People often assume intelligence tests were invented to find gifted people. The opposite is true.

1905: Binet, and a commission from the state school system

The French ministry of education gave Alfred Binet a narrow brief: identify children who were not coping in ordinary classes, so that adapted teaching could be offered. Until then the sorting was done on a teacher’s judgement, with the biases one would expect.

Binet and his collaborator Théodore Simon built an empirical scale: a series of tasks of increasing difficulty, calibrated against the age at which an average child succeeds. A child was placed by the “mental age” matching their performance.

Binet was explicit about the limits. He refused to treat it as the measurement of a fixed capacity, insisted that intelligence develops, and warned against using his scale to classify people permanently. That warning was comprehensively ignored.

1912: the “quotient” appears

The German psychologist William Stern proposed dividing mental age by actual age and multiplying by 100. A ten-year-old with a mental age of twelve scores 120.

That is where the word “quotient” comes from. Modern tests compute no quotient at all — they locate a person in a distribution — but the term stuck.

1916: the American translation, and a shift

Lewis Terman at Stanford adapted Binet’s scale for the United States: the Stanford-Binet. The instrument spread widely, but the intellectual framing changed. Where Binet saw a screening tool in the service of teaching, a part of American psychology at the time saw the measurement of an innate, heritable capacity.

That shift had grave consequences. Test results were invoked in the following decades to justify restrictive immigration policy and eugenic programmes in several countries. It is a part of this field’s history that would be dishonest to omit when you run a testing site.

1917: industrial scale

America’s entry into the war created an unprecedented need: assign nearly two million men quickly. The army commissioned group-administered tests — Army Alpha for English readers, Army Beta, non-verbal, for everyone else.

This is the birth of mass testing, with its two durable features: the multiple-choice form and the clock. The biases were enormous — Alpha items assumed familiarity with contemporary American culture — but the industrial model took hold.

1938: Raven, and the ambition of “culture-free”

John C. Raven published his Progressive Matrices. The idea was radical: grids of abstract figures, not a word of text, no apparent cultural reference, where the task is to find the missing figure.

The ambition was to measure reasoning itself, independent of language and education. The matrices became the reference task for what would be called fluid intelligence, and the template for most online tests today.

In hindsight the ambition is only partly met: it is precisely the matrices that rose most across the century, which shows they do depend on something cultural. But they remain far less culturally loaded than verbal items.

1939: Wechsler, and the end of the single score

David Wechsler in New York published an adult scale that broke with the single-number logic. His test produced several indices — verbal and non-verbal at first, more in later editions.

That was the major innovation: what interests a clinician is not the overall score so much as the profile, and particularly the gaps between indices. A wide gap between verbal comprehension and processing speed says something an average would erase.

The Wechsler scales — WAIS for adults, WISC for children — remain the most widely used clinical instruments in the world.

Where g fits in

In parallel, from 1904, the British psychologist Charles Spearman observed a robust statistical fact: performance on very different tasks is positively correlated. People who do well on number series tend to do well on verbal analogies.

Spearman inferred a general factor, written g. It is not a substance located in the brain: it is a statistical regularity. Its reality as a regularity is not disputed; what it represents is disputed a great deal, and the argument is not closed.

Raymond Cattell later proposed a distinction that has lasted: fluid intelligence, the ability to reason about the new, which declines with age, and crystallised intelligence, accumulated knowledge, which holds or grows.

What the history teaches

Tests were designed for specific uses — placing a child, assigning a soldier — and work reasonably inside those frames. They have aged badly every time they were taken outside their intended use.

Their founding author warned against the interpretation that was placed on them. Binet was right, and he was very thoroughly ignored.

And a measuring instrument is never neutral: it carries the questions of the era that produced it. That is a reason to read a score as an indication and never as a verdict — which our method page sets out for our own test.