What does the personality test measure?
The personality test is based on the Big Five (the five-factor model), the most widely used model in personality research. Each of the five dimensions splits into two aspects, so ten aspects are measured in total.
| Dimension | Aspects | Our name |
|---|---|---|
| Extraversion | Enthusiasm · Assertiveness | Energy |
| Openness to experience | Imagination · Intellectual curiosity | Outlook |
| Agreeableness | Compassion · Politeness | Decisions |
| Conscientiousness | Industriousness · Orderliness | Lifestyle |
| Neuroticism | Volatility · Withdrawal | Emotional balance |
The test is made of 60 statements written by our team, 6 for each aspect. Some statements are worded in reverse (for example “Having no plan does not bother me”), so that someone who agrees with everything does not drift to one end. Each statement is answered on a seven-point scale, from “Strongly disagree” to “Strongly agree”.
How are scores calculated?
- Answers to reverse-worded statements are flipped (7 → 1, 6 → 2 …).
- The average of each aspect and dimension is taken.
- That average is compared with a provisional reference point: an estimated average and spread for the scale, not yet measured data from people who took the test (the norm group comes with the pilot study).
- The result is shown as a percentile: by that reference point, “72%” means you score higher than about 72% of people.
The norm group is currently [PILOT STUDY: sample size, age and gender distribution, collection date]. Until the pilot is complete, percentiles are calculated with provisional values.
How do we translate it into a four-letter type?
The four-letter type summarizes four of the five dimensions: extraversion sets E/I, openness N/S, agreeableness F/T and conscientiousness J/P. The fifth dimension, neuroticism, is shown by the -H (Sensitive) or -S (Calm) suffix at the end.
If you are above 50% on a dimension, you get the letter at that end. If you are close to 50%, that letter can come out differently at different times. Reducing personality to four letters loses information, which is why we always show the percentage next to the letter.
The 16-type language is used because it is familiar; our test is not one of the trademarked type tests and has no connection to them. The type names and descriptions are our own.
Where do the sentences on the result page come from?
So that result sentences are not general statements that fit everyone, they are chosen by three rules:
- Aspect combination: sentences are chosen not from a single dimension, but from whether the two aspects of the same dimension are high or low together. Two people of the same type can therefore see different sentences.
- “From your answers” evidence: the quote under a sentence is chosen only from statements you rated 6 or 7 in that direction; the same statement is never used twice.
- Contradiction card: if you rated a statement and its opposite both highly, we do not hide it; we show it on the “Where you pull both ways” card.
Does AI write my result?
Only a small part of it. Your scores, your type and almost every sentence on your result and in your report come from fixed rules and pre-written texts, the same way for everyone.
On top of that, an AI model (Claude, by Anthropic) writes a short personal introduction and four short chapter notes that connect your particular combination of traits. It receives only your scores and the texts our rules chose; never your name, email address, account or test identifiers or your individual answers. Its notes are checked automatically before you see them: no numbers, no type codes, no health language, a length limit and our list of banned phrases. If a note fails a check or is not ready, you simply see the rule-based text.
We also use AI tools while writing, checking and translating the pre-written texts before they are published.
How does the IQ test work?
The IQ test has 30 questions that need no language, in three kinds: matrices (find the missing piece), shape series (find what comes next) and mental rotation (find the same shape). It is made for people aged 16 and over.
Questions are generated anew every time
Questions are not drawn from a fixed pool; they are generated each time with rules for shape, color, number and direction. The same rule gives the same difficulty, but the question looks different. That is why the answers cannot be shared online and memorizing gives little advantage the second time.
The score is more than the number of correct answers
Scoring uses the Rasch model; each question has a difficulty value, and the result is calculated from the difficulty of the questions you solved. Two people with the same number of correct answers can get different scores depending on the difficulty of those questions. This estimate is then converted to the IQ scale, with a mean of 100 and a standard deviation of 15.
Why do we give a range?
A 30-question test measures a score with a margin of error. We show that margin as a range in the result (for example 114–126). The middle of the range is the most likely value.
The question difficulties and norms are currently [PILOT STUDY: sample, age groups, collection date]. Until the pilot is complete, scores are calculated with provisional values.
How is the compatibility score calculated?
There are two different compatibility scores. Type compatibility on general pages (for example “INFP and ENFP”) looks at the four letters of the two types. The same outlook (N/S) and the same decision style (F/T) raise the score. Differences in energy (E/I) and lifestyle (J/P) also count as a plus, because they complete each other. When both people have taken the test, the shared result page puts your percentages on the five dimensions side by side. From those percentages, it picks where you are most alike and most different.
The compatibility score is not a prophecy. Research shows that how similar two people are explains relationship satisfaction only a little.
Reliability and validity
A test needs two things to be useful: it must measure consistently (reliability) and measure what it claims to measure (validity). The values below will be calculated in the pilot study and published here.
| Measure | What it means | Value |
|---|---|---|
| Internal consistency | Do the statements of the same dimension agree with each other? (Cronbach’s alpha) | [α, per dimension] |
| Test-retest | Does the same person get a similar result a few weeks later? | [r, interval] |
| Convergent validity | Do the results match a known open Big Five scale? | [r, IPIP comparison] |
| IQ: model fit | Do the questions fit the Rasch model? | [fit statistics] |
| IQ: measurement error | How wide is the score range? | [mean SE] |
What do our tests not do?
- They are not a health assessment; they do not assess conditions such as depression, anxiety disorders or ADHD.
- They should not be used on their own in hiring, promotion or screening decisions.
- They do not rate personality as good or bad; both ends of every dimension have strong and difficult sides.
- They are not made for people under 16.
If you have felt bad for a long time, we suggest talking to a mental health professional rather than relying on a test result.
Who made them?
The statements, questions and guides were written by our team. [EDITORIAL POLICY (KD-A15): how the texts were produced, including AI assistance, and who checks them] We name a reviewer here only once a real review has taken place.
If you see a problem with the method, write to us; we show corrections in the version history.
Version history
| Version | Date | Change |
|---|---|---|
| 1.0 | [PUBLICATION DATE] | First version: 60-statement personality test, 30-question IQ test, provisional norms. |
| 1.1 | [AFTER THE PILOT] | Norms and reliability values calculated from pilot data. |
Sources
- DeYoung, C. G., Quilty, L. C. and Peterson, J. B. (2007). Between facets and domains: 10 aspects of the Big Five. Journal of Personality and Social Psychology, 93(5), 880–896.
- Goldberg, L. R. et al. (2006). The international personality item pool and the future of public-domain personality measures. Journal of Research in Personality, 40(1), 84–96.
- John, O. P. and Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and theoretical perspectives. In L. A. Pervin and O. P. John (Eds.), Handbook of Personality: Theory and Research (pp. 102–138). Guilford Press.
- Rasch, G. (1960). Probabilistic Models for Some Intelligence and Attainment Tests. Danish Institute for Educational Research.