You get home from a dinner with someone new. On the way back you replay it: the joke that landed flat, the story that ran long, the pause after you asked about their work. By the time you’re at the door, you’ve decided they found you boring.
This post is for that replay. It explains the liking gap, a pattern found in five studies on conversations between new acquaintances. It covers what the researchers measured, why the replay runs harsher than the talk, and what to do the next day.
The short answer
After a conversation, people tend to believe the other person liked them less than they did. Erica Boothby, Gus Cooney, Gillian Sandstrom and Margaret Clark called this the liking gap. They found it between strangers in a lab, among first-year college students in a shared dorm suite, and among adults at workshops in the UK (Boothby et al. 2018).
The reason they point to sits inside your own head. Your thoughts about how you came across are more critical than the other person’s thoughts about you. You then use those thoughts as your best guess of what they think.
What the studies measured
The first study was small and plain. 36 people came to a lab at Yale in pairs and talked for about 5 minutes, with a sheet of icebreaker questions. Afterward, each person rated how much they liked their partner and how much they thought their partner liked them, on a 7-point scale. On average they rated their liking of the partner at 5.82 and the partner’s liking of them at 5.17 (Boothby et al. 2018). Both can’t be true at once, so the difference is a mistake in the guess.
The mistake was larger for shyer people. Participants low in shyness showed no significant gap. Participants of average shyness showed one, and those high in shyness showed the largest.
Length didn’t close it either. In a third study, pairs talked as long as they wanted, from 2 to 45 minutes. The gap held across short, medium and long conversations. People also thought their partner enjoyed the talk less than the partner said they did.
The signals were there
Two trained coders, who didn’t know the hypothesis, watched videos of the lab conversations. Their ratings tracked how much each person liked their partner. The participants’ own guesses didn’t track those signals (Boothby et al. 2018). The partner showed interest. The person sitting next to them missed it.
The researchers then asked people to write down their most salient thoughts after a talk. Thoughts about how the partner saw them came out more negative than thoughts about how they saw the partner. That difference went along with how little they believed they were liked.
The paper’s discussion gives a few reasons this happens. You know your ideal version of a story, so the told version falls short; the listener only hears the told version. You feel nervous and assume it shows, while the other person sees your behavior, which tends to be smooth. And replaying mistakes helps you improve next time, so the mind keeps doing it.
A later paper with Boothby as a co-author points the same way. Across three studies, people rated their conversation ability more pessimistically than their skill at other common activities. They also blamed themselves more than their partner for a conversation’s low points (Welker et al. 2023).
How long the gap lasts
The longest study followed 102 first-year students at Yale who shared dorm suites. They rated their suite mates five times over one academic year: September, October, December, February and May. The gap showed up at the first four points and was gone by May (Boothby et al. 2018).
The authors offer two possible reasons for May. By then the students knew each other well. They were also deciding who to live with the next year, and those talks may have made the liking clear. That’s their guess, and the study doesn’t test it. What the data show is the gap fading over a year of living side by side.
Where the finding stops
These studies show an average pattern. They don’t tell you what one specific person thought of you last Tuesday. The lab conversations were short, and the participants were students, lab volunteers or workshop guests meeting someone new. Close friends and family weren’t the subject. One pair in the first study was excluded because they turned out to be close friends.
So the finding is a reason to doubt the harshest version of the replay. It’s no guarantee about any one evening.
Eight questions for the replay
The replay is easier to check on paper. These questions move attention from your own performance to what the other person did. Answer them the next morning, when the first wave of the replay has passed.
- What did they ask you that they didn’t have to ask?
- When did they laugh or lean in?
- What did they tell you about themselves?
- Did the talk run longer than it had to?
- Did they mention meeting again, even loosely?
- Which moment are you replaying, and how long did it last?
- What would you think if a friend told you this same story about their evening?
- What did you like about them?
The last question matters. In the lab videos, coders could see how much each person liked their partner. What you liked about them may have shown, and the same goes the other way.
Three cards for the morning after
Three words in the Thesis deck sit close to the replay. Each card has an idea on one side and an everyday thing on the other.
Belonging has Headphones on its other side. The replay can feel like a worry about belonging: am I in, or still at the edge? Choose the idea to talk about a group where you felt half in. Choose the Headphones for a story about the moments you tuned out the room, or the room tuned out you.
Shame has a Letter on its other side. The replay can land on one sentence you wish you could take back. The Letter side lets the story come in sideways: a message you rewrote again and again, a note you never sent. If the word sits too close, choose the Letter, or pass.
Friendship has a Ball on its other side. The dorm study tracked new relationships forming over a year. The idea side can bring up how yours started, and the Ball brings up the games, teams and parks where it happened.
Try one word with someone
A conversation is also a way to check the replay. Talk with a friend about a time you misread how someone saw you, and they may tell you about theirs. The Thesis die gives six angles on one word: MEMORIES · BODY · PEOPLE · VALUES · CHILDHOOD · FUTURE. “Look at the word from that side.”
The rules start with a pass agreement: “Before you start, agree together how passing works.” The rule card also sets the tone for anyone who worries about saying too little: “A short answer is a full answer. You can always ask for more.” The three follow-ups fit every word: What did that feel like? Who else was there? Has it always meant that to you?
The game ends with one last round: “each player names one thing they discovered about someone at the table today.” For someone who replays every talk, that round is useful. You hear out loud what another person noticed about you.
What to do today
Pick the conversation you’re replaying. Write down one thing the other person did that showed interest: a question, a laugh, a story they didn’t have to share. Then send them a short message about something they said. The dorm study’s gap closed over a year of seeing each other, and a second conversation is where you start.
See the rules and the box, or open the free online table. The cards above have their own pages: Belonging, Shame and Friendship. For playing with a few close people, see Thesis for friends. For the studies behind the deck, see the research page. If the worry starts before the talk, see why deep conversations feel awkward. For the fear of having said too much, see the vulnerability hangover. For curiosity as a skill, see how to be more curious about people.