The science, plainly told

Serious science, delivered by a cute brain. This page is the serious half: what the model is, what we do with it, and where it stops.

1. Why the Big Five and not a type

The Big Five is the personality model academic psychology actually uses. Five traits (Openness, Conscientiousness, Extraversion, Agreeableness, Neuroticism), each a dial rather than a switch, recovered again and again across languages and decades. Where you sit on the dials predicts real outcomes: how satisfied you are in relationships (Dyrenforth et al. 2010) and how satisfied the other person is (Malouff et al. 2010), how satisfied you are in your job (Judge, Heller & Mount 2002), how you perform at work (Barrick & Mount 1991), even how lonely you're likely to feel (Buecker et al. 2020).

Type-based tests sort you into boxes with a cut-off down the middle. Sit near the middle, and one answer flips your type. Studies of the best-known four-letter test find 39–76% of people get a different type within five weeks (Pittenger 2005). That isn't a flaw in the people. It's what happens when a dial is reported as a switch. We think the four-letter tests are fun. We wouldn't build a week of your life on one.

2. How Mindpuzzled works

Twenty statements, four per trait, all from the public-domain IPIP item pool: ten come from the Mini-IPIP short form (Donnellan et al. 2006) and ten are Big Five markers from the IPIP's 50-item set. You answer each on a five-point scale. Reverse-keyed items are flipped, the four items per trait are summed, and you get five scores from 4 to 20. That's the whole scoring. No hidden weighting, no algorithm that "learns" you.

Your Archetype's name comes from your scores by a fixed rule: the trait that leads and the one that shapes it, so the same scores always get the same name. Then AI writes about you. It receives your five scores and nothing else for the insight, and writes one insight about how your two most extreme traits interact in an ordinary week. If you subscribe, the same scores (plus the goal you picked in onboarding, and from week two what you actually did the week before) are what the Profile and each weekly Chapter are written from. AI writes the words; the Big Five and your answers decide what they say.

What it isn't: not therapy, not a diagnosis, not a chatbot. It cannot know anything your scores and your own actions didn't tell it. A rubric bans lines that would be true of anyone, and we check generated content against it.

One more thing, on the Career clarity track inside the Journey: an optional, separate set about interest. It's the O*NET® Mini Interest Profiler, all 30 activities word for word, each answered from Strongly Dislike to Strongly Like and summed, as published, into six kinds of work. It's scored on its own and never touches your five scores or your Archetype; it's drawn as a map over the work conditions your five traits point to. Its developers report each five-item scale at an alpha of .70 to .75, and the same top area as the 60-item form for 73% of people (Rounds et al. 2016). We claim nothing beyond that.

This page includes information from the O*NET Career Exploration Tools by the U.S. Department of Labor, Employment and Training Administration (USDOL/ETA). Used under the CC BY-ND 4.0 license. O*NET® is a trademark of USDOL/ETA.

The six plain names, the map and its readings are ours; the items, answer scale and sums are the O*NET® Mini Interest Profiler's.

See it on yourself: twenty questions, free.

3. Honest limitations

  • Twenty items is a short form. Four items per trait is enough to place you on each dial and to be stable across retakes, and it is noisier than a 100- or 240-item inventory (Donnellan et al. 2006). We chose twenty because twenty honest answers beat a hundred careful ones from someone who gave up at sixty. When we show you a re-test, we'll show the noise band next to the change, so "no change" reads as a real result.
  • Self-report has known biases. You can answer as the person you'd like to be. You'd get an Archetype for her, and a Journey written for her. The Mascot says it at question one: answer like nobody's watching.
  • Traits predict; they don't determine. A high Neuroticism score makes hard weeks more likely to feel hard. It doesn't decide what you do about them.
  • We have no clinical validation of our own, and we don't claim one. The model is validated; our twenty-item implementation is built from published IPIP items, half of them the Mini-IPIP short form; the AI narrative is a reading, not a measurement.
  • Traits can move, a little, with effort. Doing trait-relevant things repeatedly shifts the trait; intending to doesn't (Hudson et al. 2019). Across interventions the average change is modest, about a third of a standard deviation over roughly six months, and largest for emotional stability (Roberts et al. 2017). That is exactly the size of thing one Daily Action a day inside the Journey is designed for, and exactly why we don't promise transformation.
The Mindpuzzled Mascot — a friendly puzzle-piece brain — waving hello

4. The no-dark-patterns pledge

  • The price is on one screen before you pay: $14.99 a month, $7.99 a week, $49.99 a year, and the per-week math beside each. No trial that renews on its own. No countdown timers, no fake discounts, no "offer expires". The Paywall looks the same tomorrow.
  • Cancel anytime, on a cancel page we built ourselves; you keep access to the end of what you paid for; your Archetype stays free after that.
  • Your answers and your Profile are private. Sharing your Archetype card earns you nothing, so there's no reason to pressure you into it.
  • No invented counts, stars or testimonials. When a number appears on this site, we can show you where it came from.
  • If a line in your Profile is wrong, you can tell us, and a human reads it.

The only count we publish is finished Assessments, read from our own database, and it stays hidden until there are enough of them to mean anything.

5. Citations