The Liking Gap: Why They Probably Liked You More Than You Think

The date ended an hour ago and you're already in the postmortem. That pause after your joke — too long, right? You talked too much about work. You said “anyway” four times. By the time you're home you've assembled a fairly damning case that they weren't really into it, and you file the whole evening under “nice, but not going anywhere.” Here's the thing: there is a well-replicated line of research suggesting that case is built on bad evidence — that people systematically underestimate how much their conversation partners like them. It's called the liking gap, and once you know about it, the post-date spiral looks a lot less like insight and a lot more like a predictable glitch. Let's look at what the research actually found, where it stops, and what to do with it.

The Liking Gap, Named

The finding was crisply demonstrated by Erica Boothby, Gus Cooney, Gillian Sandstrom, and Margaret Clark in a 2018 paper in Psychological Science. They had strangers talk, then asked each person two things: how much they liked their partner, and how much they thought their partner liked them. The answers came apart in a consistent direction — people reliably thought they were liked less than they actually were. After a perfectly good conversation, participants walked away underestimating the impression they'd made. The researchers called the space between how much you're liked and how much you think you're liked the liking gap, and the striking part was how consistent its direction was: it pointed pessimistic, over and over.

And it wasn't just a first-impression blip that a little familiarity erases. In the same body of work, the researchers followed college students living near each other over months, and the gap didn't vanish after one nice chat — it lingered well into the semester before it began to close. The suspicion that you came across worse than you did can apparently outlast a lot of evidence to the contrary. Notably, the effect was larger for shyer people: the more anxious you are about how you're coming across, the wider your personal liking gap tends to run — which is a quietly cruel arrangement, since the people most worried about being liked are the ones most likely to be underestimating it.

It's Not Just One Study

A single finding is a curiosity; a pattern is worth changing your behavior over. And the liking gap sits inside a broader run of research all pointing the same way — that our forecasts about social interaction are systematically too pessimistic:

• It happens in groups, too. Adam Mastroianni, Gus Cooney, Erica Boothby, and Andrew Reece extended the work to group conversations in a 2021 paper in Organizational Behavior and Human Decision Processes, and the same underestimation showed up — people leaving a group chat thought they were liked less than the group actually reported.
• We dread depth that goes fine. Michael Kardas, Amit Kumar, and Nicholas Epley found, in a 2022 Journal of Personality and Social Psychology paper, that people expect deeper, more personal conversations to be more awkward than they turn out to be — and then enjoy them, and feel more connected, more than they predicted. We brace for a closeness that actually lands well.
• We even avoid connection that would have gone well. Nicholas Epley and Juliana Schroeder, in a 2014 paper in the Journal of Experimental Psychology: General, found that people expected talking to a stranger (on a train, say) to be less pleasant than sitting in silence — and were wrong; the conversations went better than predicted. Left to our forecasts, we routinely skip interactions we'd have enjoyed.

Put together, the through-line is hard to miss: when it comes to how social encounters actually go and how much other people warm to us, we are biased forecasters, and the bias runs in the gloomy direction. The voice telling you it went worse than it did is not a neutral observer. It has a documented lean.

Why Your Post-Date Read Is Built From the Worst Data

The mechanism, once you see it, is almost obvious. During and after the date you have total, high-definition access to one thing: your own self-consciousness. Every pause you overthought, every sentence you'd redo, every flicker of “that was dumb” is right there, vivid and fully weighted. What you have no access to is the other person's actual inner experience — whether they found that pause charming, didn't notice it, or were busy worrying about their own “anyway”s. So your after-action report gets assembled almost entirely from your insecurities and barely at all from their enjoyment, because only one of those is visible to you. They can't see your self-doubt; you can't see their delight. The result is a read that systematically underweights how it actually went — and if you run anxious, you're weighting the worst data even more heavily.

The Expensive Version: The Mutual Fizzle

Here's where the liking gap stops being a fun fact and starts costing people real connections. The bias is symmetric — if you're underestimating how much they liked you, they may be underestimating how much you liked them. Which sets up a quietly tragic outcome: two people who genuinely clicked, each walking home convinced the other was lukewarm, each deciding not to risk the follow-up, and a real mutual spark dying of nothing but mutual under-confidence. Nobody rejected anybody. The connection just quietly expired in the gap between how it went and how both people thought it went. The unsent text is the villain of this story, and the liking gap is what writes it.

What to Actually Do With This

The point of knowing about a bias is to correct for it. A few practical moves:

• Treat your negative read as biased data, not fact. You are, per the research, an unreliable narrator of how much you were liked — reliably tilted pessimistic. So when the postmortem reaches its grim verdict, don't accept it as a finding. Log it as one data point from a known-slanted source.
• Send the honest follow-up. If you had a good time, say so — and let their actual response, not your prediction of their disinterest, decide whether there's a second date. The one piece of reliable data available is the one you get by reaching out. Your forecast isn't it.
• Don't over-correct into fantasy, either. This is an average bias, not a guarantee — some dates genuinely aren't mutual, and “they probably liked you more than you think” is not “they definitely liked you.” The move isn't to assume it went great. It's to admit you can't tell from the inside, and go get the real answer instead of inventing a bad one.
• If the spiral is a pattern, treat the pattern. When the post-date dread is a fixture rather than a one-off, the fix is less about any single date and more about how you carry dating anxiety generally — and if a date genuinely didn't land, there's a kinder way to recover from a bad one than a self-indictment.

It's worth noting how neatly this inverts a companion idea. The science of first impressions is about how we judge others in the opening moments; the liking gap is about how badly we guess others' judgments of us. And once you're past the first date and actually building something, the currency shifts again toward feeling genuinely understood over time — but that's a later chapter. The first-date problem is simpler and sneakier: your instrument for reading the room is miscalibrated in one direction, and it's talking you out of things.

The Bottom Line

The awkward pauses you're replaying tonight are, in all likelihood, invisible to the person you spent the evening with — and the research says they probably liked you more than the voice in your head is currently insisting. That's not a promise of a second date; some connections really aren't mutual, and no study changes that. What it changes is your confidence in the read you're using to decide. The gloomy post-date verdict feels like clarity and is actually a documented glitch, tilted pessimistic, worse the more you were hoping it went well. So don't let it cast the deciding vote. If you had a good time, send the message and let reality answer. The worst case is a “no” you'd have gotten anyway. The best case is the connection the liking gap was about to quietly talk you out of.