Short, and still not overstating
The short version is not a replacement for the full one. Its job is to find, quickly, which axes are clear and which are wobbling.
What was wrong with a fixed ten
Put the four MBTI indices together with attachment anxiety and avoidance and there are six axes to check. Ten questions cannot even put two on each. The old short version gave each MBTI axis two, and attachment anxiety and avoidance one apiece. Waver on that one question, or be pulled by how your love life is going that day, and the last letter of a 64-type code is settled by a single item.
And when the two questions pointed opposite ways, a total of zero was still forced into one side or the other. Putting precise-looking figures on the result screen — an "attachment health score", a national average — does not make questions this short able to support that precision. We concluded the problem was not the presentation but the allocation of questions and the way the result was worded.
How the adaptive 12–18 questions work
The first twelve are two questions on each of six axes. Where the two point the same way, that axis is reasonably clear even in a short quiz. Only where they split, or where "can't say" comes up often, does one more question arrive, asking about a different situation. At most six of those, so the whole thing is 12 to 18 questions.
| Axis | When the short version fits | Both sides of the result |
|---|---|---|
| E–I | How you recharge after being with people, and whether you sort thoughts out aloud or inside | Leaning extravert / leaning introvert |
| S–N | Concrete examples against possibilities; a method you have used against a new connection | Leaning sensing / leaning intuition |
| T–F | How you weigh a consistent standard against the effect on people when deciding | Leaning thinking / leaning feeling |
| J–P | The comfort of a closed plan against the comfort of options left open | Leaning judging / leaning perceiving |
| Attachment anxiety | Anticipating rejection; watchfulness about replies and shifts in distance | Lower / higher |
| Attachment avoidance | Asking for help; the sense of losing autonomy to closeness | Lower / higher |
Confidence is not an accuracy rate
The confidence on the result screen is an internal figure, 0 to 100, for how far the answers on one axis lined up in the same direction. It is not the probability that this is really your type, and not a ranking against other people. Low confidence is not a failure: it means the axis shifts with the situation, or sits near the middle, or has not been asked about enough.
Low confidence on E and I might mean you think by talking at work and recover alone at the weekend. Low confidence on J and P might mean you keep to deadlines but want the daily method left free. Knowing which situations make you waver is more use in a life than settling on four letters.
How to read the four attachment categories
Attachment combines the two axes, anxiety and avoidance, into an indication of secure, anxious, avoidant or fearful-avoidant. With only two or three questions each, though, it is swayed by the relationship you are in now, by a recent argument, by the days just after a breakup. The full 46-question version checks several situations; the short one is only a door in.
- Higher anxiety: More anticipating abandonment, more checking their reaction, more seeking reassurance again.
- Higher avoidance: More wariness about depending on someone, more sense of losing freedom to closeness, more cutting feeling off.
- Both mid-range: Close to a boundary; the letter changes easily with who you are with and when.
Why we ask which MBTI you think you are
If another quiz, or long self-observation, has already given you an MBTI, you can set it beside the short estimate. A mismatch does not mean one of them must be wrong. If it differs on an axis the short version was unsure about, there were probably not enough questions. If it differs on a confident axis, that is material for asking whether you answered as the self you would like to be, or whether you had work in mind rather than private life.
Once enough has accumulated, we will publish the agreement rate per axis between the MBTI people call themselves and the short estimate — not only an overall percentage, but which axes suffer most from being shortened, which is what we will use to improve the extra questions. The self-reported MBTI is never mixed into scoring.
How the questions were rechosen
In a short quiz, a question carrying several meanings at once widens the error. "I like parties" takes in not only extraversion but sensitivity to noise, whether you have friends there, and how you feel about drinking. "I like a plan" depends on more than J–P: on how anxious you are, on what your job obliges. So the items were rewritten as situations that show how you take information in, how you decide, how you handle a schedule and how you adjust distance from people — not as bare preferences.
Each MBTI axis has one question asked from the opposite direction, so that a habit of agreeing with everything does not tip the result one way by itself. Attachment anxiety and avoidance are separate axes, so they are not bundled into a single "I am bad at relationships" item: watchfulness about rejection is kept apart from watchfulness about depending on someone and letting them close.
Personality and circumstances cannot be fully separated
However well the items are written, answers are affected by circumstances. Someone who has worked with customers for years has the technique to speak to a stranger even if they lean introvert. Someone whose plans change constantly for childcare or nursing may be handling it flexibly even if they lean J. Asking not whether you can do a thing but whether it recharges you afterwards, and which state feels natural and costs least, makes it a little easier to tell a role from a preference.
How to compare when you retake it
If you want to check a result, leave two to four weeks rather than taking it again straight away. Retaking it with the first result in mind leads to steering towards it, or deliberately away from it. Use the same basis — your ordinary self over the last six months — and compare the direction and confidence per axis rather than the type code.
- Same letter, same direction: that axis is being reported fairly consistently.
- Same letter but lower confidence: there may have been more exceptional situations lately.
- A different letter with low confidence both times: treating it as the middle band is the natural reading.
- A different letter with high confidence both times: check whether the situation you had in mind, or your circumstances, have changed.
How the anonymous data reaches the articles
The optional statistics store the estimate, the aggregated value on each of the six axes, how many questions were answered including the extra ones, and — if you gave one — the four letters you call yourself. Answers to individual items are not stored. Once there is enough, we will count whether E–I agrees readily even in the short version, how far the result letter wobbles for people whose attachment anxiety sits mid-range, and which axes the people who get many extra questions have in common.
Articles will not show a combination with fewer than ten cases, and will state plainly that these are tendencies among this site's users, who opted in themselves. They are not treated as a distribution of the general population or as causation. The data is not used to label an MBTI type as attachment-unstable; it is used to find which items are easily misread and what gets lost in shortening.
How it is scored, and where that stops
Seven-point answers are converted to −3 through +3, with reverse-worded items sign-flipped. An axis whose two base questions total between −2 and +2 gets one confirming question. The four MBTI axes take their letter from the sign; attachment anxiety and avoidance convert the axis mean to a 0–100 figure and give an indication of the four categories. Confidence comes from the absolute mean of the answers within an axis. This is a simple algorithm of this site's own; it is not a standardised instrument and not a medical diagnosis.
MBTI and attachment style are separate theories, and the 64 types that combine them are this site's own content categories. Do not use a result to decide hiring, treatment, a course of study or a partner. If distress, or a pattern that keeps repeating with other people, is affecting your life, take the specific difficulty to a professional — not the name of a type.
Short version or the 46-question one?
If you want to compare with friends for fun, to get the shape of it first, or you cannot spare the time, the short version suits. If your result feels wrong, if two or more axes came out low-confidence, or if you want to look closely at trouble in relationships and attachment, use the 46-question one. Leaving a few days rather than taking both the same day, and keeping the same basis — your ordinary self over the last six months — makes them easier to compare.
FAQ
Do extra questions mean a bad result?
No. It is only checking, in a different situation, an axis where the two base questions came out near the middle. Getting a lot of them may mean you switch between modes depending on the setting.
My type changes every time.
A low-confidence axis is probably straddling the boundary. Rather than the whole code, look separately at the axes that stay and the axes that move. The full version has a middle band too.
Should I answer as my friends see me?
Go by your ordinary self over the last six months, not by what others say. Don't limit it to work or to relationships — across several situations, choose whichever is more often true.
If I share the result URL, can people see my answers?
The shared URL contains only the five-character result code. Your answers to individual questions, and the MBTI you say you are, are not put into it.
