Inferring MBTI Type From Chat Records

What parts of a chat are predictive of personality type, and why you should be skeptical of 95% confidence claims.

Apr 6, 2026 · 8 min read

MBTI is contested — we know

The psychometric validity of MBTI is debated. That said, the four-axis framework is a useful shared vocabulary, and inferring type from behavior (rather than a self-report quiz) is at least honest about its source: the words the person actually wrote.

Which axis is easiest to predict

E/I is the cleanest. Extraverts initiate more, send more messages per session, and talk about people more often. Introverts send longer but fewer messages and talk more about ideas and objects. The signal is strong.

Which axis is hardest

J/P. The judging-vs-perceiving axis maps to planning and structure — things people do outside of chat. You can only infer it indirectly, from cues like how often they reference schedules, deadlines, and plans. Confidence should stay low.

Why the confidence scores matter

Any tool that confidently declares 'you are an INTJ' from a chat is overclaiming. The honest version of this shows confidence per axis — 'I=85%, N=60%, T=55%, J=52%' — so the reader can weigh the signal.

Compatibility analysis is the useful part

The type inference is a party trick. The real value is in the compatibility analysis: given two inferred types, the report explains where your communication styles align and where they clash, with concrete examples from the chat itself.