say so than imply a false precision. This is exactly why we, and our disclaimer, keep urging you to read the range rather than fixate on a single digit.
How to read your result
The headline number is a relative position, not a measure of how "smart" you are in any absolute sense. A score of 100 is the average, not a poor result, and most people sit between 85 and 115. Above 130 falls in the top 2% or so; below 70 in the bottom 2%; everything between is the broad, ordinary middle of the curve. Just as important as the headline is the four-domain breakdown: it tells you the shape of your reasoning, which is more useful for understanding yourself than the single figure ever is. We explain how the underlying items and weights are built on our methodology page.
What the score does not mean
A score describes performance on a particular set of reasoning tasks and nothing beyond that. It does not measure your worth, your potential, or your future; a high number guarantees no success, and a low one forecloses nothing. It is not a clinical figure and detects no condition, gift, or disorder. It says nothing about creativity, emotional intelligence, conscientiousness, or judgment, none of which a reasoning test can see, and all of which shape a life at least as much as reasoning does. For any official purpose, a result from a licensed professional is what counts, not an online score, a point we make again in our disclaimer & responsible use page.
Keeping norms current
Norms are not fixed. Average measured scores have drifted upward across generations, an effect named after James Flynn, so a reference sample left untouched slowly goes stale and quietly inflates everyone's percentile. We therefore recompute our norms as response data accumulates, so that your percentile is compared against a current sample rather than an aging one. If we ever change the scoring approach in a way that affects results, we will say so. Questions about how your particular result was produced are welcome through contact.