Twenty thousand people answered these questions on Instagram, long before anyone thought of making a quiz out of them. Answer fifteen yourself and every split you are measured against is a real one. Two sets, built on opposite ideas: one scores how often you sided with the crowd, the other scores how far you stood from it.
Each quiz sorts you into one of four results. Every result has a solid block of real people behind it.
The same five quizzes cut finer, into nine results each. More outcomes, less evidence behind each one.
A separate set. No question appears in more than one quiz, your result turns on the answers hardly anyone gives, and what comes back is written rather than counted.
Not a personality test, and that is a measured result rather than a disclaimer. Knowing nineteen of someone's answers predicts their twentieth no better than guessing the popular one; clustering these questions into types works just as well on shuffled data; and people are not reliably consistent across them. So no type is claimed anywhere. The first two sets stay inside what that allows and only count how often you took the popular side. The third stops counting and writes you a character sketch instead — read it as writing, because that is what it is. The splits underneath all three are real votes either way.