Psychology › Section G · Thinking and intelligence › Topic G.5
Measures of intelligence
An IQ score is a single number, but behind it lie more than a century of test design, statistics and some of psychology's darkest history. This topic explains how intelligence tests are built and scored, what a score does and does not mean, and how testing has been used and misused.
A tool with two histories
Intelligence tests help identify children who need extra support at school, assess the effects of brain injury, and inform some legal decisions. They also have a troubling past: early in the twentieth century they were used to justify immigration restrictions and forced sterilisation. Throughout, remember one rule: only trained professionals give and interpret these tests, and no single score is ever enough to diagnose anyone.
Galton and Binet
In the 1880s Francis Galton, a cousin of Charles Darwin, measured thousands of visitors to his London laboratory on reaction time, grip strength and sensory sharpness, expecting these to reveal mental ability. They did not predict school success well. Galton also coined the word "eugenics" in 1883 for his idea of improving the population through selective breeding. The first practical test came from Alfred Binet and Théodore Simon in France in 1905. Asked by the French government to identify children needing special help at school, they built reasoning, memory and comprehension tasks, graded by the age at which typical children passed them. Binet warned that scores should guide help, not fix a child's future.
From mental age to IQ
Binet's tests gave a child a mental age, the age level of the tasks the child could pass. The German psychologist William Stern suggested dividing mental age by chronological age, and Lewis Terman at Stanford University multiplied by 100 in his 1916 Stanford-Binet test, giving the intelligence quotient:
A child performing exactly at age level scores 100. Terman also standardised the test on American children. A modern version, the Stanford-Binet fifth edition, is still in use, though it no longer calculates IQ this way.
Worked example: the ratio IQ and its flaw
1. An 8-year-old passes the tasks typical of 10-year-olds. Ratio IQ = 10 ÷ 8 × 100 = 125. 2. A 12-year-old performs at the level of a typical 9-year-old. Ratio IQ = 9 ÷ 12 × 100 = 75. 3. Now try an adult. Performance on these tasks stops rising in the late teens, so dividing by an age of 40 would make almost every adult's IQ fall year by year, which is nonsense. 4. Conclusion: the ratio works only while abilities grow steadily with age, which is why modern tests compare a person with others of the same age instead.
Wechsler and the deviation IQ
David Wechsler published the Wechsler-Bellevue scale in 1939 and introduced the deviation IQ, which compares a person with others of the same age, scaled so each age group averages 100 with a standard deviation of 15. His tests now come in versions for adults (the WAIS, whose fifth edition appeared in 2024), children aged 6 to 16 (the WISC-V) and young children (the WPPSI-IV). The WISC-V groups its subtests into five index scores, covering verbal comprehension, visual-spatial skill, fluid reasoning, working memory and processing speed, and an overall Full Scale IQ, mirroring the hierarchical models in the previous topic.
What makes a test trustworthy
A test must be standardised: given and scored the same way every time, with norms from a large, representative sample. It must also be reliable and valid, ideas covered in the research methods section. Intelligence tests are highly reliable and predict school grades moderately well, though they leave out motivation and creativity. Norms also go out of date. James Flynn documented that average raw scores rose by about 3 IQ points per decade in many countries during the twentieth century, the Flynn effect, so tests must be re-normed regularly. Since the 1990s the rise has stalled or reversed in some countries, including Norway and Denmark.
The bell curve
Deviation IQ scores follow an approximately normal distribution, the bell curve. With a mean of 100 and a standard deviation of 15, about 68% of people score between 85 and 115, within one standard deviation of the mean, and about 95% score between 70 and 130, within two. That leaves about 2% below 70 and about 2% above 130. Every score also carries measurement error, so psychologists report a range, such as 104 to 112, rather than treating one number as exact. A difference of a few points between two people usually means very little.
Worked example: reading a score
Use a mean of 100 and a standard deviation of 15. 1. A score of 115 is one standard deviation above the mean. Half the population is below 100 and another 34% lies between 100 and 115, so about 84% score lower: roughly the 84th percentile. 2. A score of 85 is one standard deviation below: about 16% score lower. 3. A score of 130 is two standard deviations above: about 98% score lower. 4. Check: the percentiles are symmetric, since 84 and 16 add to 100, as the bell curve requires.
Testing and the eugenics movement
In the United States, Henry Goddard translated Binet's test and from 1913 used it on newly arrived immigrants at Ellis Island, reporting high rates of "feeble-mindedness" in small, selected groups tested through interpreters while tired and frightened. During the First World War Robert Yerkes led the Army Alpha and Beta tests of about 1.75 million recruits, and results were later cited by those campaigning to restrict immigration, as the Immigration Act of 1924 did. In Buck v. Bell (1927) the Supreme Court upheld a Virginia law allowing the sterilisation of people judged unfit; an estimated 60,000 or more Americans were sterilised under such laws. Historians now regard these uses as scientifically unsound and deeply harmful.
The ends of the curve
A low score alone never defines intellectual disability (intellectual developmental disorder in DSM-5-TR). Clinicians also require significant limitations in adaptive behaviour, the everyday conceptual, social and practical skills of living, beginning in the developmental period; current manuals rate severity by adaptive functioning rather than by IQ. In Atkins v. Virginia (2002) the Supreme Court barred executing people with intellectual disability, and in Hall v. Florida (2014) it ruled that a strict cut-off of 70, ignoring measurement error, was unconstitutional. At the other end, Terman's study of about 1,500 high-scoring Californian children from 1921 found them generally healthy and successful, though his sample, chosen partly by teacher nomination, was mostly white and middle-class.
Common mistakes to avoid
First, using the ratio formula for adults: modern tests use deviation IQ, comparing people with their own age group. Second, treating a score as exact: it carries measurement error and is best read as a range. Third, assuming a score below 70 by itself means intellectual disability: adaptive functioning and early onset are also required. Fourth, forgetting that norms age, as the Flynn effect showed. Fifth, presenting early testing as neutral science: it was used to support eugenic laws. Sixth, trying to judge anyone's ability from an online quiz: only standardised tests given by trained professionals are meaningful.
Key takeaways
Galton's sensory tests failed to predict school success; Binet and Simon's 1905 test aimed to identify children needing help. Terman's 1916 Stanford-Binet used ratio IQ, mental age divided by chronological age times 100, which fails for adults. Wechsler's deviation IQ sets each age group's mean at 100 and standard deviation at 15: about 68% score 85-115 and 95% score 70-130. Good tests are standardised, reliable and valid, and need re-norming because of the Flynn effect. Testing was misused to support eugenics, as in Buck v. Bell (1927). Intellectual disability requires limits in adaptive behaviour, not just a low score.
Flashcards
Flashcard
Measured reaction time and senses in the 1880s; coined "eugenics"
Test your understanding
Question 1 of 5
Using the ratio formula, what IQ does a 10-year-old with a mental age of 12 have?
Question 2 of 5
Why do modern tests use the deviation IQ instead of the ratio IQ?
Question 3 of 5
Scores are set to a mean of 100 and a standard deviation of 15. Roughly what percentage of people score between 85 and 115?
Question 4 of 5
Which statement about intellectual disability matches current professional practice?
Question 5 of 5
What does the Flynn effect imply for intelligence tests?
