Cattell Culture Fair Intelligence Test Explained: Cattell IQ and SD 24

A Cattell IQ is a score from Raymond Cattell's Culture Fair Intelligence Test (CFIT, also written CFT or CCFIT) on a scale with a mean of 100 and a standard deviation of 24, not the 15 used by the Wechsler tests. That is why the same number means different things: a Cattell score of 148 is two standard deviations above the mean, the 98th percentile, and equals an IQ of 130 on the standard SD-15 scale. This guide gives the conversion, explains what Mensa and the Triple Nine Society now accept, describes the CFIT scales and the German CFT 20-R, and sets out what the research says about how culture fair the test really is.

What is the Cattell Culture Fair Intelligence Test?

The CFIT is a nonverbal test of fluid reasoning, the ability to solve new problems without leaning on vocabulary or schooling. It grew out of Cattell's theory of fluid (gf) and crystallized (gc) intelligence; our guide to CHC theory covers how that split became the backbone of modern test batteries, and our blog explains fluid versus crystallized intelligence. Cattell's first paper on a test meant to cut cultural influence appeared in 1940; American Mensa's own psychologist dates the assessment as designed in the late 1940s and adapted over time. We could not find the publisher's manual for the English scales, so what follows rests on Mensa, the Triple Nine Society, the German publisher Hogrefe and published studies.

The English scales are commonly described as three: Scale 1 for young children, Scale 2 for roughly ages 8 to 13 and average adults, and Scale 3 for above-average adults, with parallel forms A and B on scales 2 and 3. Scale 2 has four subtests: Series (continue the sequence), Classification (pick the odd figure out), Matrices (complete the grid) and Conditions (a topology task). The items are all pictures. A related instrument, the CFT 20-R, is easier to document and is covered below. For another nonverbal test see Leiter-3 scores explained, and for the best-known matrix test, Raven's matrices.

What is a Cattell IQ, and why does it use a standard deviation of 24?

American Mensa's supervisory psychologist puts it plainly on a Mensa page: the Cattell IIIB and the old form of Raven's Advanced Progressive Matrices have a mean of 100 and a standard deviation of 24, so Mensa requires a score of 148, the 98th percentile. The same page says the Stanford-Binet and many school tests used a mean of 100 and a standard deviation of 16, which made Mensa's line 132 on those, and gives the example that an IQ of 178 on Cattell equals 152 on the Stanford-Binet. We could not find a Cattell or IPAT text that says why 24 was chosen. The conversion is a z-score: divide the distance from 100 by 24 to get the number of standard deviations, then multiply by 15.

• Cattell 100 = IQ 100 (50th percentile)
• Cattell 116 = IQ 110 (75th)
• Cattell 124 = IQ 115 (84th)
• Cattell 132 = IQ 120 (91st)
• Cattell 140 = IQ 125 (95th)
• Cattell 148 = IQ 130 (98th); Mensa's line
• Cattell 156 = IQ 135 (99.0th)
• Cattell 164 = IQ 140 (99.6th)
• Cattell 172 = IQ 145 (99.87th)

Read it the other way and the trap is plain. A Cattell 140 sounds like a genius score and is the 95th percentile, no rarer than an IQ of 125. For the equivalent problem between the SD-16 and SD-15 scales, see our guide to standard deviation 15 versus 16, and use the score converter and percentile calculator for any other value.

What Cattell score do Mensa and the Triple Nine Society accept?

American Mensa's table of accepted tests still lists a Cattell IQ of 148, among tests given by private psychologists, beside 130 for the Stanford-Binet 5 and the Wechsler Full Scale IQ. Its own testing changed, though. In a notice dated 21 February 2022, Mensa's supervisory psychologist announced that American Mensa would discontinue all administration of its Culture Fair test battery on that date, because these nonverbal assessments were, she wrote, no longer considered valid instruments to measure IQ. The score is still accepted from a psychologist; Mensa no longer gives the test itself. See also our guide to the Mensa IQ score requirement.

