5 Personality Test Myths the Research Quietly Killed

A lot of what people repeat about personality tests is structurally wrong — not in a hot-take way, but in the boring sense that the research stopped supporting it decades ago and the culture has not caught up. Some of these myths are inherited from the MBTI marketing of the 1980s. Some come from the way personality tests get covered online. A few are just intuitively appealing in a way that does not survive contact with the data.

Here are five of the most common, with the research that quietly killed them.

Myth 1: "Personality tests measure who you are"

The intuition: you take a test, the test reports your personality, the result is the truth about you.

What the research says: a personality test measures the average of your behavior across many situations, not who you "are" ¹. Behavior varies across contexts widely enough that any single moment of your day may land far from your average. The trait score is the center of a distribution. Your moment-to-moment self can be anywhere along the curve.

Fleeson's 2001 work on density distributions is the clearest version of this finding. He had hundreds of participants rate their own behavior several times a day for two weeks. Each person's behavior ranged across most of the trait scale within those two weeks. What was stable was the average, not the moment ¹.

The practical translation: a personality test is measuring tendency, not identity. "You score at the 71st percentile on Extraversion" is closer to "the center of your social-energy distribution sits there" than to "you are an extravert."

Myth 2: "Your personality type is stable"

The intuition: once you know your type, you know your type. People are basically who they are.

What the research says: this is mostly wrong on two levels.

First, MBTI four-letter types specifically are not stable on retest. When the same person takes the MBTI a few weeks apart, 39–76% of them get a different four-letter type — mostly because they sit near the cutoff on one of the four dichotomies, and small fluctuations in mood or interpretation flip the label ². This is a structural problem with using sharp either/or categories on traits that actually sit on a smooth scale.

Second, even continuous trait scores change slowly over the adult years. Most people become slightly more conscientious, slightly more agreeable, and slightly less neurotic between their 20s and their 60s ³. The changes are real but slow — several percentile points across a decade, not a personality overhaul. Big life events and deliberate effort can accelerate the shifts modestly ³.

The practical translation: your continuous Big Five score this year is a fairly good estimate of next year. The four-letter MBTI type may not be the same in a month.

Myth 3: "If the description fits, the test must be measuring something"

The intuition: when a test report describes you accurately, that is evidence the test works.

What the research says: this confuses Barnum accuracy with measurement accuracy ⁴. A description that is vague, mostly positive, and broadly worded will feel personally accurate to almost anyone who reads it — even when the description has nothing to do with the test result.

Bertram Forer demonstrated this in 1948 by giving all his students the same fake personality description (assembled from horoscopes). The average accuracy rating was 4.3 out of 5. The experiment has been replicated dozens of times since ⁴.

The practical translation: "this description fits me" is not strong evidence that the test measured anything. The honest test is whether some of the description does not fit. If everything lands as accurate, the description is probably doing Barnum work, not measurement work.

Myth 4: "Type-based tests and trait-based tests are basically the same thing"

The intuition: the MBTI puts you in a type, the Big Five gives you scores, but they are measuring the same stuff, just with different labels.

What the research says: they overlap a lot but the format difference matters in ways the popular framing usually misses ⁵.

The MBTI uses sharp cutoffs on four traits and reports a four-letter type. The Big Five reports continuous scores on five traits. Three concrete things change ^{2 5}:

• The MBTI loses information at the cutoff. A 49% and a 51% on E/I get different letters but are essentially the same person. The Big Five does not do this.
• The MBTI does not measure Neuroticism at all. The trait that most strongly predicts day-to-day emotional experience is missing from the type label entirely ⁵.
• The MBTI types are unstable on retest in ways the trait scores are not. Because the type depends on which side of a cutoff you land, small fluctuations flip the label. Continuous scores fluctuate too but do not flip categories ².

The practical translation: a type label and a trait score are not interchangeable. The trait score carries more information and is more stable. The type label is more memorable and more shareable. Both are doing something, but they are not the same thing.

Myth 5: "A personality test can predict job fit, relationship fit, or life outcomes"

The intuition: take the test, find the careers and partners that match your type, and you are set.

What the research says: this is partly true and mostly oversold.

Some traits do predict some outcomes. Conscientiousness is a consistent predictor of job performance across most occupations ⁶. Neuroticism predicts how someone responds to stress and is a notable predictor of relationship problems ⁷. Big Five trait scores carry real signal for outcomes that play out over years.

But the size of the effects is usually modest, and the typical "type X works best in career Y" framing dramatically overstates it. A 1991 National Academy of Sciences review concluded that there was not enough evidence to justify using the MBTI in career counseling ². Follow-up reviews on the MBTI's predictive validity for job performance have reached similar conclusions ⁵.

The Big Five does better than the MBTI on prediction, but even at its best, trait scores tell you about tendencies and likelihoods — not destinies. A high-Conscientiousness person can struggle in the wrong role. A low-Conscientiousness person can succeed in the right one ⁶.

The practical translation: trait scores are useful inputs for thinking about fit. They are not a matching algorithm. The right grip is "this trait pattern makes some paths easier and some paths harder," not "this type belongs in this job."

What this leaves you with

Take all five together and a different picture of personality testing emerges:

• The test measures the center of a distribution, not your identity.
• The score is stable enough to lean on within a year, slow-changing over a decade.
• A description that fits is necessary but not sufficient — also notice what does not fit.
• Continuous trait scores carry more signal than four-letter types.
• Trait scores predict tendencies, not destinies.

That is a smaller set of claims than the marketing version of personality testing usually makes. It is also the one the research supports.

Take the Big Five assessment (12 min) →

References