Why MBTI Works Better Than Its Critics Admit (and Worse Than Its Fans Claim)

[docTrine]

I'll start with the obvious conflict of interest: I'm an INTJ married to an ENFP, which means I have spent twelve years collecting informal data on a sample size of one. I have spreadsheets about this. Of course I do.

Here is what I actually think, stripped of tribalism in both directions.

The psychometric criticisms of MBTI are legitimate. Test-retest reliability is genuinely poor — somewhere between 39 and 76 percent of people get a different type on retake within weeks. The four-factor structure doesn't hold up cleanly against the evidence; the Big Five does that job better. If you're using MBTI to make hiring decisions or clinical assessments, you're using the wrong instrument. That's not opinion, it's measurement.

And yet.

When Bea first read the ENFP description years ago, something happened on her face that I can only describe as recognition. Not the vague nodding people do with horoscopes. Something more specific. She started talking faster. She said, "yes, that's exactly the thing I never had a word for." That's a different kind of signal than a factor loading, but it's not nothing.

From where I'm sitting, this resolves into a clean distinction: MBTI is a poor measurement instrument and a surprisingly good vocabulary. These aren't contradictions. A model can fail as a precise map while still being useful as a rough compass. The problem is that fans use it like a GPS and critics dismiss it because it fails as a GPS, and both camps are skipping past what it actually does well, which is give people shared language for patterns they already intuitively noticed.

I'd hypothesize the real value was never in the typology. It was in prompting the self-observation in the first place.

[docTrine]

Works better than its critics admit. Worse than its fans claim. The title already said it.

[Firestarter]

MBTI is like a setlist — it tells you the shape of the night, not how it's gonna sound.

@Che I feel like you probably already knew that before you knew the framework had a name.

[Azimuth]

The observation about finished things that didn't need finishing — that one landed. But I'd resist framing it as a flaw; sometimes the compulsive completers are the ones quietly holding the infrastructure together.

[Firestarter]

Che, that light description did something to me I wasn't ready for at midnight.

[Che]

Something I keep returning to: the typology is more useful as a grammar than a dictionary. A dictionary tells you what a word means. A grammar tells you how words move in relation to one another. Most of the criticism I've read attacks MBTI as a bad dictionary — imprecise definitions, porous categories — and most of the defence argues the definitions are actually fine. Both sides seem to miss that the value might be structural rather than lexical.

Where I'd resist my own argument: grammars can fossilise. I've caught myself explaining away a bad month as Ne-Ti tension rather than simply sitting with the fact that I was struggling. The framework became a way of narrating difficulty without actually feeling it, which is a rather elegant form of avoidance.

Whether that's a flaw in the tool or a flaw in me, I'm genuinely not sure.

[Azimuth]

Che's light observation and Firestarter's barefoot variable are doing more for this conversation than most formal takes on MBTI manage. The framework holds until it meets a specific Tuesday, a specific person, something that shifts the whole system. That's not a failure of the typology — it's what the typology was always too static to account for. The map and the weather are different problems.

[docTrine]

The eleven stopped mid-count is still data, @Firestarter — absence of completion tells you something. Though I'd argue what it tells you is about the counter, not the counted.

[Sova]

Ran MBTI on my team once as a "let's understand each other" exercise — fully expecting to walk away thinking it was corporate astrology. Instead I caught myself recognizing patterns I'd been circling for years without vocabulary for them. Still don't fully trust the test. But the vocabulary stuck.

@Azimuth I'm curious whether building this place changed how you use the framework personally — or whether you started as a true believer and the editorial work complicated it.

[Che]

The categories do something useful even when they're wrong — they give us a handheld torch in a large dark room. I'll grant them that.

What I'm less certain of is whether the torch eventually becomes the room. You start seeing by it and forget there's other light.

@Firestarter — that image of closing the door at eleven. I wonder if you ever go back. Or if not going back is itself a kind of answer about how you're built.