The history of IQ testing

Cattell coins "mental tests"

James McKeen Cattell publishes "Mental tests and measurements" in Mind and proposes giving the same series of tests to large numbers of people, so that results can be compared. The phrase "mental test" sticks.^[1]

Spearman proposes general intelligence, g

Charles Spearman notices that scores on very different school and sensory tests all correlate positively, and argues that one general ability runs through them. He calls it g, and it is still central to intelligence research.^[2]

What the g factor is

Binet and Simon publish the first intelligence scale

To identify schoolchildren who needed extra help, Alfred Binet and Théodore Simon publish 30 tasks arranged in order of difficulty, in L'Année Psychologique. It is the first practical intelligence scale.^[3]

The Binet-Simon scale is grouped by age

The revised scale has 58 tasks arranged by the age at which a typical child passes them, from 3 to 13. A child's result can now be read as a mental level, the idea that becomes "mental age". A further revision follows in 1911.^[4]

Stern proposes the intelligence quotient

The German psychologist William Stern suggests dividing mental age by chronological age, so that a child's standing does not depend on how old they are.^[5]

How IQ is calculated, then and now

Terman publishes the Stanford-Binet

Lewis Terman at Stanford adapts the Binet scale for American children and multiplies Stern's ratio by 100. The result, mental age over actual age times 100, is the original "IQ".^[6]

The Stanford-Binet test today

The US Army tests its recruits

A committee chaired by Robert Yerkes builds two group tests: Army Alpha, written, and Army Beta, a picture test for men who could not read English. Sources put the number of recruits tested at between 1.5 and 2 million. No intelligence test had been given at this scale before, and the tests' language and cultural bias drew criticism at the time and since.^[7]

Terman starts his study of gifted children

Terman begins following high-scoring children. By 1928 the study has 1,528 of them, and it follows them for most of their lives. It is one of the longest-running studies in psychology.^[8]

Buck v. Bell: the Supreme Court backs forced sterilisation

The US Supreme Court upholds Virginia's law allowing the sterilisation of people labelled "feeble-minded". Carrie Buck had been classed that way partly on the strength of an intelligence test. Later evidence showed the label was wrong, and Virginia formally apologised in 2002.^[9]

Eldon Wonderlic writes a 12-minute test

As a graduate student at Northwestern, Eldon F. Wonderlic writes a short cognitive test for hiring. He begins selling it in 1937, and the Wonderlic is still used by employers today.^[10]

The Wonderlic test explained

Raven's Progressive Matrices appear

John C. Raven publishes the first versions of his matrix test: patterns with a missing piece, no words needed. Because no words are needed, the same test can be given across languages.^[11]

How Raven's matrices work

Wechsler publishes the Wechsler-Bellevue

David Wechsler's adult test reports a deviation IQ: a person's standing against people of their own age, set to an average of 100 and a standard deviation of 15. Almost every modern IQ score works this way.^[12]

Why IQ uses a standard deviation of 15

Mensa is founded in Oxford

Roland Berrill and Lancelot Ware found a society for people who score in the top 2% on an approved test. Mensa dates its founding to 1 October 1946.^{[13, 14]}

Mensa qualifying scores by test

The first Wechsler test for children

The Wechsler Intelligence Scale for Children (WISC) brings the deviation IQ to school-age testing. It is now in its fifth edition.^[15]

How to read WISC-V scores

One entrance test for every US service

Following the Selective Service Act of 1948, the services move to a single screening test, the Armed Forces Qualification Test (AFQT). Its name survives as the score inside the ASVAB.^[16]

The WAIS replaces the Wechsler-Bellevue

The Wechsler Adult Intelligence Scale keeps the 11 subtests of the earlier scale with new items, administration and norms. It is revised in 1981, 1997, 2008 and 2024.^[12]

The WAIS test explained

The Stanford-Binet drops the ratio IQ

The 1960 revision of the Stanford-Binet replaces mental age divided by actual age with a deviation IQ, as Wechsler had done. The ratio IQ falls out of use.^[17]

