Personality types do not exist as separate kinds of people, as far as current research can tell. Traits such as extraversion vary smoothly from one person to the next, and a type is a label put on where your scores fall. Some combinations of traits are more common than others, so types can still be a handy summary, as long as you read them as a shorthand for scores rather than a box you belong in.
The useful question is not whether types are real, but what a type label keeps and what it throws away.
What Would Have To Be True For Types To Exist
A type is a real category only if people fall into groups with a gap between them, the way most people are either left handed or right handed. If that were true of personality, scores on a scale would pile up at the two ends and thin out in the middle.
There are three ways to check this:
| Check | What a real type would show | What studies find for normal personality |
|---|---|---|
| Shape of the scores | Two humps with a dip in the middle | One hump, heaviest in the middle |
| Taxometric analysis | A hidden category under the scores | A smooth dimension |
| Cluster analysis | Groups that come back in every sample | A few loose groups, with many people in between |
What Taxometric Studies Found
Taxometric analysis is a family of statistical methods designed to tell a true category apart from a smooth scale. It was made for exactly this question.
Arnau and colleagues (2003) applied it to three type questionnaires, including the Myers-Briggs Type Indicator. The preferences behaved like continuous dimensions, not hidden categories.
The broader picture is the same. Haslam, Holland and Kuppens (2012) reviewed 177 taxometric articles with more than 533,000 participants in total. Category findings became much rarer in newer and better designed studies, and normal personality gave little persuasive evidence of any true category. Their conclusion was that most of what personality psychologists measure is dimensional.
What About Studies That Found Personality Types?
Some studies do report types, and they are worth reading closely because they find something narrower than the word suggests.
Asendorpf and colleagues (2001) found three repeating patterns, which they called resilient, overcontrolled and undercontrolled, in children and adults across several samples and methods. Gerlach and colleagues (2018) analysed four large data sets with more than 1.5 million participants and reported at least four robust types.
These are patterns in trait scores: combinations that more people share than you would expect. They are not walled-off groups. The Gerlach team also showed that common clustering methods produce many false types alongside the real ones, which is one reason so many type systems claim to find their own groups. Neither study supports the idea of 16 distinct kinds of people.
How Much A Type Label Leaves Out
To see what a label does to the underlying scores, we worked through the scoring of our own 16 personality types test with a short script.
Each of the four scales has 12 statements, giving 49 possible scores per scale. That is 5,764,801 different score profiles, and the test sorts all of them into 16 labels. Each type covers between 331,776 and 390,625 score profiles. Two people with the same four letters can have very different scores, and two people with different letters can be one answer apart. Our post on whether MBTI is scientifically valid gives figures for that gap.
Our Big Five personality test takes the other route. Each trait has 41 possible scores, and the result shows the score itself, with a reading of low, middle or high (35% or less, 36 to 64%, 65% or more). Nothing is forced to one side. In a model where every answer is equally likely, about 78% of results on a trait land in the middle reading, and the scores form a single peak at 50% with no dip. That is a model of the scoring, not visitor data, but it shows the shape a smooth scale produces.
Types Vs Traits: Which Tells You More?
| Type result | Trait scores | |
|---|---|---|
| What you get | A label, such as INFP | A score on each trait |
| Middle scores | Pushed to one side | Shown as they are |
| Easy to remember | Yes | Less so |
| Change on a retest | Letters near the middle flip often | Scores move a little |
| Best for | Starting a conversation about preferences | Seeing where you really sit |
A type is easier to talk about. Traits keep more of the information. Our guide on how accurate online personality tests are explains why trait scales are the better studied of the two.
How To Use A Type Without Being Boxed In
- Read your scores first. A letter based on a score near 50% is a mild lean, and both sides may describe you.
- Treat the type description as a starting point. Keep the sentences that fit and drop the ones that do not.
- Expect neighbours to fit too. Types that share three letters are often close in scores.
- Do not use a type to rule anyone out. A label cannot tell you what someone can do, only what they said they prefer.
- Check with a trait test. If both point the same way, the pattern is more likely to be stable.
Frequently Asked Questions
Are personality types real?
Not as separate categories. Research shows personality traits vary smoothly, and type labels are cut-offs on those scales. Some trait combinations are more common, which is why a few broad types turn up in large studies.
Why do personality types feel so accurate?
A type description is written to fit people with a clear lean, and most of us recognise ourselves in parts of it. The parts that fit stand out more than the parts that do not, so the whole description feels right.
Is the Big Five a type system?
No. The Big Five gives a score on each of five traits instead of a category. Some studies group Big Five scores into common patterns, but the traits themselves are scales.
How many personality types are there?
There is no agreed number, because the count depends on where you draw the lines. The 16 types come from splitting four scales in two. Large data studies have found three or four common trait patterns, not 16 separate groups.