Average IQ of Lawyers: What the LSAT, Bar Exam and Army Data Show

What is the average IQ of lawyers? No one has measured it. The figure most pages repeat, about 128, comes from a 1945 study of Army Air Force enlisted men who had worked as lawyers before the war, scored on a scale whose standard deviation was 20, not the 15 of modern IQ tests. On today's scale that 127.6 would be roughly 121, and it describes wartime recruits, not practising lawyers. What lawyers do take is the LSAT and the bar exam, and both are published by the bodies that run them: LSAT test takers averaged 152.66 in 2024-25, and 63% of everyone who sat a bar exam in 2025 passed. Neither number has an IQ equivalent.

Where does "lawyers average 128" come from?

From Thomas and Margaret Harrell's 1945 paper in Educational and Psychological Measurement, which grouped Army General Classification Test (AGCT) scores for 18,782 enlisted men in the Army Air Force Command by the civilian job each had held before enlisting. In the slice of the table our guide to IQ by occupation prints, the mean for accountants is 128.1 and for lawyers 127.6, second on that list. Those men were a wartime sample of one branch, one sex and one racial group who had held the job, not a sample of practising lawyers, and the AGCT is not a modern IQ test.

The scale is the part charts leave out. Our occupation guide gives the AGCT's mean as 100 and its standard deviation as 20. If so, a mean of 127.6 sits 1.38 standard deviations above average, which is the 92nd percentile of the AGCT population and about 121 on a scale with a standard deviation of 15. We could not open the 1945 paper to check its printed standard deviation or the number of lawyers in the sample, so treat 121 as arithmetic, not as a published IQ. Our IQ score converter does this change of scale for any score.

Later figures are no firmer. A 2009 report in the trade paper The Lawyer on a University of Bristol study said lawyers' ability scores had drifted toward the average between the 1958 and 1970 birth cohorts, and added that a score of 100 is not an actual IQ and that rescaling would leave lawyers at "just 108". That 108 is the newspaper's rescaling of an index, not a measured IQ. And the finding that high earnings do not require top ability, from Swedish conscription data, is explained in our piece on the IQ of billionaires.

What does an LSAT score say about ability?

The LSAT is scored 120 to 180. LSAC publishes the distribution of the scores it reports, and the table for the 2023-24, 2024-25 and 2025-26 testing years gives the percentage of scores below each value:

• 150: 36.56% of scores are lower
• 160: 71.06%
• 165: 85.17%
• 170: 94.48%
• 173: 97.59%
• 175: 98.72%
• 178: 99.57%

Two details change how to read the table. First, LSAC staff have said it does not consider the number of people taking those tests and includes tests cancelled by the test taker, so it counts tests, not people. Second, the people are a self-selected group. LSAC's May 2026 technical report shows 96,036 test takers in 2022-23, 107,062 in 2023-24 and 123,973 in 2024-25, with a mean of 151.87, 152.16 and 152.66 and a standard deviation of 10.63, 10.62 and 10.96. The mean was 150.94 in 2018-19. A 170 is therefore the 94th percentile of people who chose to take a law school entrance exam, not of the population, the same trap we describe for the MCAT. LSAC says the share of scores from 160 to 180 has been rising for about seven years while the share from 120 to 145 has been falling, which it links to eight test dates a year instead of four, and so more chances to delay or retest.

The rising scores show up at the top schools. Columbia's 2026 entering class had an LSAT of 171 at the 25th percentile, 173 at the median and 175 at the 75th, from 8,591 applicants for 424 places. Those are scores of enrolled students at one school, and our analysis of Harvard students shows what such admissions data can and cannot say about intelligence. There is no official conversion from LSAT to IQ; our blog explains why in GRE and LSAT scores to IQ.

Does the LSAT predict law school grades?

Yes, moderately. LSAC's 2020 to 2024 correlation study reports, as the mean across schools, a raw correlation of .42 between LSAT score and first-year average, .24 for undergraduate GPA alone and .50 for the two combined. Corrected for the fact that schools see only the applicants they admit, a restriction of range, the figures are .60, .41 and .66. A correlation of .42 means the LSAT shares about 18% of its variance with first-year grades. That is a statement about law school performance, not about general intelligence, and we found no published correlation between the LSAT and an IQ test.

Two recent changes affect the meaning of the score. From the August 2024 test, LSAC removed the Analytical Reasoning section, popularly known as logic games, and replaced it with an extra Logical Reasoning section; it says the test is as hard as before. And LSAC announced that from the August 2026 LSAT the vast majority of test takers must test in person, with limited exceptions, after about 60% had been testing in a test center. Some descriptions of the LSAT online still refer to logic games.

What does the bar exam tell you?

The National Conference of Bar Examiners (NCBE) reports that jurisdictions counted 67,442 people taking the bar exam in 2025, with an overall pass rate of 63%. First-time takers numbered 46,356 and 76% passed; repeaters numbered 21,086 and 34% passed. By sitting, first-timers passed at 79% in July 2025 (38,642 people) and 60% in February 2025 (7,714). First-time status is counted by the reporting jurisdiction only.

• Passing scores on the Uniform Bar Exam run from 260 in states such as Alabama, Iowa and Minnesota to 270 in states such as Alaska, Arizona and Colorado, so the same score can pass in one jurisdiction and fail in another
• The July 2026 national mean on the Multistate Bar Examination was 140.7, from 44,472 examinees, about 76% of them probably first-timers
• The NextGen bar exam reports a single score from 500 to 750 and lets each jurisdiction set its own passing score; it was first given in July 2026 in 10 jurisdictions, and the large states follow in July 2028

We found no NCBE or LSAC page giving an IQ equivalent for any bar score, and a pass rate is a share of candidates, not a measure of ability on the IQ scale. Passing a bar exam also depends on the jurisdiction, the sitting and whether it is a first attempt.

Are lawyers smarter than doctors?

No study we found compares the two professions on one test, and there is no conversion between an LSAT and an MCAT score. Each test is normed on its own self-selected applicant pool, so a percentile on one cannot be moved onto the other. The closest consistent statement is the pattern in our guide to IQ by occupation: scores rise with occupational level on average, and the spread within every occupation is wide. Whether IQ tests may be used to hire is a separate legal question, covered in are IQ tests legal for hiring.

How to check any "IQ of lawyers" claim

• Name the test and the scale: an AGCT score of 128 is not an IQ of 128
• Ask who was tested: enlisted men with a past job are not practising lawyers, and LSAT takers are applicants
• Ask for the standard deviation before comparing two numbers
• Check the year: the AGCT table dates from 1945 and the LSAT lost a section in 2024
• Treat a percentile as a rank within the people who took the test, not within the population

The defensible answer to "average IQ of lawyers" is that nobody knows, and that the numbers in circulation are a 1945 Army score on a different scale, a newspaper rescaling and percentiles of a self-selected pool. To see what a score looks like on the standard scale, use the IQ percentile calculator or take the IQ Metrics IQ test.

Common questions