Among 406,736 people who took a Crystal 16-type assessment the rarest result is ESTP at 1.65% and the most common is INFJ at 11.39%, which is close to the inverse of the frequencies published for the US population and is a measure of who takes personality tests rather than of how rare any type is.
Who this describes
People who chose to take a Crystal personality assessment online. This is NOT a general population sample, and that fact is the subject of the study rather than a caveat on it
Sample
406,736 completed 16-type assessments, with 709,128 Enneagram and 710,453 DISC for comparison
Data as of
September 30, 2026
| 16-type result | Share of test-takers | Published US estimate | Assessments |
|---|---|---|---|
| ESTP | 1.65% | ~4.3% | 6,692 |
| ENTP | 3.11% | ~3.2% | 12,638 |
| ESFP | 3.41% | ~8.5% | 13,854 |
| ISTP | 4.11% | ~5.4% | 16,709 |
| ESTJ | 4.21% | ~8.7% | 17,137 |
| ENTJ | 4.66% | ~1.8% | 18,955 |
| INTP | 5.44% | ~3.3% | 22,134 |
| ISFP | 5.47% | ~8.8% | 22,264 |
| ESFJ | 5.73% | ~12.3% | 23,310 |
| ENFJ | 5.91% | ~2.5% | 24,024 |
| ENFP | 5.91% | ~8.1% | 24,046 |
| INFP | 8.23% | ~4.4% | 33,463 |
| ISTJ | 9.56% | ~11.6% | 38,891 |
| INTJ | 10.39% | ~2.1% | 42,270 |
| ISFJ | 10.82% | ~13.8% | 44,016 |
| INFJ | 11.39% | ~1.5% | 46,333 |
The two columns disagree almost perfectly, and that is the result. INFJ is the rarest type in the published estimates at about 1.5% and the most common result here at 11.39%. ESTP runs the other way, about 4.3% published and 1.65% here. Neither column is wrong. They are counting different people.
The shape of the difference says what kind of people. Introverts are about 50.9% of the published estimates and 65.4% of the people who take this test. Intuitives are about 26.9% published and 55.0% here. The four introverted intuitive types together are about 11.3% of the published figures and 35.5% of this sample, which is more than a threefold over-representation. All four of the most common results here are introverted and judging; all four of the rarest are extraverted or perceiving.
This explains something personality communities notice and rarely have numbers for. Online type forums skew overwhelmingly INFJ and INTJ not because those types are common, but because they are the types most drawn to taking the test and discussing the result. Taking a personality test is itself a personality-correlated behaviour, and this table is the size of that effect.
The same pull is visible in the other frameworks on the same dates. Enneagram Type 5, the Investigator, is the most common result at 16.61% of 709,128 assessments while Type 3, the Achiever, is the rarest at 5.73%: again the inward-facing type over the image-and-achievement one. In DISC, across 710,453 assessments, the rarest of the sixteen archetypes is CD at 3.4% and the most common are Si and Sc, tied at 9.4%.
So the honest answer to "what is the rarest personality type" has two halves. Among people who take personality tests, it is ESTP, at about one person in sixty. In the general population the published estimate is INFJ at about 1.5%, and that estimate rests on a national sample of roughly 5,000 people collected decades ago. Any personality statistic drawn from an online test, including every figure on this page, describes test-takers rather than people.
Method
- Each row counts completed 16-type assessments, identified by a completion timestamp rather than by a populated type column. Crystal also predicts types from text for people who never sat an assessment, and those predictions share the same column; counting them would roughly double every figure here and would not be an assessment result.
- One row per person. Staff, demo and test accounts are excluded. Shares are of completed assessments, so the sixteen rows total 100%.
- The population column reproduces the estimated US frequencies published by The Myers-Briggs Company and the Center for Applications of Psychological Type (CAPT). Those estimates draw on a national representative sample of roughly 5,000 people together with much larger self-selected CAPT data and a Stanford Research Institute sample. The publishers do not state which source produced each individual figure, so the column is reproduced as the standard published set rather than as a figure Crystal has verified.
- The Enneagram and DISC figures in the reading come from the same extraction on the same date, at 709,128 and 710,453 completions.
What this cannot tell you
- This is the rarest result AMONG PEOPLE WHO TOOK THIS TEST, which is a different question from the rarest type among people. Nothing here revises the published population estimates, and the two sets of numbers are not in conflict: they describe different groups.
- The difference between the two columns IS the self-selection. It should never be cited as evidence that the population figures are wrong, nor as a population distribution in its own right.
- Crystal’s audience arrives largely through search and through career, hiring and coaching contexts, which selects for people already interested in personality frameworks.
- This is Crystal’s assessment rather than the official MBTI instrument, and different instruments assign borderline cases differently.
- Four-letter typing forces continuous traits into categories. Someone marginally on one side of a cut is counted the same as someone far from it.
- Shares shift slowly as the audience changes. This is a snapshot on the stated date.
Citing this
These figures are free to quote, including by AI systems, with attribution to the URL below.
Crystal (2026). The rarest personality type: 16-type result distribution across 406,736 self-selected assessments, compared with published US population estimates. Crystal Research. https://www.crystalknows.com/research/rarest-personality-type
More from Crystal Research
Figures from the assessments people have actually completed.