The Triple Nine Society, which admits the top 1 in 1,000, has changed too. An older route, a combined raw score of 85 across forms A and B, is marked not valid for test dates after 2 January 2015. The live lines are Cattell Scale IIIB 173 and the German CFT 20-R at 146. A 173 on the SD-24 scale is a z-score of 3.04, the 99.88th percentile or about 1 in 850, and about 145.6 on the SD-15 scale; a 146 on the CFT 20-R, if that test uses SD 15, is z = 3.07 or about 1 in 920. The two lines land at nearly the same rarity, which fits SD 24 for the IIIB and SD 15 for the CFT 20-R. TNS also says Mensa UK's Cattell IIIB results top out at 161, which is z = 2.54, about the 99.45th percentile.

What is the CFT 20-R, and how does it differ?

The CFT 20-R (Grundintelligenztest Skala 2, revised) is the German publisher Hogrefe's descendant of Cattell's test. Hogrefe lists it for children and adolescents aged 8;5 to 19;11 and adults aged 20 to 64, in two parallel parts of four subtests each (series, classifications, matrices and topological conclusions), with 56 items in part 1 and 45 in part 2. The full test takes about 50 minutes with the introduction and the short form, part 1, about 35 to 40. Hogrefe reports norms for more than 4,300 young people and 5,858 adults for part 1, an internal consistency of .95, retest correlations of .80 to .82 over three months, and correlations of .57 to .73 with other intelligence tests. It reports IQ, T-values and percentile ranks, and it is supplied in English, Danish, Croatian, Dutch, Norwegian, Polish, Swedish, Spanish and Czech as well as German, to buyers who can show a qualification. Hogrefe does not state the IQ standard deviation on the page we read; SD 15 is probable but not confirmed.

An older version of this page called the CFT 20-R a German-language revision for ages 8 to 17. That was wrong on both counts: the items are figural and the age range runs to 19, with adult norms.

Does the CFIT really remove cultural bias?

No test does, and the best-known study of the CFIT found items that did not. Nenty (1986), a paper to the American Educational Research Association, gave Scale 2 Form A to 600 American, 231 Indian and 603 Nigerian students, mainly aged 13 to 15, and used four bias-detection methods. It reported that 59% of the items, 27 of 46, tended to be biased, and concluded the test still showed cultural bias; it also found that the claim of measuring a single second-order fluid intelligence factor was supported, and said the test may not have universal validity. That is a conference paper on one scale, not a peer-reviewed journal article.

Wordless items do not mean wordless testing. In a 2025 PLOS ONE study of Scale 3 Form A with 53 undergraduates in Russia (HSE University) and 53 in the United Arab Emirates (American University of Sharjah), the Russian group took the test in a back-translated Russian version and the UAE group in English. A 2025 study of Scale 3 forms A and B with 1,240 students in Misurata, Libya, reported weak discriminatory power for many items in both forms and built a modified version. On the g factor, a study of 500 Dutch seamen found that the Cattell Culture Fair Test, which consists of just four matrix-reasoning tasks of similar methodology, was the exception to g correlations of at least .95 among five batteries. And fluid tests are the ones most affected by the Flynn effect: a meta-analysis of 271 samples (almost 4 million people, 1909 to 2013) found gains of 0.41 IQ points a year for fluid tests, about 4 points a decade, which is why norms expire. That figure is not specific to the CFIT.

Our look at whether IQ is just pattern recognition covers what figural items measure, and our guide to second language and IQ test scores covers why language-light tests are used with people still learning a language. We found no independently published correlation between the Cattell and Raven tests for this guide, although a 2025 Libya study reports comparing a modified CFIT with Raven's matrices.

Is the CFIT still used?

In research, yes: the two 2025 studies above used it, and Hogrefe revises the CFT 20-R (the second edition is from 2019). In admission testing the picture is narrower. Mensa stopped giving its Culture Fair battery in 2022 and the Triple Nine Society retired its raw-score route in 2015, so a Cattell score now usually means one obtained from a psychologist or a Hogrefe-licensed examiner. For a score you can compare on the standard scale, take the IQ Metrics test.

The CFIT set out to measure reasoning on a level field, and it is mostly nonverbal even where it falls short of erasing cultural influence. The IQ Metrics test reports a score on the familiar mean-100, standard-deviation-15 scale with its percentile shown beside it.

Common questions