IIF Certified Assessment
Start IQ Test
IIF Certified Assessment
Nearly every "IQ by profession" chart online traces to one 1945 study of wartime enlisted men, scored on a scale that is not even the one modern IQ tests use. Here is the real table — and what still holds up.
If you have seen a chart online ranking IQ by job — accountants and lawyers near 128, teachers in the low 120s, unskilled trades down in the 90s or below — you have seen a repackaging of one specific dataset from 1945, almost never credited to its source.
The original source is a 1945 paper by Thomas W. Harrell and Margaret S. Harrell, published in Educational and Psychological Measurement, analyzing Army General Classification Test (AGCT) scores for 18,782 white enlisted men in the Army Air Force Command, broken down by the civilian occupation each man held before enlisting. It is real data, honestly reported — and also a wartime sample of one branch, one sex and one racial group, tested on a specific 1940s military instrument, a limitation the data has carried since the day it was collected. A separate 1947 analysis by N. Stewart, covering AGCT scores of Army personnel more broadly and published in the journal Occupations, found the same occupation-level pattern independently, which is part of why the shape of the ranking is taken seriously even where the exact 1945 figures are not.
The full table runs to more than 70 occupations, each with its own sample size, mean, median and range. A representative slice, drawn from the original paper as reproduced in a widely cited later review:
The AGCT was normed to a mean of 100 with a standard deviation of 20, not the mean-100, standard-deviation-15 scale that modern deviation IQ tests use. That difference matters arithmetically, not just historically. A score of 128 on the AGCT’s SD-20 scale sits 1.4 standard deviations above the mean, which our own percentile calculator places at roughly the 92nd percentile. Treat that same "128" as if it came from a modern SD-15 IQ test instead, and it implies 1.87 standard deviations above the mean — close to the 97th percentile. Same number, two different rarity claims, depending on a scale nobody printed on the chart.
The number on a viral "IQ by profession" chart is usually real. The scale it was measured on almost never survives the trip from a 1945 journal table to a social media post — and the scale is the part that decides what the number actually means.
The specific 1945 numbers should not be quoted as a current fact about any occupation today. What has held up is the broader pattern behind them. A comprehensive 2004 review by Frank Schmidt and John Hunter, published in the Journal of Personality and Social Psychology, reports that occupation-level mean scores on the U.S. Employment Service’s General Aptitude Test Battery correlate between .90 and .95 with independent occupational-prestige ratings, and that general mental ability correlates roughly .65 with individual occupational level attained — rising to about .72 once corrected for measurement error. Data from both world wars, drawn from draftee samples that were unusually representative of the male population at the time, showed the same shape Harrell and Harrell found: scores rising with occupational level, and the spread of scores narrowing as that level rises.
The same battery’s data, drawn from 425 job-performance validity studies covering more than 32,000 workers, shows the ability-to-performance link scales with a job’s own complexity rather than applying at one fixed strength everywhere:
The AGCT was not an isolated instrument. General mental ability tests have been used in hiring since the end of the First World War, and the most widely used one today is a direct descendant of that same testing tradition: the Wonderlic Personnel Test, a 12-minute, 50-item paper-and-pencil test drawing on verbal, quantitative and spatial material in roughly equal measure. Our piece on converting a Wonderlic score to an IQ estimate covers that specific test in detail — the same conversion-scale caution that applies to the 1945 AGCT table applies there too, since the Wonderlic has its own scale, not a shared one with either the AGCT or a modern deviation IQ test.
No comparably large, modern dataset scores a nationally representative population on a single standardized test and records detailed occupation alongside it the way the wartime draft data happened to. Related, narrower pictures do exist on this site: our pieces on whether IQ predicts income and on average IQ by education level cover adjacent modern data using the same kind of large-sample research this occupational question would need. Until an equivalent dataset exists for occupation specifically, the honest version of this chart is the general finding — ability tracks occupational level, reliably and substantially — not a specific score next to a specific job title.
Rather than trust an 80-year-old number, see where your own score actually sits. Take the full IQ assessment or run a score through our percentile calculator for a result anchored to a real, current reference population.
Find your IQ score now! →The next time that chart resurfaces, the two facts worth adding are the ones it always leaves out: the data is from 1945, and the scale it used is not the one being assumed when someone reads it today. If you want a score anchored to the present rather than to a wartime draft roster, our IQ test reports one, with the percentile attached.
Almost all trace back to Harrell and Harrell’s 1945 study of Army General Classification Test scores for 18,782 white, male U.S. Army Air Force enlisted men, broken down by their pre-enlistment civilian occupation.
No. The AGCT was normed to a mean of 100 with a standard deviation of 20. Modern deviation IQ tests use a mean of 100 with a standard deviation of 15, so identical numbers on the two scales correspond to different percentiles.
Accountants scored highest, with a mean of 128.1, and teamsters scored lowest, at 87.7. Professional and clerical occupations consistently outscored unskilled trades throughout the table.
Yes. A widely cited 2004 review found general mental ability correlates roughly .65 with occupational level attained, one of the largest relationships documented in the social sciences, even though the specific 1945 occupational scores are long outdated.
Corrections: spotted an error? Email corrections@iqmetrics.org and we will update this story and note the change here.
The meta-analytic correlation between measured intelligence and earnings is around 0.23, which leaves roughly ninety-five per cent of the differences in pay to everything else. A 2026 analysis of two large US cohorts adds a second finding: your specific strengths carry a third to a half of the weight that general ability does.
Tables ranking average IQ by degree level circulate constantly and are read as though schooling simply sorts people. Pooled longitudinal evidence puts the correlation at about 0.56, and a meta-analysis of more than 600,000 people found that education also raises scores.
Identical twins raised in separate homes end up more alike on IQ than fraternal twins raised in the same one. That single, replicated pattern is the whole logic behind how heritability gets estimated — and behind why the number is so often misread.
Our IIF-certified assessment reports your score with its scale, percentile and confidence range — and a breakdown of the cognitive domains behind it.
Start IQ Test →
