Personality Tests at Work: What's Legit and What's Not

"So, what's your MBTI?" at the end of the interview

Here's a true story. A candidate cleared the first two rounds, things were going great, and in the final interview someone asked it casually: "By the way, what's your MBTI?" The candidate said "INTP." The interviewer's face shifted a little. "Ah. We kind of need a more social person on this team." The candidate didn't get the job.

Forget for a second whether that's legal where you live. Just on the merits of the test itself, it makes no sense. Four letters are supposed to tell you whether someone will do good work? Honestly, those same four letters might flip if you retook the test next month.

So let's sort this out. Of all the personality tests companies use, which ones have some actual footing, and which are closer to reading tea leaves. And then how to keep using them without wrecking the people who take them.

Why MBTI shouldn't touch your hiring decisions

This isn't me saying MBTI is no fun. I like joking with friends about who's a total ENFP. The problem starts when you wire it into a serious decision like "should we hire this person."

First, the same person retakes it and gets a different answer. The rate at which a test gives you the same result on a second sitting is called test-retest reliability, and with MBTI it's shaky. Plenty of people see one or two of their four letters flip after just a few weeks. Your mood, a rough day at work, how your relationship is going that week, all of it nudges the answers. A hiring tool whose output wobbles every time you measure is a rubber band pretending to be a ruler.

Second, and this is the bigger one, MBTI type doesn't predict whether someone will perform on the job. That property is called predictive validity. The idea that ESTJs make better managers or INFPs make better designers, that clean correlation just isn't in the data. Job performance comes from actual skill, experience, motivation, and fit with the team, not from whether someone leans E or I.

Third, MBTI chops people into two bins. Someone who's 51 percent extraverted and someone who's 99 percent extraverted both get stamped with the same "E." In real life those are two very different people. Tons of people sit near the middle on these traits, and forcing them all to one side throws away most of the useful information.

There's a fourth thing people skip over. The four-letter system creates a false sense of completeness. Sixteen tidy boxes feel like they cover everyone, so a hiring manager stops looking for what the boxes leave out, like how the person actually handles conflict, deadlines, or being wrong. A neat label can be worse than no label, because it makes you stop asking questions.

So what actually holds up better

There's no perfect test. I'll just say that plainly. But there are less sloppy ones.

The closest thing to a consensus in psychology is the Big Five. It looks at personality along five dimensions: openness, conscientiousness, extraversion, agreeableness, and neuroticism. The key difference is that it scores you instead of sorting you. You're not "a conscientious person" versus "an unconscientious person." You're, say, around 70 out of 100 on conscientiousness. And of those five, conscientiousness in particular tends to link to performance across almost every kind of job, fairly consistently. If you want the longer version of how these trait models work, I unpack it in the psychology guide hub.

But there's something far stronger than any personality test for hiring: watching someone do the actual work. These are called work-sample assessments. Give a developer a real coding task. Have a marketer sketch a campaign plan. Hand a designer a fake brief and see what they make. When you think about it, it's obvious. If you want to hire a good driver, you watch them drive. You don't hand them a personality quiz.

Structured interviews hold up reasonably well too. Every candidate gets the same questions in the same order, and the answers get scored against criteria you decided on beforehand. That narrows the gap where "I just got a good feeling from this one" sneaks in. It's less fun than freewheeling chat, but it's fairer, and fairness is the whole point.

Notice the pattern across all the defensible options. They tie back to the job, they resist the interviewer's gut, and they treat people as points on a scale rather than members of a tribe. That's the bar. Anything that fails all three should make you nervous.

When "I'm just a J" turns toxic

Here's the part I actually wanted to get to. Even if a company never uses these tests for hiring, a team can still misuse them in ways that quietly poison things.

The most common one is using a type as an excuse. A coworker keeps blowing off meeting times and laughs it off with "I'm a P, I'm just bad with schedules." In that moment MBTI becomes a permission slip for not doing the job. Missing commitments isn't a personality type. It's missing commitments. "I'm a T, I can't really do empathy" is the same trick. Empathy isn't a letter. It's effort you decide to spend.

The flip side is using a type to cage someone else. "She's an I, let's not make her present." Maybe she presents well. Maybe she wants the chance. Reading four letters and quietly pulling the opportunity off the table means somebody loses a shot at growth without ever being asked. That's how a label becomes a ceiling.

So does that mean a team should ban the whole thing? No. There's a way to do it right.

Use the result as the start of a conversation, not a diagnosis. Picture a team workshop where people talk about their type and share how they actually operate: "I absorb feedback better when it's written down," "I'm calmer in meetings if I get the agenda beforehand." That's not boxing anyone in. That's building a shared language for understanding each other. And here's the catch that proves the point: two people with the same INTJ result might feel completely opposite about meetings, so what matters in the end isn't the letters. It's the self-description the person gives you in their own words.

The whole thing comes down to direction. A type should never mean "this is who I am, so cut me slack." It should mean "I work better under these conditions, so let's adjust toward them." The first is an excuse. The second is collaboration. Same test result, opposite outcomes, and the difference is entirely in how the team chooses to hold it.

One more honest admission: I've caught myself doing the excuse version. Skipped a tedious follow-up and half-thought "eh, not my type." It's a comfortable lie. The fix is boring but it works: name the actual behavior, not the letter behind it.

Red flags in workplace testing

Let me hand you a practical checklist. If a company you're joining, or already work for, uses personality testing in any of these ways, treat it as a warning light.

A single test score deciding pass or fail is bad. In any serious evaluation, a test should be one input among several, never the thing that eliminates a person on its own. The moment someone says "rejected for a low score," something's broken.

If the test vendor throws around phrases like "99 percent accurate" or "scientifically proven," get suspicious. The people who actually build sound assessments tend to be upfront about the limits. The ones insisting they're flawless are usually the flimsiest.

Not showing you your own results, or refusing to explain them, is a problem. It's a test about you, but you don't get to see it and the company does? That's closer to surveillance than assessment.

Using the same test for hiring, promotion, and firing all at once is another signal. When one tool claims to settle every decision, it usually means someone is over-trusting it, or hiding their decisions behind it so they don't have to own them.

And a subtler one: a test can work against certain groups of people. Wording that lands differently depending on someone's culture or background can skew results in ways nobody intended. Good companies actually think about this and check for it instead of assuming the test is neutral.

There's also the consent angle people forget. Were you told what the test was for, who would see it, and how long it'd be kept? If a test got slipped into onboarding paperwork with no explanation, the issue isn't the test. It's that nobody treated you like an adult who gets to know what's being measured about them.

Personality tests aren't evil. A hammer that smashes a thumb isn't a bad hammer, it's bad swinging. Used as a tool for understanding yourself, a test is great. Used as a blade for cutting people, it's not. If you're in the mood for more self-exploration that isn't about your job at all, there's a pile of it over on the blog.

For the record, every test on Selvora, and this article, is for fun and self-reflection. None of it is meant to inform hiring, HR, or legal calls. If you genuinely have to evaluate people at work, use validated tools and get real expert help. What's here is more like a light mirror, the kind that gets you to look at yourself one more time. Nothing more serious than that.