Practice IQ test in the Mensa admission format

The admission criterion is a percentile rather than a score: 98th percentile of the general population, per Mensa International's published admission criterion — The criterion is a percentile, not a score. The score that reaches it differs by test.. Everything confusing about the familiar numbers follows from that one fact. A percentile has to be expressed on some scale before it becomes a number, and the scales in circulation do not share a standard deviation — so one rank comes out three ways:

National organizations round those and publish the rounded figures: American Mensa states 132 on the Stanford-Binet and 130 on the Wechsler scales, British Mensa states 148 on the Cattell III B. Nobody is disagreeing with anybody — 130 and 148 are one rank described in two units, and a person holding both scores holds one result. This is also why an IQ figure quoted with no scale beside it cannot be checked, and why the second question to ask about any such number is when the test was normed. Test performance drifts upward against fixed norms at ≈ 3 IQ points — Flynn (1987), Massive IQ gains in 14 nations: What IQ tests really measure, Psychological Bulletin — and the drift is largest on exactly the abstract matrix items that free online tests copy, so an undated norm table quietly inflates everyone it scores.

The items on this page are generated rather than reproduced, and the reason is the same reason supervision exists: a published matrix item that has escaped into circulation no longer measures anything, because whoever has seen it is remembering rather than reasoning. So the generator works from the construction rules the literature describes instead — a value held constant along a row, a value stepping down a column, a value distributed so that each of three appears exactly once in every row and once in every column. Two attributes here never share a rule, which is what makes the nine cells nine distinct drawings; an item where two cells coincide usually has a second answer somebody can argue for, and that is a scoring fault rather than a hard item. What is deliberately missing is the fourth format the culture-fair batteries use — the topology items, where a dot has to be placeable inside one shape and outside another — because a generated version of those is only unambiguous when the geometry is checked by hand, and an item with two defensible answers is worse than an item that is not there.

A supervised session looks nothing like this page, and it is worth knowing the shape of it before you sit one. A proctor administers approved instruments to a room or to one person, with photo identification checked, strict timing on each separately timed section, no calculator, no phone and no going back to a section that has closed; the score goes from the tester to the organization rather than to you, and you are told only whether the qualifying rank was reached. The second route in is documented prior evidence — a score you already have from an approved, supervised administration of an approved test, sent in by the institution that holds it. The list of approved tests and the accepted dates are published by each national organization and they differ, so the only reliable version of that list is the current one on the site of the organization you are applying to. What is not a route, anywhere, is a screenshot of a browser result.

If the thing you actually want to know is narrower than a global figure, the neighbors measure it more directly and none of them takes twelve minutes. Reasoning of the strict kind — does this conclusion follow from these premises, or does it merely happen to be true — is the deductive reasoning test, and the number half of an admission battery is the math aptitude test, with the quick math test for the same arithmetic under a tighter clock. If what cost you items above was the pace rather than the reasoning, that is a separate thing and the processing speed test reports it on its own.

A reaction time here is the interval between the frame that painted the stimulus and the timestamp the browser attached to your key, both read from the same monotonic clock. What neither can see is the display pipeline behind it, so on a 60 Hz screen roughly 16 ms of every figure below is the machine rather than you. That is the timing floor: two numbers closer together than that are the same number, and this page reports no precision it cannot support.

That floor is why the only millisecond figure on this page is a rule rather than a result: an answer arriving within 150 ms of an item appearing is discarded, because light has to reach the retina and a signal has to reach the finger before any choice is possible. Everything the run reports back is a span of seconds — a reasoning item takes tens of them — and a quantity that large has no business being printed to the millisecond.

Not affiliated with, endorsed by, or connected to the owner of Mensa. Mensa is a trademark of its owner and is used here only to name the assessment this page prepares for. The questions on this page are our own: no part of the published test is reproduced, and a score here is not comparable to a score from the real instrument.

This is a measurement exercise, not a clinical assessment. It reports what you did on this page against a stated reference and nothing more — it cannot establish giftedness, a learning difficulty, or eligibility for anything. Only a qualified professional, working with more than a browser, can make that judgment.