A high-IQ society is an organization that limits its membership to people who have attained a specified score on an IQ test, usually in the top two percent of the population (98th percentile) or above.[1][2] These may also be referred to as genius societies.[1][3] The largest and oldest such society is Mensa International, which was founded by Roland Berrill and Lancelot Ware in 1946.[4][3]
High-IQ societies typically accept a variety of IQ tests for membership eligibility; these include WAIS, Stanford-Binet, and Raven's Advanced Progressive Matrices, amongst many others deemed to sufficiently measure or correlate with intelligence. Tests deemed to insufficiently correlate with intelligence (e.g. post-1994 SAT, in the case of Mensa and Intertel) are not accepted for admission.[5][6][7] As IQ significantly above 146 SD15 (approximately three-sigma) cannot be reliably measured with accuracy due to sub-test limitations and insufficient norming, IQ societies with cutoffs significantly higher than four-sigma should be considered dubious.[8][9][10]
Some societies accept the results of standardized tests taken elsewhere. Those are listed below by selectivity percentile (assuming the now-standard definition of IQ as a standard score with a median of 100 and a standard deviation of 15 IQ points). Since the 1960s, Mensa has experienced increasing competition in attracting high-IQ individuals, as various new groups have emerged with even stricter and more exclusive admissions requirements.[11] Notable high-IQ societies include:
Name | Established | No. of members | Approx. no. of countries | Eligibility / Rarity | Approx. IQ |
---|---|---|---|---|---|
Mensa International | 1946 | ≈ 145,000 (as of 2022[update])[12] | 100 | Top 2 percent of population (98th percentile; 1 person out of 50) | 130 |
Intertel | 1966 | ≥ 1,700 (as of July 2024[update])[13] | 40 | Top 1 percent (99th percentile; 1 out of 100) | 135 |
Triple Nine Society | 1978 | ≈ 1,900 (as of September 2022[update])[14] | 46 | Top 0.1 percent (99.9th percentile; 1 out of 1,000) | 146 |
norm tables that provide you with such extreme values are constructed on the basis of random extrapolation and smoothing but not on the basis of empirical data of representative samples.
[Curve-fitting] is just one of the reasons to be suspicious of reported IQ scores much higher than 160