IQ of Nobel Prize Winners: What Was Measured and What Was Guessed

The 2026 Nobel announcements run from 5 to 12 October. So far the Nobel Foundation has named three prizes: Physiology or Medicine on 5 October to Karl Deisseroth, Peter Hegemann and Georg Nagel for optogenetics, Physics on 6 October to Francis Halzen for the IceCube neutrino observatory, and Chemistry on 7 October to Henri B. Kagan and Kenso Soai for non-linear effects and autocatalysis in asymmetric synthesis. Literature, Peace and Economic Sciences follow. Every year the same question trends alongside them: what is the IQ of a Nobel Prize winner? Search results give figures of 150, 160 or more. This guide sets out what was ever measured, what was estimated, and what the best studies of very bright children actually found.

Has anyone measured the IQ of Nobel Prize winners?

Almost no one. There is no registry of laureates' test scores, and laureates are not a sample anyone gave an IQ test to before they won. The evidence comes in three kinds: a handful of reported childhood scores, one 1953 psychological study of eminent scientists, and long-running studies of gifted children that followed them to adulthood. We look at each, because each answers a different question and each has a weakness.

The reported scores are modest. In a 2020 paper in the journal Intelligence, Russell Warne, Ross Larsen and Jonathan Clark cite Shurkin's 2006 biography of William Shockley, who shared the 1956 Nobel Prize in Physics for the transistor, for scores of 129 and then 125 on childhood intelligence tests, and a score at the 90th percentile (about an IQ of 119) as an adult. They add that Richard Feynman's and James Watson's scores were in the 120s, citing Root-Bernstein (2015), a source we could not open. The Feynman figure of 125 comes from James Gleick's biography, which we also could not open; a research summary from the test company Riot IQ notes that the name of the test and the original record "have never surfaced". Treat these as reported scores, not verified ones.

What did the Terman study miss?

Lewis Terman began his Genetic Studies of Genius in the 1920s. His first volume (1925) says an IQ of 140 on the Stanford-Binet "was set as the provisional lower limit" for children under eleven, and that by 1 May 1924 the files held 1,444 subjects. By 1928, according to a Stanford Magazine history, he had 1,528 children, including siblings and some below the cutoff. Nobel Prize winners are famously absent from the group. Stanford's Albert Hastorf is quoted as saying that "none of the Terman kids ever won a Nobel or Pulitzer".

The story that tightened the criticism is that two future Nobel physicists, Shockley and Luis Alvarez (Physics, 1968), were tested for Terman and did not qualify. Warne and colleagues trace the claim to Shockley himself, in a 1980 interview with the journalist Syl Jones, and note that Shurkin repeated it in 1992 after interviewing both men. We searched the full text of Terman's first volume and found neither name, so the claim rests on later accounts. Alvarez's score is not public.

Does missing them discredit IQ testing?

Not on the numbers. Warne, Larsen and Clark ran a simulation of 168,000 schoolchildren, 1,000 runs for each of 330 conditions, selecting the top 1,528 on measured IQ. They set the base rate from the data: among U.S. children born from 1901 to 1916, the rate of winning a Nobel Prize was 1 in 627,728. They assumed the lowest IQ a Nobelist could have was between 1.3 and 1.6 standard deviations above the mean (IQ 119.5 to 124), a range that Shockley's scores support, and a correlation between IQ and winning a Nobel of 0 to .20. Under those conditions Terman had a 53.0% to 83.2% chance of selecting neither future laureate, a 16.2% to 39.7% chance of selecting at least one, and a 0.6% to 7.3% chance of selecting both.

The authors also point out that Terman's instruments were mostly verbal group tests, with no screening for mathematical or spatial ability, the strengths physicists tend to show. Their conclusion is that "predicting high levels of eminence requires measuring a variety of relevant cognitive and non-cognitive variables". The paper was a 2019 preprint that was peer reviewed and published in 2020 in Intelligence, volume 82, article 101488; we read the preprint, whose numbers may differ slightly from the journal text.

What did Anne Roe find in 1953?

The nearest thing to measured data on eminent scientists is Anne Roe's The Making of a Scientist (1953). She studied 64 leading U.S. scientists in physics, biology and the social sciences, interviewing them for hours and giving each a test she called the Verbal-Spatial-Mathematical (VSM) test, "one which had to be especially constructed" because ordinary intelligence tests were too easy for them. They were eminent scientists, not Nobel laureates as a group; the only Nobel winner she mentions is one subject who "has always wanted to be a farmer".

• Verbal test: median score approximately equivalent to an IQ of 166, with the lowest and highest at about 121 and 177
• Spatial test: median equivalent to an IQ of 137, with a highest score near 164; scores fell with age (correlation -.40)
• Mathematical test: lowest about 128, median 154, highest 194

Roe is candid about the weakness. These equivalents "have been arrived at by a series of statistical transformations based on assumptions which are generally valid for this type of material but which have not been specifically checked for these data", and she notes that the verbal test was probably not difficult enough for some of the men. So the three medians are estimates on three different tests, and they differ by nearly 30 points. They also show that "an IQ" is not one number: the same group sits at 137, 154 or 166 depending on which ability is converted. Our guide to where celebrity IQ numbers come from shows how often a single figure like this is repeated without its caveats.

What about the brightest children followed for decades?

The Study of Mathematically Precocious Youth (SMPY) selected children who scored extremely high on the SAT before age 13. In a 2013 paper in Psychological Science, Harrison Kell, David Lubinski and Camilla Benbow surveyed 320 of them in midlife. Of these, 142 (44%) held doctoral degrees, 59 had produced 392 refereed STEM publications, and 117 had made 820 software or patent contributions. According to Warne and colleagues, "none has won a Nobel Prize" among these and other cognitively elite cohorts, although one SMPY member earned a Fields Medal; that statement dates from their 2019 manuscript, and we have not checked for later winners. We cover the same study in does IQ above 120 stop mattering.

So what is the IQ of a Nobel Prize winner?

The defensible answer is a range with a floor, not a number. Reported childhood scores sit in the 120s; Warne and colleagues argue the minimum for a laureate is probably 1.3 to 1.6 standard deviations above the mean, which is an IQ of about 120 to 124; and Roe's converted medians for eminent scientists, built on tests she wrote for them, run from 137 to 166. None of these is an average over laureates. A score of 150 or 160 attached to a named winner is, on this record, a guess. For the arithmetic of how rare high scores are, see how rare is a high IQ score and the ceiling on an IQ test.

What should you check before sharing a laureate IQ list?

• Whether the score was measured by a named test, or estimated from a biography
• Whether the figure is a childhood score, an adult score or a conversion of a test the person never took
• Whether the list gives a source other than another list
• Whether a ceiling or a conversion is doing the work: Roe's tests were built for her sample
• Whether it treats an IQ as the cause of the prize rather than one of several ingredients

Our IQ Metrics test reports on the standard scale, which you can read against the IQ bell curve. For the person behind the label, see the profiles of Albert Einstein and Richard Feynman and the scientists collection, and for what the word means in practice read what is a genius IQ level.

Common questions