Mecca Bingo readers with a taste for competition tend to reach the same question sooner or later: what does it actually take to join Mensa, and is the number people quote as high as it sounds. The answer is tidier than the folklore around it. Entry rests on one threshold, expressed as a percentile, and every accepted route is simply a different way of demonstrating that you cleared it.
This page covers the entry standard itself, the supervised papers used in Britain, why the qualifying number looks so much larger than the one quoted elsewhere, the route for people who already hold a suitable score, and what the free online quizzes circulating under the Mensa name are and are not. Fees, booking arrangements and session availability are set by the national society and change from time to time, so those are worth checking directly rather than trusting any third-party page.
- The standard is the 98th percentile: roughly one person in fifty.
- British Mensa uses the Cattell III B paper, where that threshold prints as 148.
- The same percentile is 130 on Wechsler scales and 132 on Stanford-Binet.
- Only supervised sittings count; nothing self-administered at home is admissible.
- Prior scores from recognised supervised instruments can be submitted for review.
- Widely shared online quizzes labelled Mensa-style carry no admission value whatsoever.
The Entry Standard in Plain Terms
Mensa was founded in Oxford in 1946 with a deliberately narrow admission rule and has kept it ever since: candidates must reach the 98th percentile on an approved, supervised intelligence assessment. That is the entire requirement. There is no interview, no nomination, no subject knowledge and no separate standard for different backgrounds.
Stating the rule as a percentile rather than a score is the crucial detail. Approved instruments print different numbers for the same position in the distribution, so a fixed score threshold would be meaningless across the accepted list. One in fifty adults meets the standard, which makes it selective without being rarefied; most people know several qualifying individuals without knowing it.
Why the British Number Is 148
The figure that circulates in Britain is 148, and it startles people who have read elsewhere that 130 is the mark. Both are right. The Cattell III B paper used by British Mensa is built on a scale whose standard deviation is 24, while Wechsler instruments use 15 and the Stanford-Binet fifth edition uses 16.
Two standard deviations above a mean of 100 is the 98th percentile on every one of them. With a spread of 15 that lands on 130. With 16 it lands on 132. With 24 it lands on 148. Nobody scoring 148 on Cattell has outreached someone scoring 130 on Wechsler; the two are the same performance in different units. Anyone comparing figures from different sources without checking the scale is almost certainly comparing nothing at all, which is the central point of the material on how IQ test scores are scaled.
| Instrument | Standard Deviation | Qualifying Score | Percentile | Notes |
|---|---|---|---|---|
| Cattell III B | 24 | 148 | 98th | Supervised paper used in Britain |
| Cattell Culture Fair III A | 24 | 132 | 98th | Non-verbal, sat alongside the above |
| Wechsler adult scales | 15 | 130 | 98th | Common in prior evidence submissions |
| Stanford-Binet 5 | 16 | 132 | 98th | Wide age range, reaches both tails |
| Raven's Advanced Matrices | Percentile reported | 98th percentile | 98th | Reported directly as a percentile |
The Supervised Sitting
The standard route is to book a supervised session and sit the society's papers under invigilation. Sessions run at venues around the country on published dates, with a modest fee, and are administered by trained supervisors. Candidates normally need to be over a minimum age for the adult sitting, with separate arrangements for younger applicants.
The Two Papers
British sittings have long used two instruments in one session. Cattell III B is language-loaded and draws heavily on vocabulary, verbal reasoning and comprehension alongside its abstract items. Cattell Culture Fair III A is deliberately non-verbal, built from matrices, series and classification problems that owe nothing to English usage. Candidates clear the standard by reaching the 98th percentile on either paper; the better of the two results is the one that counts.
That structure is fairer than a single paper would be. A candidate whose first language is not English, or whose schooling was interrupted, can still demonstrate reasoning ability on the culture-fair paper without being penalised by vocabulary depth. The two papers are made of exactly the material described among the standard IQ test questions, with the same rule families appearing under strict timing.
On the Day
Sessions are short and businesslike. Identification is checked, instructions are read aloud, practice items are worked through together, and then each paper runs to a hard time limit. Results arrive by post or email some weeks later rather than on the day, and successful candidates receive an invitation to join. Unsuccessful ones receive their percentile without further comment.
Retakes are restricted. The society limits candidates to a single supervised attempt in most circumstances, precisely because repeated exposure to the same item bank inflates performance without reflecting any change in reasoning. Treating the sitting as a one-off is therefore the correct mental model, and it argues for arriving rested rather than arriving having crammed.
The Prior Evidence Route
Candidates who already hold a qualifying result from a recognised supervised instrument can submit it for evaluation instead of sitting the papers. Mensa maintains a long list of accepted assessments, and documentation is reviewed for administration date, edition and administering professional. An evaluation fee usually applies.
Two details catch people out. The score must come from a supervised administration by a qualified administrator, so an unproctored version of an otherwise accepted instrument does not count. And the edition and standardisation year matter, because norms drift upward across generations and a score produced against decades-old norms overstates current standing. Scores obtained in childhood are often accepted, though the review is more careful the younger the candidate was.
Online Quizzes Wearing the Name
Several free tests circulate under Mensa branding. The best known is the Norwegian society's public matrix test, which runs thirty-five pattern items against a twenty-five minute limit and returns an estimate immediately. It is a well-made puzzle set and a fair taste of the format, and it has no admission value at all, a point the society itself makes plainly.
Other pages borrow the name with far less justification. Anything self-administered, untimed in practice, or repeatable is a self-assessment tool regardless of its branding. Used as a warm-up before a supervised sitting they are genuinely useful, and one honest run through a timed iq test will tell you far more about your readiness than a dozen casual attempts at the same quiz. What none of them can do is produce evidence anybody will accept.
Preparing Sensibly
Preparation has a real but bounded effect. Meeting the item families beforehand removes the cost of decoding an unfamiliar layout under time pressure, which is worth a few points to almost anyone. Practising timing discipline, so that no single hard item eats four minutes, is worth more. Sleep, daylight and a session booked at a sensible hour matter more than either.
What does not work is grinding item banks for months in the hope of manufacturing a qualifying score. Fluid reasoning is not a topic that can be revised, gains from drilling attach to the specific items drilled, and the single-attempt rule leaves no room to convert practice into a second bite. Understanding the broader landscape of home papers, including how a free IQ test online is normed and where its estimates go astray, is a better use of the weeks beforehand than repetition.
After the Result
Clearing the standard buys access to a society: local groups, special interest groups covering everything from cryptography to hillwalking, publications, and national gatherings. That is the whole proposition, and members who enjoy it are the ones who wanted the community rather than the certificate. It confers no professional advantage and no standing that anybody outside the society is obliged to care about.
Falling short is worth keeping in proportion too. The threshold is arbitrary in the sense that any threshold is, it is drawn on one occasion under one set of conditions, and it sits inside a measurement error band several points wide. Someone missing by two points is statistically indistinguishable from someone clearing by two. Families sometimes ask whether children should be tested with membership in mind, which shifts the question onto quite different ground, covered on the page about an IQ test for kids.
