Why your MBTI result changes: same person, same day, two types

Tuesday, ten in the morning, sixty questions, four letters. You read the description and most of it fits.

Tuesday, nine in the evening, you take it again to show a friend how it works, and the third letter has changed. Same person. Same day. Nothing happened in between except a commute, a difficult meeting and a plate of pasta.

This is the most common complaint about the whole framework and it is treated as evidence that the thing is worthless. It is better understood as the framework showing you something true that it is not designed to display.

The cutoff is doing the damage

Each of the four pairs is scored as a running total. Your answers push a dimension one way or the other, and at the end there is a number, and then the number is compared to a line in the middle and reported as a letter.

That last step throws almost everything away. Somebody who lands heavily on one side and somebody who lands one answer past the middle both get exactly the same letter, in the same font, and are handed the same description. The framework reports a dichotomy; what it actually measured was a position on a scale.

So a person sitting near the middle of a dimension is in an unstable place by construction. One or two answers moving is enough to cross the line, and one or two answers move for reasons as small as which example came to mind when you read the question. Nothing about the person changed. The reported summary of them did.

This is worth saying clearly because the usual reaction, that the test is broken, gets the situation backwards. A near-even score is an accurate measurement of somebody who is genuinely near even. How the sixty questions turn into four letters goes through the arithmetic, and the margin is the part of the result worth looking at hardest.

How often this happens

Often enough that it is a known property rather than a curiosity. David Pittenger, reviewing the published research in 1993 in the Journal of Career Planning and Employment, reported that studies retesting people after a few weeks found as many as half were assigned a different type on the second attempt.

That is one review, and it is the one I will cite, because it is named and can be looked up. I am not going to add the other figures that circulate on this subject, since most of them are quoted without a source attached and I cannot check them.

The important nuance is that a changed type usually means one letter changed, not four. Somebody who was reported one way and is now reported another has typically moved across a single line, and quite often the same line each time. That is a much smaller event than the label makes it look, and it is invisible in a summary that only reports four letters.

Which letters move

Not all four are equally slippery, and knowing which one moved tells you something.

The middle two are the least stable, and the second letter especially. Asking somebody whether they work from what is in front of them or from patterns is asking them to observe something they have never had occasion to compare, and most people honestly believe they do both. S and N explained by a coffee order is my attempt at making that pair concrete enough to answer.

The third letter moves for a different reason: the questions behind it are the ones most tangled up in how people were raised to describe themselves, so the answers are partly a report of a preference and partly a report of what somebody thinks is decent to prefer. That is covered in why T and F is not logic versus emotion.

The last letter tends to be the steadiest, because it attaches to visible behavior: whether you booked the thing, whether the deadline is a target or an edge. It is easier to report a habit than an inclination.

The first letter is steady for most people and unreliable for a particular group: those who are sociable and easily drained, or shy and refilled by company. Their answers genuinely point both ways, because the questions tend to conflate how much you like people with what being around them costs.

The other four reasons

Beyond the cutoff, there are causes that are entirely about the circumstances of taking it.

The first is which version of yourself is answering. Sitting at a desk with a work problem still open produces different answers than sitting on a sofa on Sunday, because the examples that arrive when you read a question come from whatever you were just doing. This is one of the reasons a result from a corporate workshop so often feels slightly off, which why Americans meet MBTI at work retreats goes into.

The second is mood and tiredness. A hard day pushes people toward the answers that describe a hard day, not because they are confused but because the recent evidence is the loudest.

The third is who you are comparing yourself to. Almost every question of this kind is implicitly relative. Are you talkative? Compared with your family, or with your colleagues? Change the reference group and a genuine answer changes with it.

The fourth is that you have read the descriptions. Once somebody knows what the letters mean, they are answering with a destination in mind, sometimes without noticing. The second attempt is never as clean as the first for exactly this reason, and it is a cousin of the recognition problem in why every type description feels like it was written about you.

None of these four are flaws in a particular questionnaire. They are what happens when the measuring instrument is a person describing themselves from memory, which is the only instrument available for this subject. Any test of this kind, from any publisher, at any price, is reading the same imperfect source.

What a changed result is worth

It is not proof that you have no personality and it is not proof that the framework is nonsense. It is a fact about a specific dimension, and the fact is that you sit near the middle of it.

That is genuinely useful information, and it is more useful than a clean letter would have been. It tells you that on that dimension you have a real, working capacity in both directions, which people with a strong score in one direction often do not. Reading both descriptions and taking what fits from each will get you closer to something true than reading one of them and ignoring the other.

The practical rule I use: if a letter changed between attempts, treat that dimension as undecided rather than picking the result you liked more. If three letters held steady and one moved, the three that held are the part worth taking seriously.

MBTI is a trademark of The Myers-Briggs Company, which has no connection to this site. I run it by myself and I am not a psychologist, so this is not a diagnostic instrument and a changing result is not a symptom of anything. The sixteen type pages are descriptions to compare yourself against, which is all any of them were ever meant to be, and taking it again in a week with nobody watching is a better use of ten minutes than arguing with the version you got at nine in the evening.