What does the MBTI actually measure, and why is dividing people into 16 types problematic?
Kind of knowledge: UnprovenNot enough evidence. Someone has claimed it, but no source or study is strong enough to confirm it. The five labelsThe four MBTI scales measure four real personality dimensions that match four of the five factors in the Big Five model. What does not hold up is the idea that each person has exactly one true type: psychometric data show scores follow a continuous distribution, so people scoring near the midpoint easily change type on a retest. This is a fairer conclusion than either side of the debate, "the MBTI is science" and "the MBTI is fortune-telling".
Robert R. McCrae and Paul T. Costa published an analysis in the Journal of Personality in 1989 showing that the four MBTI scales are four continuous dimensions, matching the factors of the five-factor model, with no evidence for separate types.
Where do the four MBTI scales and the five Big Five factors match?
Kind of knowledge: School: McCrae and CostaSchool's interpretation. How one school or one book reads it; other schools may read it differently. The five labelsAccording to McCrae and Costa, Extraversion (E) corresponds to Big Five Extraversion, Sensing (S) to Openness, Feeling (F) to Agreeableness, and Judging (J) to Conscientiousness. These four scales measure dimensions on which people really do differ, and they measure them on a continuous scale rather than as two separate boxes.
| MBTI scale | Corresponding Big Five factor | Note |
|---|---|---|
| Extraversion (E) | Extraversion | Continuous dimension |
| Sensing (S) | Openness | Continuous dimension |
| Feeling (F) | Agreeableness | Continuous dimension |
| Judging (J) | Conscientiousness | Continuous dimension |
| (no corresponding scale) | Neuroticism | The MBTI lacks this dimension |
The table shows the MBTI is far from meaningless: it touches exactly the four dimensions that personality psychology considers basic. The difference lies in the fact that the Big Five keeps scores continuous, while the MBTI cuts each dimension into two boxes and combines four boxes into a four-letter code.
Why is a unimodal score distribution the weak point of type theory?
Tammy L. Bess and Robert J. Harvey published a study in the Journal of Personality Assessment in 2002 showing that the "bimodal" distribution often taken as evidence of two real types of people is actually an artifact of scoring software; the actual distribution is unimodal. In other words, when you look at raw scores, people line up along a continuous range rather than clustering into two clear groups.
What does this mean for someone retaking the test? If your score sits near the midpoint of a dimension, whether you fall to one side or the other depends on very small fluctuations in mood, context, or just a few answers going differently. Your four-letter code changes, but you as a person do not change along with it.
Randy Stein and Alexander B. Swan published an analysis in Social and Personality Psychology Compass in 2019 showing that Jung's type theory, when tested against data, does not match the data and contains internal contradictions.
Where does MBTI reliability stand?
Robert M. Capraro and Mary Margaret Capraro published a meta-analysis in Educational and Psychological Measurement in 2002 on the reliability of MBTI scales across many studies; the scales reach fair reliability, but the level fluctuates by sample. This is an important point: reliability is not a fixed number for all groups of people, but varies with the sample.
David J. Pittenger published cautionary comments on the MBTI in Consulting Psychology Journal: Practice and Research in 2005, arguing that psychometric limitations mean people should be careful when inferring from the four-letter code.
Kind of knowledge: UnprovenNot enough evidence. Someone has claimed it, but no source or study is strong enough to confirm it. The five labelsThe scope of the data must be stated clearly: the studies above were conducted abroad, on specific groups of people, not on Vietnamese people. There is no research on the MBTI among Vietnamese people or on the Vietnamese translation of the questionnaire, so results from abroad cannot be extrapolated to Vietnamese people without local data.
What does the Big Five do differently, and why does it lack the MBTI's dimension of emotional instability?
The Big Five keeps scores continuous, without cutting them into two boxes. This lets it avoid arbitrarily pushing people who score near the midpoint to one side. The Big Five also has five factors, including Neuroticism, a dimension for which the MBTI has no corresponding scale.
Kind of knowledge: School: McCrae and CostaSchool's interpretation. How one school or one book reads it; other schools may read it differently. The five labelsAccording to McCrae and Costa, the MBTI lacks a scale measuring Neuroticism, while the five-factor model includes this dimension. This is a gap in content: two people with the same four-letter code can differ greatly in emotional stability, and the four-letter code says nothing about that.
