My type changed between two tests a month apart. Did my personality change or was one of them wrong?

Almost certainly neither. One of your four axes sat close to the middle, and a letter derived from a midpoint is a coin flip dressed up as a category.

Here is the mechanism. Traits like extraversion are distributed as one hump with most people in the middle — not as two clusters. A four-letter instrument cuts that hump at the midpoint and reports which side you landed on. If you score near the edge, the letter is stable and means something. If you score near the middle, two items answered differently on a different day moves you across the line, and the report gives you the new letter with exactly the same confidence as the old one. It never tells you which situation you were in, and that omission is the entire source of the confusion you are describing.

So E flipping to I most likely means your extraversion sits near the middle — which is a real and reportable finding about you, and a more accurate one than either letter. It means contact neither drains nor charges you reliably; it depends on who and how long. That is genuinely useful to know, and the code cannot express it.

One way to check: take a dichotomy scale that reports how far from the midpoint each axis sat, rather than only the letter. If E–I comes back at 48 and your other three are at 70-plus, you have your answer and you can stop wondering. Real personality change over a month does happen — after an illness, a bereavement, a move, a new medication — but it shows up as a shift in degree across several related things, not one letter swapping while everything else holds.

Worth adding: the direction of your flip is the common one, and there is a mundane reason for it. E-to-I between two takes is often about the fortnight before each take rather than about you. Self-report items ask how you generally are, and people answer from what is recent and available — a busy social stretch pulls the answers one way, a quiet stretch pulls them back. That is not you being inconsistent, it is the instrument sampling a window.

If you want the version that does not have this problem, look for measures that are not asking you to summarise yourself: a task that records what you did, or a scale that asks about specific recent instances rather than general tendencies. And when you do retake something, note what the preceding two weeks looked like. If the letter tracks the fortnight, you have learned something more precise than a type.