How to Tell Who's Good at the Job

For most of the last century, the way we decided who to hire was, and I mean this without exaggeration, a form of magic.

We read the bumps on people's skulls. There were bureaus — real ones, with letterhead — that would feel a job applicant's head and tell an employer what work he was suited for. When that went out of fashion we moved to handwriting: the slant of your loops was supposed to reveal whether you had leadership in you, and in some countries a graphologist still gets a look at your cover letter before you do. We studied the shape of faces. We put enormous weight on the firmness of a handshake. And mostly, then and now, we sat a stranger down for half an hour, had a nice conversation, and hired the one who "felt right" — which, if you watched closely, usually meant the one who reminded the boss of a younger version of the boss.

Here is the thing about all of it. None of it worked. The skull bumps predicted nothing. The handwriting predicts nothing. The firm handshake predicts nothing you'd want to bet a company on. And the warm, confident conversation that every one of us has staked a career on — the job interview as it is actually conducted, unscripted, off the cuff — is one of the weakest tools in the whole kit. We built the most consequential decision in a person's working life on a coin flip we'd dressed up in a blazer.

That should be the end of a sad story. It isn't, and this is the part almost nobody knows.

Somebody actually kept score

While the skull-readers were doing their thing, a small and deeply unglamorous corner of psychology was asking a genuinely hard question and — this is the difference — keeping score. Can you predict, before you hire someone, how well they'll actually do the job? Not guess. Predict, and then check whether you were right.

It turns out you can. Not perfectly — people are not that predictable — but far, far better than a handshake, and we now know with real precision which methods work and which are theater. A hundred years of studies, pooled and re-pooled and argued over and corrected, have produced something rare: a cumulative science, one where each generation actually built on the last instead of starting over. It can tell you that a structured interview roughly doubles the accuracy of the casual one. It can tell you that a work sample beats a personality quiz. It can even tell you how to write down criteria for two thousand different jobs without running two thousand separate studies. This is advanced stuff — closer to real measurement than most of what passes for data in a company — and it has been sitting there, finished and load-bearing, for decades.

You have almost certainly never heard of it. That's not your fault. It doesn't get taught. It's buried in journals with names like Personnel Psychology, written in a dialect of statistics designed to keep outsiders out. If you took chemistry in high school you learned how atoms bond, which is beautiful and which you will use approximately never. Nobody sat you down and said: okay, that aside — here's the science of how a stranger will decide whether you get the job you want, and here's how much of it is nonsense. And that second thing is going to happen to you, by a rough count, a few dozen times before you retire.

The part that should make you angry

So we have a solved problem — solved enough to matter — sitting next to a decision that shapes every organization on earth and every working life inside it. You'd expect the gap between the two to be small. It's enormous.

The people whose entire job is this — the human-resources function, the recruiters, the hiring committees — often don't know the science exists. Or they've heard a version of it that's fifteen years out of date. Or they know and can't get anyone to change the process, so the panel still gathers in a room, "just has a feel," and calls it judgment. The single most defensible finding in the field — structure your interview and it works twice as well — is free, takes an afternoon to adopt, and most companies still don't. The answer is on the shelf. The room where the decision gets made is still running on the handshake.

That's the part that gets me, and it's why this series exists. Not that the science is hard — plenty of important science is hard. It's that this particular science is usable, and cheap, and about all of us, and we're still deciding people's futures with the professional equivalent of reading tea leaves, while the manual sits unopened one department over.

Where this goes

So this is a series about the hidden science of telling who's good at the job — what we actually know, how we came to know it, and why the distance between what's known and what's done stays so wide.

Some of it will surprise you. Some of it will make you a little angry about an interview you once sat through, or one you once ran. You don't need any of the statistics to follow it; I'll bring the numbers along and translate them as we go. Start anywhere. If you want the single fact that best captures how far the field has moved — and how wrong the thing most people "know" about it now is — start with The Numbers Moved. If you want the one change you could make to your own hiring on Monday, start with Structure Is the Tool.

Either way, the point is the same. There is a real answer to how do you tell who's good at the job. It's more interesting than you'd think, more settled than you'd guess, and more ignored than it has any right to be.