Mensa IQ Test: Entry Thresholds, Formats and Admission Routes
Mensa is a membership society with a single entry condition: a demonstrated score at or above the 98th percentile on an approved standardised intelligence assessment. Everything else about the organisation — its branches, its publications, its gatherings — sits downstream of that one requirement. This page from Megaways Casino covers what the threshold means in practice, the two routes to demonstrating it, and what applicants tend to misunderstand before applying.
The 98th percentile requirement
The criterion is expressed as a percentile rather than as a score, and that is deliberate. Different instruments use different standard deviations, so the same rarity corresponds to different numbers depending on which test produced them.
| Instrument type | Standard deviation | Approx. 98th percentile score |
|---|---|---|
| Wechsler scales (WAIS, WISC) | 15 | 131 |
| Stanford-Binet, 5th edition | 15 | 131 |
| Cattell III B | 24 | 148 |
| Raven's Advanced Progressive Matrices | 15 | 131 |
The Cattell row is the one that causes trouble. Its unusually wide standard deviation of 24 means the qualifying figure of 148 sounds dramatically higher than the Wechsler figure of 131, while representing exactly the same position in the population — roughly one person in fifty. Someone quoting a Cattell score alongside a Wechsler score without stating the scale is comparing quantities that are not comparable. The mechanics behind this are set out on our page about the average IQ score.
The point that gets missed: the threshold is a rarity, not a number. Any score, from any instrument, at any deviation, qualifies if it sits at or above the 98th percentile of that instrument's norms.
Route one: the supervised admission session
Most applicants take a supervised test administered by the national branch. Sessions run at booked venues on published dates, cost a modest administration fee, and take between two and three hours including registration and instructions.
What the session involves
The British branch has historically administered two instruments in a single sitting: the Cattell III B, which is language-based and includes verbal reasoning items, and the Cattell Culture Fair III A, which is non-verbal and built from figural and matrix material. An applicant reaching the threshold on either qualifies. That two-instrument structure is genuinely useful, because it gives people whose strengths lie clearly on one side of the verbal/non-verbal split two independent chances rather than one blended one.
Practical details worth knowing in advance: bring photographic identification, arrive with time to spare because late arrivals are not admitted once instructions have begun, and expect strict timing on each section. Results arrive by post or email within a few weeks. The result is pass or fail against the threshold; some branches release the percentile, others do not.
Retake policy
Supervised admission testing is normally available once. The reasoning is straightforward — practice effects on this material are large, and allowing repeated attempts would mean admitting people on the strength of familiarity rather than reasoning. Anyone unsure whether they are close to the threshold should establish that before booking, not afterwards.
Route two: prior evidence
If you have previously been assessed on an approved instrument by a qualified administrator — an educational psychologist, a clinical psychologist, or an occupational psychologist working within their competence — that documentation can be submitted directly for evaluation. The evidence is checked by an appointed psychologist and either accepted or not.
What is accepted:
- A full report from a recognised standardised battery, on official letterhead, naming the instrument, the edition, the date and the administrator's professional registration.
- Childhood assessments, in many cases, since scores are reasonably stable from around age six onward — though the accepting branch decides, and older reports are scrutinised more closely.
- Certain group-administered tests where administration conditions were properly controlled and documented.
What is not accepted, without exception:
- Any self-administered or browser-based result, regardless of how well constructed the platform is or what score it produced.
- Screening tools designed for other purposes, or abbreviated versions of full batteries.
- Reports without a named, qualified administrator and a verifiable instrument.
This exclusion is absolute and it catches people out. A carefully built web assessment can produce a figure close to a supervised one under good conditions, but no unsupervised score is verifiable, and verification is the entire point of the evidence route. Our page on online IQ test accuracy covers why the gap between web and supervised results is systematic rather than random.
Preparing without distorting the result
There is a real tension here. Familiarity with item formats improves performance, which means anyone who has met progressive matrices before has an advantage over someone meeting them cold — an advantage that has nothing to do with what the test is trying to measure. Removing that inequality by familiarising yourself is reasonable. Drilling until you have memorised the underlying rule set produces a score that measures your preparation.
Preparation that is defensible:
- Meeting each item type once so that reading the instructions does not consume test time.
- Practising working under a clock, since timing pressure is a substantial part of the difficulty and being unused to it costs points that reflect nothing.
- Handling the logistics — sleep, food, arrival time, knowing the venue — so that state effects are minimised.
Preparation that undermines the exercise:
- Working through hundreds of matrix items until the transformation rules are automatic.
- Rehearsing on the specific instrument you will sit, where copies can be found.
- Repeated retakes of the same iq test until the number reaches a target, which trains the item set rather than the capacity.
The distinction is whether you are removing an artefact or manufacturing a score. Taking a well-constructed free test once to see the formats is the former; our page on free IQ test options discusses what those short assessments can and cannot tell you beforehand.
A realistic self-check before booking
Because the supervised route is normally a single attempt, it is worth forming an honest estimate first rather than discovering the answer on the day.
Useful signals: consistent placement in the upper tail across several well-constructed short assessments taken on separate days under proper conditions; a documented prior score from schooling or occupational testing that sat near the threshold; sustained ease with novel abstract problems rather than with familiar ones. Unreliable signals: a single flattering web result; being told as a child that you were bright; performing well on material you have practised extensively.
The threshold admits roughly one person in fifty. That is uncommon but not rare — in a secondary school of a thousand pupils it is about twenty of them. People frequently over-estimate the exclusivity and under-estimate how many ordinary, unremarkable lives sit above the line, which distorts both the decision to apply and the reaction to the outcome.
If the estimate lands close to the boundary, the two-instrument structure of the supervised session is genuinely in your favour — reaching the threshold on either qualifies, and applicants with an uneven verbal/non-verbal profile benefit substantially from that arrangement.
What membership actually offers
It is worth being clear-eyed, since the threshold generates expectations that the society does not set out to meet. Membership provides local and national gatherings, special interest groups covering a wide range of subjects, member publications, and an international network with reciprocal arrangements across national branches.
It does not provide career advancement, and listing membership on a CV has no established effect on hiring outcomes. It does not provide access to opportunities unavailable elsewhere. It does not certify expertise in anything — the threshold measures performance on reasoning tasks, which is a narrow thing.
People who get value from membership generally want the social side: conversations with people who enjoy the same kinds of problems, and gatherings organised around shared interests rather than shared professions. People who join expecting a credential are usually disappointed, and the disappointment is predictable from the entry criterion itself.
Younger applicants
Supervised admission testing has a minimum age, typically around ten and a half years, varying by branch. Below that age, children can be admitted on the basis of a report from an educational psychologist using an age-appropriate instrument such as the WISC-V or the British Ability Scales.
Whether that is a good idea for any particular child is a separate question from whether it is possible. Assessment for a child is worth doing when it answers a question about their education — the reasoning behind that is set out on our page about IQ tests for kids — and worth thinking twice about when the question it answers is really the parents'.