How do I know if someone really likes me?

You know someone really likes you by what they do for you specifically over a few dates: they make plans that happen, they remember what you told them, they follow up. One great night can’t tell you, and neither can a feeling, but a pattern of small, specific effort usually can.

The good news is that you’re probably better at reading this than you give yourself credit for. The tricky part is keeping two lists apart in your head: what they’ve shown you, and what you’re hoping it means.

How do you tell real interest from someone just being nice?

Look for effort that’s aimed at you, not effort they’d spend on anyone.

Picture Dani, 34, three dates in with Marco. Marco is charming with everyone: the bartender, the host, the couple at the next table. That’s lovely, and it tells Dani very little. What tells her more is that on the third date, Marco asked how her dad’s surgery went, something she’d mentioned once, briefly, on the first date.

Research backs up the instinct to look for the specific. In a speed-dating study by Paul Eastwick, Eli Finkel and colleagues, 156 people went on four-minute dates with 9 to 13 others each.3 People who were especially drawn to one particular person tended to be liked back by that person.3 People who were drawn to everyone at the event were liked less.3 Interest that’s spread around reads differently from interest that’s aimed.

So the useful question isn’t “are they warm?” It’s “are they warm toward me in ways they wouldn’t be toward just anyone?”

Why does it feel so hard to trust my own read?

Partly because you’re probably underestimating how well things are going. Erica Boothby, Gus Cooney, Gillian Sandstrom and Margaret Clark ran five studies on what they called the liking gap: after ordinary conversations, people consistently thought their partner liked them less than the partner actually did.1 It showed up with strangers in a lab, at a workshop with members of the public, and between college roommates over a whole year.1 Those weren’t dates, but the pattern is familiar. You replay the pause after your joke and miss the three times they laughed.

And here’s the other side: you can read interest better than the worry suggests. In a 2009 study, Skyler Place, Peter Todd, Lars Penke and Jens Asendorpf showed observers video clips from real speed dates.2 People could predict how interested a dater was from clips as short as 10 seconds.2 They weren’t perfect. Predictions were sharper for men’s interest than women’s, which were just above chance, and some daters were simply hard to read.2

So your read is a real tool. It just works better on more footage. A full evening, and then another one, gives you far more to go on than a single text.

Why does what I want change what I see?

Because wanting is a lens, not just a feeling. Emily Balcetis and David Dunning showed people an ambiguous figure that could be read as the letter B or the number 13.4 People tended to see whichever one led to the outcome they wanted, and they weren’t aware there was another way to see it.4

That was a lab task about juice and smoothies, not dating. But it’s a useful picture of what happens on the walk home from a good date. “We should do this again” can look like a plan. A slow reply can look like shyness. Neither reading is crazy. They’re just the version your hope prefers.

Fear bends things too, in the other direction. David Buss and Martie Haselton argued that people’s guesses about the other sex can lean in predictable directions, depending on which mistake would feel worse.5 In their studies, women tended to underestimate men’s commitment, and men tended to overread women’s sexual interest.5 Whatever you think of the evolutionary explanation, the takeaway is practical: your read can be tilted by what you’re scared of getting wrong.

What actually counts as a sign?

None of these prove anything about what’s in someone’s head. They’re things you can see, which makes them better than guesses.

• Plans with a day attached. “Thursday at 7?” beats “we should hang out sometime.”
• Follow-through. Plans that happen, and if one falls through, a counteroffer.
• Memory. They bring back something you said a week ago.
• Curiosity. They ask about your life, then ask a second question about the answer.
• Consistency over a few weeks. Roughly the same person on date four as on date one.

Notice what’s not on the list: how fast they text, how intense the first date felt, or a single sweet message at midnight. Those can be lovely. They’re just thin evidence.

What if I genuinely can’t tell?

That’s a real answer too, especially early on. Some people are hard to read, as the speed-dating clips showed.2 If you’re three or four dates in and still unsure, you’re allowed to ask a light, direct question: “I’m enjoying this. Are you?” It takes about four seconds, and it gives you better information than a week of decoding.

And keep the stakes in proportion. You’re not trying to find out if they’re the one. You’re trying to find out whether to go on another date.

What to do this week

• Keep two short lists after each date. “Shown” (only things they did) and “Hoped” (what you’d like it to mean). Read the first list when the second one gets loud.
• Make one plan yourself. Then see how they respond. Do they confirm, show up, suggest the next one?
• Count specific effort, not charm. One remembered detail outweighs a lot of general warmth.
• Assume you came across better than you think. The research suggests you probably did.1
• Ask, lightly, if you’ve been guessing for weeks. A plain question is kinder to both of you than a long silence.

If the worry about whether people like you is constant, keeps you up, or makes it hard to get through ordinary days, that’s worth bringing to a licensed mental-health professional. You deserve support with it, not just better guessing.

What doesn’t the research show?

• The liking-gap studies weren’t about dating. They were about conversations between strangers, workshop members and roommates.1
• The speed-dating studies used brief, staged dates, mostly with younger people, not three weeks of getting to know someone in your 30s.23
• The ambiguous-figure study was about perception in a lab, not romance.4
• The error-management work describes average tendencies and offers one evolutionary explanation that other researchers debate.5
• None of it can tell you what one particular person feels. Only time and a direct conversation can do that.