7 min read · Updated 2026-06-01
How IQ tests are scored: from raw answers to a 3-digit number
The path from your raw answers to an IQ score involves a norming sample, a bell curve, and a bit of statistics. Here's what happens step by step.
You finish the test, click submit, and a three-digit number appears. Where did it come from? Here's the actual pipeline.
Step 1 — Count raw correct answers
Every question is scored as right or wrong. Some tests weight harder items more than easier ones, but at the simplest level you get a raw score: the number you got right.
Step 2 — Compare to the norming sample
The test publisher previously ran the same items on a large reference group — the norming sample. From that group they know: what's the mean raw score? What's the standard deviation? Your raw score is compared to that distribution.
Step 3 — Convert to a z-score
A z-score tells you how many standard deviations you are from the mean. If the norming sample averaged 20/30 correct with an SD of 4, and you got 24, your z-score is +1.0 — one standard deviation above average.
Step 4 — Rescale to IQ points
The z-score is then transformed onto the IQ scale, where the mean is 100 and the standard deviation is 15. The formula is simple: IQ = 100 + 15 × z. A z of +1.0 becomes an IQ of 115. A z of −0.5 becomes an IQ of 92.5, rounded to 93.
Step 5 — Add a confidence interval
Good tests report a confidence interval, not just a single number. Typical measurement error is around ±5 points at the 95% confidence level. A reported IQ of 115 really means "we're 95% sure your true score is between 110 and 120."
What about per-domain scores?
The same statistical process runs on each domain independently. Your verbal, numerical, spatial, and logical sub-scores each get a raw count, a comparison to the norming sample, a z-score, and a rescale. The composite IQ is a weighted average of the four.
Why online tests vary so much
The norming sample matters more than most people realise. A test normed against "visitors to a random website" gives different scores than one normed against "a representative sample of the general adult population." If you take three online IQ tests and get three different numbers, most of the difference comes from the norming samples, not from you.
Frequently asked questions
- Why is my score different on different online tests?
- Mainly because the tests use different norming samples. A test normed on a self-selected online audience will give different results than one normed on the general population.
- What is a z-score?
- A z-score measures how many standard deviations a value is from the mean. A z-score of 0 is exactly average; +1 is one standard deviation above the mean.
- How accurate is a single IQ test?
- Even well-designed tests have measurement error of roughly ±5 IQ points at the 95% confidence level. Shorter online tests have wider error bands.