IQ explained · 6 min read
Why a free online IQ test cannot give you a real IQ
A real IQ score needs a norm sample, controlled administration, calibrated items and a reliability estimate. A browser quiz has none, yet sites print a number anyway.
In short
A real IQ score is a position in a norm sample, obtained under controlled conditions on calibrated items with a known reliability. A browser quiz has none of these: it does not know who else took it or under what conditions, its items were not calibrated on a representative sample, and it cannot state a confidence interval. The number some sites print is a formula applied to your raw score, often tuned to flatter. What a browser test can report honestly is raw accuracy by item type.
What a standardized score needs
- A norm sample: a large group, sampled to represent a population by age and other characteristics, whose scores define where 100 and 15 sit.
- Controlled administration: the same instructions, timing, materials and environment for everyone, usually with a trained administrator.
- Calibrated items: each item's difficulty and discrimination measured on the sample, so that the test is neither too easy nor too hard and the items measure one underlying ability.
- A reliability estimate: evidence that the score would come out similar on another day or a parallel form, expressed as a standard error of measurement.
Professional standards for tests describe these requirements in detail. Without them, a number labelled "IQ" is not comparable with a number from a real test, even if both are near 100.
How a website produces an IQ anyway
A site can take your count of correct answers and map it onto a bell-curve-shaped scale by formula: so many correct equals 100, each extra item adds a few points. The mapping is a guess about how a population would perform, made without a population. Some sites tune it so that most visitors receive scores above 110, because a flattering result is shared more often. Others charge for the number after the free test, which tells you what the number is for.
Speed adds another distortion. Timing over the internet varies with the device, the connection and whether the page was already loaded, so a timed browser score mixes reasoning with hardware.
What a browser test can honestly report
It can report how many items of each kind you solved, with the items, your answers and the correct answers shown. It can tell you which rule families you recognised and which you missed, which is exactly what practice needs. It can also be a fair introduction to the item formats that real tests use, so that the formats are not new on the day a real test matters.
That is what the tests under this hub do. The free IQ-style test reports raw accuracy by domain and tier with every answer reviewed, and no IQ.
When you need a real score
If a decision depends on an IQ score, the only route is a professionally administered test with a written report, confidence intervals and subscale results. Educational psychologists, clinical psychologists and, for admission purposes, some high-IQ societies administer or accept such tests. A browser result is not evidence in that setting, and no site should tell you otherwise.
Sources
- The Standards for Educational and Psychological Testing (AERA, APA and NCME, 2014) set the professional requirements for test norms, reliability, validity and fairness.
- Mensa Norway's own online test page says the test gives an indication of general cognitive abilities and is not a substitute for professional intelligence tests.
- The American Psychological Association's topic page on intelligence collects its public material on how intelligence is defined and measured.
Text on this page is original to MyTestAtlas. It explains published standards and definitions; it is not psychological, medical or admissions advice.