Is the Myers-Briggs test scientifically valid?

The answer depends on which property of the test is being judged. Reviews find the Myers-Briggs Type Indicator (MBTI) produces fairly consistent scores within a single sitting and correlates with some Big Five scales. Critics point to weak stability of four-letter types over time, a questionable either-or structure, and poorer prediction of life outcomes than the Big Five. The test's publisher defends it as reliable and validated. [1] [2] [3]

A synthesis of 193 studies from 1999 to 2024 on the current Form M found internal consistency of 0.845 to 0.921 across scales. It also found strong convergent correlations (r of 0.50 or higher) between some MBTI scales and the NEO and NEO-FFI Big Five inventories. Its authors concluded that scores had acceptable reliability and validity and may help screen for important personality constructs, while urging caution in interpreting them. [1]

An earlier systematic review and meta-analysis found satisfactory test-retest reliabilities of .75 or higher for three of the four scales, and .61 for Thinking-Feeling. It noted that most studies used college-age students, so the evidence applies mainly to that group. The Myers-Briggs organization similarly says many studies show strong reliability, though some show poor reliability. [4] [3]

The same 25-year synthesis notes that only 65 percent of participants received the same four-letter type on a retest, according to the manual's four-week studies. It found no recent test-retest or factor-analytic studies of structure beyond those in the manual. It also found that two-thirds of comparisons between the manual's national sample and aggregated later samples differed significantly in type proportions. [1]

A study of 559 U.S. participants compared an MBTI-style test with a Big Five test on 37 life outcomes, such as number of close friends, exercise frequency and life satisfaction. The Big Five was about twice as accurate, and adding MBTI-style results to Big Five results did not improve predictions. The authors used tests modeled on the MBTI, not official MBTI forms. They also found that the four traits are roughly bell-curve distributed, so sorting people into two categories discards information; predictions would be about 38 percent better without that split. [2]

On the theory side, a 2019 review in Social and Personality Psychology Compass argues that MBTI theory fails rigorous theoretical criteria because it lacks agreement with known facts and data, and says the test has existed in a parallel universe to mainstream social and personality psychology. [5]

The publisher's position is that reliability depends on the samples and conditions studied, and that many studies show strong reliability. Counseling researchers take a middle view: adequate psychometric evidence, but less enthusiastic than the manual's, with types framed as changeable preferences rather than fixed traits. Critics in personality psychology hold that the type theory and the categorical scoring are the core problems and that the Big Five predicts better. The search results do not include a survey of specialists, so how many hold each view is not established here, though the critical papers cited come from mainstream personality research. [3] [1] [2] [5]