Myths and famous claims
How historical IQ estimates were made
What is documented
Most famous historical IQ figures trace to one 1926 study. Catharine Cox estimated IQs for 301 eminent people who had already died, rating each from biographical records instead of a test. She warned that her estimates were not exact scores for any individual, and they should not be read as test results.
Want your own number? The test is 25 visual puzzles, about 12 minutes, no sign-up.
The study behind the famous numbers
In 1926 Catharine Cox published the second volume of Lewis Terman's Genetic Studies of Genius. It was the one volume that did not follow living gifted children. Instead it was a retrospective study of 301 eminent historical figures whose IQs were estimated by historiometric methods (Simonton 2020). Everyone in the sample had already died, many of them centuries earlier, and their average birth year was 1705. Since none could be tested, Cox estimated their scores from the record.
How the estimates were made
- Choosing people. Cox picked subjects from James McKeen Cattell's 1903 ranking of the 1,000 most eminent people, which ranked people by how much space reference works gave them. She kept those born between 1450 and 1850 whose fame was earned rather than inherited, which ruled out royalty (Simonton 2020).
- Gathering records. The work drew on more than 3,000 biographical sources in several languages.
- Rating. Three raters, Cox, Terman and Maud Merrill, gave each person one estimate from childhood records and one from records for ages 17 to 26.
- Correcting. When records were thin, raters pulled estimates toward 100, so Cox treated the raw figures as too low (Cox 1926). She produced corrected versions of both, for four estimates per person (Simonton 2020).
Examples from the book
| Person (raw Cox 1926 estimates) | Childhood | Ages 17 to 26 |
|---|---|---|
| Isaac Newton | 130 | 170 |
| Leonardo da Vinci | 135 | 150 |
| Charles Darwin | 135 | 140 |
| Wolfgang Amadeus Mozart | 150 | 155 |
| Johann Wolfgang Goethe | 185 | 200 |
These are the raw values printed in the book (Cox 1926). The raters did not always agree: for young Mozart they gave 125, 190 and 140, and Cox noted how hard it is to rate a musical prodigy. For the group, Cox wrote that the records alone supported an average of 135 to 145, and that after correction the true average was no lower than 155 to 165. A 1978 reanalysis that averaged her corrected estimates found a group mean of 158.9 (Walberg, Rasher & Hase 1978).
Try one · Logical Reasoning
Which tile completes the second pair?
Tap the option you think is right, or show the answer.
Show the answer
The missing panel shows a small hatched arrow pointing down on its own (option 2).
How to solve it
- Compare the first two panels and name exactly what changed.
- The whole figure turns a half turn.
- The large shape becomes hatched.
- Nothing else changes.
- Apply the same changes to the third panel: a small solid arrow pointing up on its own.
- So the missing panel shows a small hatched arrow pointing down on its own.
That was 1 puzzle. The full test has 25.
What the method can and cannot show
Cox herself warned that her corrected estimates were not meant to give exact scores for any individual (Cox 1926). Dean Keith Simonton, a leading historiometry researcher, adds three cautions (Simonton 2020):
- Raters judged each person by their own field, so Mozart's estimate rests mostly on his musical development.
- The Stanford-Binet measures general abilities all children are expected to develop, so an estimate from a biography is not the same thing as a test score.
- Terman and Cox assumed a study run backward from fame would match one run forward from childhood, and that may not always hold.
Timing matters too. The first intelligence scales date from 1905 to 1911 (Britannica, Alfred Binet), and the deviation IQ used today arrived in 1939 (Wikipedia: WAIS), after Cox's book. Her figures belong to the era of the ratio IQ.
What Cox concluded
Her estimates correlated only .25 with Cattell's eminence ranking, and .16 after adjusting for how reliable the data were (Simonton 2020). She concluded that high but not the highest intelligence, combined with great persistence, achieves more eminence than the highest intelligence with less persistence (Cox 1926). That is the study's most durable lesson, more than any single number.
See the individual pages for Newton, Leonardo and Darwin, and why there is no estimate for Einstein. For how real tests developed, read the history of IQ tests.
Sources
- Simonton, D. K. (2020). Galton, Terman, Cox: The Distinctive Volume II in Genetic Studies of Genius. Gifted Child Quarterly. doi:10.1177/0016986220921360 doi.org/10.1177/0016986220921360
- Cox, C. M. (1926). Genetic Studies of Genius, Vol. II: The Early Mental Traits of Three Hundred Geniuses. Stanford University Press (scan of original) gwern.net/doc/iq/high/1926-cox-theearlymentaltraitsof300geniuses.pdf
- Walberg, H. J., Rasher, S. P., & Hase, K. (1978). IQ Correlates With High Eminence. Gifted Child Quarterly, 22(2), 196-200. doi:10.1177/001698627802200213 doi.org/10.1177/001698627802200213
- Encyclopaedia Britannica. Alfred Binet (read via Wayback Machine, January 2025) www.britannica.com/biography/Alfred-Binet
- Wechsler Adult Intelligence Scale. Wikipedia. en.wikipedia.org/wiki/Wechsler_Adult_Intelligence_Scale