similarminds.comCOMPILER Personality System
COMPILER is a SimilarMinds personality system built around eight broad dimensions. Each dimension has two named poles; one letter from each pair forms an eight-letter type code. The high-pole initials spell COMPILER, and the sixteen pole names all have different initials.
Take the Short Test · Explore interests by type
How the system was developed
The starting point was historical Facebook-linked SimilarMinds personality data. An eight-factor model was fitted to 35 stored personality-trait scores in a balanced sample of 60,000 profiles. Factor scores combine information across the source traits using regression weights, rather than treating one original question as an entire dimension.
Financial wealth, self-rated attractiveness, excess weight, relative height and same-gender attraction were not inputs to this model. A music-preference item about the Pixies was also excluded, so musical taste was not used to define the personality factors later compared with music preferences.
The eight dimensions were then given the COMPILER labels below. The labels describe the patterns in this particular model; they do not establish that eight is the only possible number of personality dimensions, or that the named poles are mutually exclusive in everyday life.
The eight dimensions
| Pole pair | What it describes |
|---|---|
| C: Caregiving / A: Autonomy | Organizing life around care for others and close relationships versus emphasizing personal independence and freedom. |
| O: Openness / T: Traditionalism | Intellectual, imaginative and artistic exploration versus a preference for familiar approaches and established ideas. |
| M: Materialism / F: Frugality | The importance placed on appearance, money, status and expensive consumption. This is about priorities, not how wealthy someone actually is. |
| P: Prudence / S: Spontaneity | Order, planning, caution and deliberate self-control versus a more improvised, flexible approach. |
| I: Intimacy / G: Guardedness | Emotional openness, interpersonal trust and willingness to accept support versus keeping feelings and vulnerabilities more private. |
| L: Leadership / Y: Yielding | A desire to direct or influence and comfort with confrontation versus giving way or preferring someone else to lead. It is not a measure of leadership ability. |
| E: Equanimity / V: Volatility | Emotional calm and steadiness versus stronger emotional reactions and a greater need for reassurance. |
| R: Religiosity / N: Nonreligiosity | Religious faith and conventional restraint versus less reliance on religious guidance. This is descriptive, not a judgment about either outlook. |
From continuous traits to 256 types
People have a continuous score on every dimension. An eight-letter code is a compact summary: eight two-way choices produce 28, or 256, possible combinations. The code always follows the C, O, M, P, I, L, E, R dimension order.
For the historical analysis, each dimension uses its own fixed median cutoff across all 253,717 eligible profiles. This divides each individual dimension as close to half-and-half as possible while keeping tied scores together. The cutoff does not change within a type or according to someone's preferences.
Balanced individual dimensions do not produce 256 equally sized groups, because the dimensions are correlated. The type groups are useful summaries for exploring the data, not evidence that everyone falls into one of 256 sharply separated natural kinds.
The Short Test
Your results show where you lean between the two sides of each trait. Scores run from 0 to 100, with 50 in the middle; they are not a ranking against other people. Your eight-letter code summarizes those preferences. A question mark means your answers landed exactly in the middle.
What the interest explorer adds
The same historical type definitions were compared with listed music, television, movies, interests, books, activities and hobbies, religious orientation, political orientation, and education concentration. For example, the music analysis contains 139,946 usable profiles; the usable sample differs by category and type.
“Most listed” shows common entries. “Distinctive” looks for entries listed more often by the selected group than by other profiles, subject to minimum counts, effect-size checks and consistent direction across two data halves. Music results show artists and bands rather than mixing them with genres.
The explorer starts with a full type. You can also broaden a search to one selected trait pole or any combination of trait poles. Counts and comparison rates are recalculated for the combined group, not averaged from the separate type rankings.
These are historical group associations, not predictions of what a particular person will enjoy. The comparisons are not adjusted for age, gender, geography or list length. Matching a questionnaire code to a historical code remains a provisional bridge between different instruments.
Ongoing development
COMPILER is being refined through new responses, item analysis, comparisons with MOTIVES and tests of whether questionnaire scores recover the intended dimensions. Reliability, boundary calibration and prediction of new, independent preference data still need to be established. Future versions may change items, labels or scoring as that evidence develops.
Take the COMPILER Short Test · Open the interest explorer · Detailed methods and coverage