Ten-Item Personality Inventory: ten questions for the Big Five
Ten items is an absurd number for measuring personality, and that is the point. Gosling, Rentfrow and Swann published the TIPI in 2003 for research where a full questionnaire will not fit: web surveys, epidemiological studies, repeated measurement.
Each item is a pair of adjectives, answered on a 7-point scale from 1 (disagree strongly) to 7 (agree strongly). The whole thing takes about a minute, and anyone can use it for any purpose without asking permission.
Two items per dimension
| Dimension | Items | Reverse-scored |
|---|---|---|
| Extraversion | 1, 6 | 6 |
| Agreeableness | 2, 7 | 2 |
| Conscientiousness | 3, 8 | 8 |
| Emotional Stability | 4, 9 | 4 |
| Openness to Experiences | 5, 10 | 10 |
Items 2, 4, 6, 8 and 10 are reverse-scored before scoring, and each dimension score is the mean of its two items. Extraversion is therefore the average of a positive statement and a reversed negative one, with the negative item flipped first. The dimensions map onto the Big Five, with emotional stability standing in for the reverse of neuroticism.
Reliability, honestly
Two items per trait cannot produce strong internal consistency, and the original study did not pretend otherwise. Alphas in the undergraduate sample were .68 for extraversion, .73 for emotional stability, .50 for conscientiousness, .45 for openness and .40 for agreeableness, with the full scale above .55 and the range running 0.40 to 0.73 in a sample of 1,813 students.
A 2023 scoping review pooled 29 studies, 27 versions, 18 languages and 27,427 participants. The overall mean alpha was 0.53. Dimension means were .64 for extraversion, .58 for emotional stability, .56 for conscientiousness, .47 for openness and .37 for agreeableness. Only the Turkish version cleared a conventional bar on all five dimensions, at 0.81 to 0.86.
Retest reliability looks much better than internal consistency, which is the usual pattern for short scales. All 17 versions with retest data met the review's threshold, with an overall mean of 0.74 and dimension values from 0.69 to 0.78. The original English version averaged 0.71, and its own two-week retest ran from .62 for openness to .77 for extraversion.
Does it measure the same five things?
Convergent correlations with longer instruments are the main support: 0.68 with the BFI across versions, 0.53 with the NEO-FFI, 0.54 with the NEO-PI-R, and 0.56 to 0.68 against the NEO-PI-R in the original study.
Structure is where it gets shaky. Of 19 versions tested with factor analysis, 10 supported the five-factor structure, 3 supported it only after statistical adjustment, and 6 failed. The Chinese version was among the failures. A Japanese validation with 902 students reported item-pair correlations from -.22 to -.59, two-week retest values from .64 for conscientiousness to .86 for extraversion, and convergent correlations of .45 to .69 with one longer inventory and .60 to .85 with another. A separate Japanese study of 163 students found correlations of 0.46 to 0.70 with the corresponding NEO-PI-R-J dimensions, a mean of 0.59, with 28 of 30 facets pointing in the expected direction.
The Chinese version, tested in 2,223 medical students, reported alphas from 0.119 to 0.785 and supported the expected structure for extraversion, emotional stability and conscientiousness only.
What it is not for
A ten-item scale cannot support facet-level statements, and it should not be used where a full personality assessment is needed, such as counselling assessment or diagnostic work. Short scales also tend to understate how strongly personality predicts anything, so a weak correlation from the TIPI does not mean the trait is unimportant.
There are no clinical cutoffs and no norms that turn a score into a category. The published norms are unpublished data collected through a music and personality site, so they describe that sample rather than a population. A score here is a snapshot of how you described yourself today. This page is not a diagnosis and does not replace one.