Average IQ of a Physicist: How Smart Are Physicists?

Where do these numbers come from?

The estimates rest on a handful of studies, most of them decades old, plus modern inference from career selectivity.

Roe's 1952 study of eminent scientists

The most-cited primary source is psychologist Anne Roe's 1952 work, later published as The Making of a Scientist. Roe studied 64 eminent American scientists, including physicists such as Luis Alvarez and J.H. Van Vleck, and gave each a high-ceiling test built for very able adults. Her sample scored far above the general run of PhDs, with a median IQ reported around 137 on one of her tests and mathematical and verbal medians reaching the top fraction of a percent of the population. Crucially, Roe was studying the elite of science, hand-picked for landmark research, so her ~137 median describes standout scientists, not a typical working physicist.

Harmon and later field comparisons

Later work comparing fields, associated with Lindsey Harmon's 1961 analysis of doctorate holders, again placed physical scientists at the top of the ranking, with physics often cited in the neighborhood of 140 for that select doctoral group. Modern "IQ by profession" compilations, which lean on personnel-test data and meta-analyses of cognitive ability and occupational attainment (such as Schmidt and Hunter, 1998), settle on the more conservative ~125-135 band for physicists in general rather than the eminent few.

Why the range is wide

The spread between "125" and "140" is not sloppiness; it reflects which physicists were measured. Study a Nobel-caliber sample and you get numbers near 140. Estimate the whole profession, including every graduate student and lab physicist, and the average pulls down toward 125-130. Both can be true at once.

The Feynman caveat: why one number misses the point

Richard Feynman is the standard counterexample, and it is worth sitting with. He reportedly scored 125 on a school IQ test around age 12 or 13. By the tidy logic of "IQ equals genius," 125 is good but unremarkable for a physicist, roughly the field average, not the outlier you would expect from one of the century's great minds.

Yet the same person went on to top the notoriously brutal Putnam mathematics competition by a wide margin, reportedly posted record scores on Princeton's graduate admissions exams, and won a Nobel Prize. Physicists who knew him described his problem-solving as extraordinary. So what went wrong with the 125?

A few things, and they are instructive for anyone reading their own score:

• It was taken young. Higher-order reasoning keeps developing into early adulthood; a childhood score is a weak predictor of adult capability.
• The test likely leaned verbal. Feynman's genius was overwhelmingly mathematical and physical, exactly the dimension that a verbally weighted school test can undercount.
• Ceiling effects flatten the top. Ordinary tests run out of hard questions before they run out of ability, so they compress differences between the very gifted.

The honest takeaway is not "IQ is meaningless." It is that a single number, especially from a limited or age-inappropriate test, can badly misrepresent an individual even while group averages hold up fine.

The honest note: what these figures do and don't tell you

A profession-average IQ is a describing statistic, not a gatekeeper. It says the typical physicist scores high; it says nothing about whether any specific person can or cannot do physics. Fields with high average IQs still contain enormous internal spread, and traits that IQ tests barely touch, persistence, curiosity, comfort with being wrong for years, and the discipline to work a hard problem to its end, matter enormously for actual scientific output. Roe herself emphasized personality and drive alongside raw ability.

So if you are eyeing physics and worried your own score is "only" 120, remember two things: the profession averages are estimates layered on decades-old, elite-skewed samples, and Feynman's 125 is a permanent reminder that the number and the scientist are not the same. IQ is one lens on cognition, useful in aggregate, unreliable as a verdict on a single life.

Frequently asked questions

Q: What is the average IQ of a physicist?

A: Roughly 125 to 135, one of the highest of any profession. The estimate comes from studies of scientists (Roe, 1952) and modern occupational-IQ compilations. Eminent-physicist samples score higher, near 137-140, but that describes standout researchers, not the typical working physicist.

Q: Was Feynman's IQ really only 125?

A: That is the reported school-test figure, and it almost certainly undersold him. The score was taken in his early teens on a likely verbally weighted test, while his gift was mathematical. He later dominated the Putnam competition and won a Nobel Prize, showing how poorly one early number can capture a mind.

Q: Do physicists have higher IQs than doctors or engineers?

A: On average, slightly, but the fields overlap heavily. Physics and mathematics tend to top the field rankings, with engineering and medicine close behind. The differences between these high-scoring groups are small compared with the variation among individuals within any one of them.

Q: Is a high IQ required to become a physicist?

A: No single cutoff exists, and the averages describe groups, not requirements. High quantitative reasoning helps, but persistence, curiosity, and years of deliberate study drive real physics output. Many capable physicists would score below their field's average, and a score alone predicts little for any individual.

References

• Roe, A. (1952). The Making of a Scientist. Dodd, Mead & Company. Summary and score tables archived at iqcomparisonsite.com.
• Roe, A. (1952). "A Psychologist Examines 64 Eminent Scientists." Scientific American, 187(5). Full text (gwern.net archive).
• Schmidt, F. L., & Hunter, J. E. (1998). "The validity and utility of selection methods in personnel psychology." Psychological Bulletin, 124(2), 262-274.
• "Richard Feynman's IQ Score Was 'Only' 125." Forbes (2016). forbes.com.

Last updated: July 13, 2026