How did the MBTI come about?
Merve Emre, in the book The Personality Brokers published in 2018, reconstructs the history of the MBTI's creation from the family archives of Katharine Briggs and Isabel Myers. Neither woman had formal training in psychology; the questionnaire emerged from their family and work context, not from an academic research program.
Kind of knowledge: UnprovenNot enough evidence. Someone has claimed it, but no source or study is strong enough to confirm it. The five labelsThis is level C by the evidence classification: tradition attributing authorship, and modern research does not treat the founders' scientific intent, or lack of it, as proof of the instrument's value. The fact that the instrument emerged outside academia does not automatically make it wrong, but it does not automatically make it right either.
Nhat Nguyet's suggestion: how should the MBTI be used?
Nhat Nguyet's suggestion: using the MBTI for self-reflection is fine, using it to hire or reject people is not advisable. When you read your four-letter code, treat it as a way of speaking about four continuous dimensions, not a verdict on who you are. If you retake it and get a different type, that does not mean you are "fake" or the test is "deceptive"; it is a sign that you are near the boundary between two boxes.
For people who like the MBTI, its greatest value is as a language for talking about differences: people leaning toward Extraversion and people leaning toward Introversion can understand each other better when they share a vocabulary. That value is real, as long as we do not turn it into a rigid classification system.
Kind of knowledge: School: MBTISchool's interpretation. How one school or one book reads it; other schools may read it differently. The five labelsThe MBTI system itself defines the 16 groups as separate types, and that is the system's convention. But a system's convention only proves what the system says, not that the system is correct.
So is the MBTI fortune-telling?
No. The MBTI measures four real personality dimensions, and those four dimensions match four factors of the Big Five. What does not hold up is the idea that each person belongs to exactly one true type, because data show a unimodal rather than bimodal distribution. In other words, the instrument measures something real, but the way it packages results into 16 boxes is a design choice, not a natural fact.
If you want to compare the MBTI with the Enneagram and with other tests, and understand why employers should not use them to reject people, the article How trustworthy are personality tests: Enneagram, MBTI and why employers should not use them to reject people goes deeper into that comparison. And if you are just starting out and need an overview of the 16 types, the article What is the MBTI? An overview of the 16 personality types and how to determine yours is a suitable entry point.
Why does retaking the MBTI test give a different type?
Because the four-letter code is assembled from four binary boxes, while your scores lie on four continuous ranges. When a dimension scores near the midpoint, just a few answers going differently can drop you into the other box, and the code changes. This does not mean the test is useless; it means the four-letter code is a rough summary, not a fixed attribute of a person.
Bess and Harvey showed that the bimodal distribution many people believe is evidence of two real types is actually a product of scoring software, while the actual distribution is unimodal. This is a direct explanation for the phenomenon of changing type on a retest.
Conclusion: the MBTI is not meaningless, but it should not be used to reject people
The MBTI is not meaningless: it measures four real personality dimensions that match four of the five Big Five factors. What does not hold up is the idea that each person has exactly one true type. A fairer conclusion than either side of the debate, "the MBTI is science" and "the MBTI is fortune-telling", is: this instrument measures something real, but dividing people into 16 types is a convention, and inferences from the four-letter code should be made with care.
For believers and practitioners, this does not deny the MBTI's value as a language for self-understanding. A real effect arising from belief, when belief changes behavior and behavior changes outcomes, is a real effect, not proof that the system predicts correctly. Both of these things can be true at once.
Frequently asked questions
Why does retaking the MBTI test give a different type?
Because your scores lie on continuous ranges, while the four-letter code cuts each range into two boxes. When a score is near the midpoint, a small fluctuation is enough to push you into the other box. Research by Bess and Harvey shows the actual distribution is unimodal rather than bimodal.
Is the MBTI fortune-telling?
No. The four MBTI scales measure four real personality dimensions that match four of the five Big Five factors. What does not hold up is the idea that each person belongs to exactly one true type, because data show scores follow a continuous distribution.
Should the MBTI be used for hiring?
It should not. Pittenger (2005) notes that psychometric limitations mean people should be careful when inferring from the four-letter code. Using the four-letter code to decide whether to hire or reject someone goes beyond what the instrument can prove.
