Our science

A unique 100D model for each person

Personomist builds a unique profile of each person's personality from about a hundred traits. You can think of it as a high-resolution 100D psychological model. In the next steps, a person's model is compared with other models that help discover their personality from different angles.

To compute the Big Five, for example, we compare a person's personality model with the models of extraversion, conscientiousness and the other Big Five traits. To describe the Wellbeing landscape, we compare it with the models of stress resilience, ability to concentrate and ability to relax. To place a person in the career landscape, we compare their model with the models of people in more than 120 jobs. And so on: the same model can be used for a great many comparisons.

How the model is built

1. We ask about 100 questions about typical ways of behaving, thinking and feeling. Each question is about a different nuance of personality.
2. We neutralise the answers for gender and age, so that a person is compared with people of the same age and gender. For this we use associations found in the data of about 14,000 people, each of whom was also described with the same questions by someone who knows them well.
3. We analyse the patterns in the answers with a psychological model pre-trained on a large sample of people. It finds a suitable compromise between the person's own answer and what their other answers, given other people's data, predict about the same trait. Each answer thus becomes, together with contributions from many other answers, a feature of its own in our 100D model. The average person's value on each feature is 0, and people typically differ from it by about one unit. In the comparison group, this step raised the agreement between self-descriptions and other people's descriptions by more than 30% on average.
4. This is the person's basic psychological model, which can be used flexibly to draw other conclusions.
5. From other people's data we have built a model of each Big Five trait, of more than 120 jobs, of the Wellbeing landscape's traits and of much else. A person's basic model can be compared with these models, giving a similarity and showing exactly which features raise that similarity and which lower it.

What does "90% similarity" mean?

Personomist uses a very large database on what sets the people in each job apart from people in other jobs. Your similarity to each job shows how strongly your personality traits point the same way. At 50% your personality model leans neither towards it nor away. At 90% it clearly points the job's way. The percentage roughly follows percentiles: 90% means your model points the job's way more strongly than about 90% of people do.

We show the range 5–95%, because 0 and 100% would suggest a complete (dis)similarity this kind of estimate does not give. It does not take skills, education or work experience into account. Note that the results are descriptive: how similar you are to the people in that job in what sets them apart from others. They do not say which work you should do or what is right for you. There are a great many other reasons why people do one job or another.

Personomist gets to know you better and better

The first time you answer, you already get a useful profile, but it can be made better still.

• If you answer again a few days or a couple of weeks later, the quality of the information improves and the model becomes more stable and reliable. Answers given within 30 days are taken together. Psychologists call this improving reliability.
• Adding information from other sources reduces systematic biases, related for example to a person's self-perception or to how they understand the questions. It also opens the chance to see yourself through others' eyes, which no approach based only on a person's own information can offer.
• Personality is not fixed, so it is worth building a new model of yourself in future and comparing it with the earlier one.
• The personality profile can be extended with the results of other questionnaires, on wellbeing, interests and, in future, other things.

Users who have saved their results can fill in the questionnaires repeatedly, either to make the results more reliable or to follow changes. You can also invite other people to answer the questions about you. Their answers are combined, and results are shown only once at least two people have answered. Then Others' view opens, and the Combined view, which synthesises everyone's information and is usually the most accurate. In the Combined view your self-description carries the same weight as each description by someone else.

Each result is the model's prediction about you. Read it against your own experience and look for examples in your life.

What the results are calculated from

• Traits and the Big Five14,000 people

  Basis: the comparison group, about 14,000 people from the Estonian Biobank's personality data (Arumäe et al., 2025). The Big Five give a bird's-eye view: five broad dimensions from 60 of the traits, as in Anni, Vainik and Mõttus (2025). We call neuroticism Emotionality.

• Wellbeing landscape16,226 people

  Ability to concentrate, Ability to relax and Stress resilience, estimated from personality using associations found in the data of 16,226 Estonian Biobank participants (Soodla et al., 2025). These scores bring together combinations of stable tendencies that go with different levels of wellbeing, so they describe the psychological landscape behind wellbeing. How work has been going for you over the last 30 days is what the work wellbeing questionnaire shows, and in future a questionnaire on the different sides of general wellbeing.

• Jobs83,828 people

  Personomist compares your personality model with the models of 128 jobs. They rest on 83,828 people working in those jobs: Estonian job means (Anni, Vainik and Mõttus, 2025) and the public data of the whichjob.me app, from which people who have stayed in their job and are satisfied with it were taken. Each job shows how many people its model rests on.

• Work rolesAI ratings

  Work-role models rated on the same traits by several AI models. They are not based on data from people.

• Free timeUniversity of Tartu

  University of Tartu research data on what people with different personalities do in their free time. About 2,000 people took part, describing more than 5,000 free-time activities. We will soon add at least as much data again.

• Recommendationsyour goal

  The habit recommendations were written by clinical psychologists to match your personality profile. A recommendation weighs where you are now and how far that is from how you want to be, which you describe in the ideal-self questionnaire; until you do, your model is compared with what is generally considered desirable.

• Work wellbeingBAT-23

  The Burnout Assessment Tool, an internationally used questionnaire grounded in research literature. The questions are about wellbeing at work over the last 30 days and are brought together into five scores. The scores' risk-level cut-offs currently come from the BAT manual and rest on a sample of 1,403 people in Flanders, but we plan to make them more thorough. The results are not diagnoses. Work wellbeing recommendations come with the result, written by a work psychologist to match the pattern of your answers.

• ComingUniversity of Tartu

  A non-clinical wellbeing questionnaire, with norms from University of Tartu research.

Who is behind it

Personomist's science comes from the personality and wellbeing research group at the University of Tartu's Institute of Psychology. The team includes clinical psychologists. The research and development were co-funded by the European Union (ASTRA+ measure).