MBTI (Myers-Briggs Type Indicator)

What is the MBTI?

The Myers-Briggs Type Indicator (MBTI) is a personality assessment that reports preferences using a four-letter type, such as INFP or INTJ. Katharine Cook Briggs and Isabel Briggs Myers developed it from Carl Jung's ideas about psychological types. The official instrument, the theory behind it, and online quizzes using similar letters are different things.

The letters can be memorable. The harder question is what they justify. A type description may give you language for a familiar pattern without establishing that people fall into distinct natural categories or that your type determines your career, relationships, or ability.

What the four letters mean

In the Myers-Briggs framework, each position represents one of two preferences:

• Extraversion (E) or Introversion (I): An orientation toward the outside world of people and activity or toward the inner world of reflection.
• Sensing (S) or Intuition (N): Greater attention to concrete information and details or to patterns and possibilities.
• Thinking (T) or Feeling (F): An emphasis on impersonal principles or on values and consequences for people when making judgments.
• Judging (J) or Perceiving (P): A preference for structure and closure or for keeping options open in dealing with the outside world.

These terms have meanings within the framework. Thinking does not mean intelligent, Feeling does not mean irrational, and Judging does not mean judgmental. The model describes preferences rather than a monopoly on a way of acting.

Four pairs with two choices produce 16 combinations. For example, INFP combines Introversion, Intuition, Feeling, and Perceiving. People with the same letters can still differ in many important ways. The tests and assessments collection includes selected type descriptions; those descriptions are explanations of type language, not separate assessment instruments.

Selected type descriptions on this site: ENFP, ESFJ, ESFP, ESTJ, INFP, INTJ, INTP, ISTP.

The official instrument and other four-letter tests

The official MBTI is a proprietary assessment. Its versions have changed over time: the publisher distinguishes Form M, Form Q, and the later Global assessments in its instrument history. A claim about one version should not be silently applied to every other version, language, or online quiz.

Step I reports the four-letter type; Step II adds 20 facets, five within each preference pair. Facets are narrower aspects of a preference, so Step II can describe differences between people who receive the same four-letter result. The official feedback process also asks people to consider their best-fit type: whether the reported preferences match their understanding of themselves. Receiving a result and discussing whether it fits are therefore distinct parts of the process.

16Personalities is not the official MBTI instrument. Its own framework explanation describes the NERIS model, which adapts Big Five dimensions, uses five scales, and adds the Assertive/Turbulent distinction. A result such as INFP-T therefore identifies a different system. Similar labels do not make the questionnaires, score meanings, or evidence interchangeable.

Types and dimensions: a worked example

Imagine a hypothetical scale that runs from 0 to 100. A reporting rule assigns everyone below 50 to category A and everyone at or above 50 to category B. Someone scoring 49 and someone scoring 51 receive different labels, despite being two points apart. People scoring 51 and 90 receive the same label, despite being 39 points apart.

This is an illustration of categorization, not the MBTI scoring system. It shows the information a category can hide. A small change around a boundary can change a label without showing a large psychological change. To establish distinct types, researchers need more than the ability to calculate a category: they need evidence that the proposed categories describe the underlying differences well.

A Big Five profile reports differences by degree. That does not make every Big Five questionnaire good, but it preserves a distinction that a four-letter label alone cannot show: how far apart two measured scores are.

What does the evidence support?

Do the four-letter categories represent distinct kinds of people, or do the underlying scores describe differences by degree? In a 1989 study, Robert McCrae and Paul Costa examined MBTI scores alongside self-reports and peer ratings on the NEO Personality Inventory, a trait questionnaire. Their sample included 468 adults. They found that the MBTI indices captured aspects of four Big Five dimensions, but reported no support for genuinely dichotomous preferences or qualitatively distinct types. “Dichotomous” means divided into two separate kinds, rather than varying along a range.

Evidence for distinct types would have to support grouping people into separate categories, rather than merely dividing a range of scores at a chosen boundary. Researchers can compare whether a category-based or continuous description better accounts for patterns of responses and related traits; the appropriate test depends on the claim and measurement.

In the earlier hypothetical example, the scores of 49 and 51 could reflect a small difference; assigning different labels would not establish two fundamentally different kinds of person. McCrae and Costa interpreted their findings as four relatively independent dimensions rather than distinct types. That is the conclusion of this study, not proof that every conceivable way of grouping personality is impossible. It predates later MBTI versions and is not a direct evaluation of the current Global assessments.

A separate question is whether the questions within each scale give consistent information. The publisher's research discussion reports internal-consistency coefficients, called Cronbach's alpha, of .87–.89 for the four Global Step I preference scales in a global sample of 16,773 people. Alpha summarizes how the responses to a scale's questions relate to one another. These publisher-reported values support consistency among the questions in that sample; .89 does not mean that 89% of people received an accurate type. Questions can produce consistent scores without proving that the 16 categories are natural divisions of personality.

What happens when people take the instrument again? The same publisher summary reports Global Step I scale test–retest correlations of .81–.86 in two samples: 588 people retested within six weeks and 1,133 retested after seven to fifteen weeks. These correlations describe how closely the first and second sets of scale scores correspond across people. A positive correlation near 1 means people who scored higher initially tended to score higher again. It does not mean that everyone's numerical score stayed exactly the same.

Keeping the same four-letter label is a different test of consistency. Separately, the publisher's discussion of whole-type results in global assessments reports about half retaining all four letters and 90% retaining three or four. The latter figure includes the people who kept all four: it does not mean another 90% kept three. Someone whose result changes from INFP to INTP retains three letters but not the complete type. As the boundary example illustrates, even a small shift in a score can sometimes change a category. Scale correlations and exact type agreement therefore answer different questions, and neither is a percentage-accuracy score.

These findings address consistency and the interpretation of scores. They do not establish whether an MBTI workshop improves communication: that question needs evidence comparing outcomes from the workshop with an appropriate alternative. See reliability for consistency and construct validity for evidence about what a score means.

What should you do with a result?

The MBTI Code of Ethics rejects using results to screen job applicants and warns against directing someone toward or away from a career, relationship, or activity solely on type. The publisher also states that MBTI is not a mental-health diagnostic instrument. These are limits on use, not merely objections from critics.

If a description helps you notice that you prefer time to prepare before a difficult conversation, investigate that specific pattern. Try receiving an agenda beforehand or writing down questions. Whether the change helps is a practical question you can observe. You do not need to prove that all people with your letters respond in the same way before trying it.

Before paying for an assessment or using a report to make a consequential decision, identify the actual instrument and read the test evaluation guide. If a report includes numerical comparisons, check its scoring explanation and reference group; a percentage displayed by one provider need not mean what a percentile means elsewhere.

Disclosure

Jason Hreha owns Twofold, a personality-assessment product. That commercial interest is relevant to this site's discussion of other assessments. Research about personality models or other instruments does not establish the quality of Twofold's tests. The evaluation criteria and disclosure apply the same questions to each instrument and distinguish provider descriptions from research findings.