Fluid and crystallised intelligence

John Horn and Raymond Cattell set out the split between fluid reasoning, solving problems you have not met before, and crystallised intelligence, what you have learned.^[18]

Fluid vs crystallised intelligence

The ASVAB begins in high schools

The Armed Services Vocational Aptitude Battery starts as a school testing programme. In 1976 every service adopts it for selection and job assignment.^[16]

The ASVAB explained

Griggs v. Duke Power limits hiring tests

Duke Power required the Wonderlic and a mechanical test for transfers. The Supreme Court rules that a test which screens out Black applicants is unlawful unless it is related to the job, whatever the employer intended.^[19]

Larry P. v. Riles bars IQ tests for special-class placement

A federal court in California rules that IQ tests may not be used to place Black students in classes for the "educable mentally retarded". The ruling is affirmed on appeal in 1984.^[20]

Gardner proposes multiple intelligences

In Frames of Mind, Howard Gardner argues for several separate intelligences, such as musical, bodily and interpersonal, instead of one general ability. It is the best-known challenge to a single g.^[21]

Flynn shows scores have been rising

James Flynn compares American norm samples from 1932 to 1978 and finds a rise of 13.8 IQ points in 46 years. Three years later he reports gains of 5 to 25 points in a generation across 14 nations. The rise becomes known as the Flynn effect.^{[22, 23]}

The Flynn effect, country by country

Carroll's three-stratum model

John Carroll re-analyses 461 data sets and arranges abilities in three layers: g at the top, about eight broad abilities beneath it, and many narrow ones below. Combined with Horn and Cattell's work it becomes the CHC model behind most modern batteries.^[24]

The Bell Curve and the reply

Herrnstein and Murray's book reignites argument over IQ, class and race. In December, 52 researchers sign "Mainstream Science on Intelligence" in The Wall Street Journal; 48 others asked to sign refused. Its claims about genetic group differences are not accepted by mainstream science today.^[25]

The APA task force reports

The American Psychological Association publishes "Intelligence: Knowns and unknowns", a survey by 11 researchers chaired by Ulric Neisser that separates what is established about IQ from what is disputed and what is unknown.^[26]

WAIS-III adds matrix reasoning

The third edition adds Matrix Reasoning, Letter-Number Sequencing and Symbol Search, and introduces index scores for verbal comprehension, working memory, perceptual organisation and processing speed.^[12]

WAIS index scores explained

Atkins v. Virginia

The Supreme Court rules that executing a person with intellectual disability is unconstitutional, and leaves states to decide who qualifies. IQ scores become evidence in capital cases.^[27]

Stanford-Binet 5 and WISC-IV

Both major tests for children and adults get new editions in the same year. The SB5 now uses a standard deviation of 15, like the Wechsler tests.^{[17, 28, 29]}

WAIS-IV

The fourth edition of the adult Wechsler scale is standardised on 2,200 people. It has 10 core subtests and stays the standard adult test until 2024.^{[12, 30, 31]}

Hall v. Florida: no hard cut-off at 70

The Supreme Court strikes down Florida's rule that a score above 70 ends the inquiry into intellectual disability, because every IQ score has a margin of error. The same year the WISC-V is published, normed on 2,200 children.^{[27, 32]}

IQ confidence interval calculator

A century of IQ gains, measured

A meta-analysis of 271 samples and almost 4 million people in 31 countries puts the 20th-century rise at about 3 points per decade, largest on fluid tests and slowing in recent decades.^[33]

The Flynn effect explained

The reversal happens inside families

Using about 730,000 Norwegian conscript scores, Bratsberg and Rogeberg show that scores fell for men born after 1975, and that the fall appears between brothers too. The cause must be environmental, not a change in who has children.^[34]

The NFL drops the Wonderlic

After decades of use, the league stops giving the Wonderlic at its Scouting Combine. Some teams still give it themselves.^[35]

NFL Wonderlic scores

WAIS-5

Pearson publishes the fifth edition, normed in 2023 and 2024 on a sample matched to the US census. Adults who took both scored about 1.9 points higher on the WAIS-IV, a smaller gap than earlier editions showed.^{[36, 37]}