The eternal search for our other half (and what the science covers)

Everyone wants to know if compatibility is real. Whether there's an actual science behind "we just click," or whether it's a story we tell after the fact to explain something that was mostly luck. It's the question behind every dating app's matching algorithm and every "are we right for each other" conversation at two in the morning.

We don't have the research to tell you who to marry. But the same science tells us a lot about who you should work for.

What compatibility means

Compatibility isn't chemistry, and it isn't a gut feeling you either have or don't. In the research, it's the measurable overlap and complement between two behavioral profiles: where being alike helps, where being different helps, and where neither matters much at all. That's a more useful definition than "it just works," because it tells you what to actually look at.

The research behind it comes from a specific place, though: studies of people and their managers, not people and their partners. Person-supervisor fit, specifically, the match between someone and the individual they report to at work. That's the evidence we're pulling from here, and it's worth being upfront that it doesn't generalize to romantic pairing. Nobody's run the dating version of this study. So consider the rest of this a look at what fit predicts at work, where the evidence is genuinely strong.

What fit predicts

Here's the finding worth sitting with: person-supervisor fit predicts job satisfaction and supervisor satisfaction strongly, at correlations of roughly 0.44 and 0.46. It predicts overall job performance far more weakly, at around 0.18, weakly enough that the estimate's confidence interval includes zero.

In plain terms: a good fit with your manager makes it much more likely you'll be satisfied and stick around. It's a much weaker signal for how well you'll actually perform the work. Fit gets sold, informally, as a performance lever. The evidence supports it as a retention lever. Those are two different promises, and only one of them holds up.

Similar in some ways, different in others

The more specific finding is where it gets interesting, because "similar people work well together" turns out to be only half true.

On warmth, similarity is what predicts good outcomes: people who are similarly warm, or similarly reserved, tend to get along and perform well together. On dominance, the opposite holds: complementarity is what predicts good outcomes. Two highly assertive people in the same dynamic tend to compete rather than collaborate. Two unassertive people tend to stall, waiting for the other to take the lead. What works is one of each.

So "opposites attract" isn't false. It's just narrower than the phrase suggests. It's true on one specific axis, dominance, and not a general principle you can apply everywhere.

Why this isn't one score

If fit predicts satisfaction and retention well, and performance only weakly, then blending everything into a single compatibility number quietly averages a strong signal against a weak one and reports the result as though it were one thing. That's why compatibility isn't reducible to a single score, done properly. Different outcomes need different signals, and collapsing them into one number hides exactly the distinction that matters.

The takeaway

Compatibility is real, and it's measurable, but the evidence we have is about work, not romance. It predicts whether you'll be satisfied with your manager and want to stay far more reliably than it predicts how well you'll perform. And it isn't about being similar across the board. It's similarity on warmth, complementarity on dominance, and neither one is the whole story on its own.