You leave the conversation and the audit starts. The thing you said that landed oddly. The pause where you should have had something. The story that went on too long.
By the time you get home you have concluded it went badly, and you file the person under someone who probably found you tedious.
It has a name and a number
Erica Boothby, Gus Cooney, Gillian Sandstrom and Margaret Clark (2018) ran a set of studies on exactly this. After conversations, they asked people how much they liked their partner and how much they thought their partner liked them.
People consistently underestimated how much they were liked and how much the other person had enjoyed the conversation. The researchers called it the liking gap, and the important detail is that it is one-directional. Nobody systematically overestimates. The error goes the same way for almost everyone, which means your read is not neutral information.
The conversation you are replaying and the conversation they are remembering are not the same conversation.
It also persists. It is not a first-meeting effect that resolves after a week.
Why your evidence feels so strong
Because you have access to a channel they do not: your own internal commentary. You heard the pause as agonising because you were inside it. They heard a pause.
You are also the only person in the room who remembers every alternative sentence you rejected. Your version of the conversation is the transcript plus the outtakes. Theirs is just the transcript.
What to do with it
Method at a glance
The Low-Stakes Reconnect
- Reference something specific from the last time you talked.
- Keep the ask small or nonexistent.
- Send it, even if it has been a long time. Especially if it has been a long time.
The reason the method is a message rather than a mindset exercise is that the gap does not close by being told about it. It closes by being contradicted.
People consistently underestimate how much someone will appreciate being reached out to, and the effect gets stronger, not weaker, the longer the silence has run.
Every message you send that gets a warm reply is one data point against a belief that has never been tested. Arguing with the feeling directly does not work. Collecting evidence does.
The specific thing matters
The second version proves you were paying attention, which is most of what makes a message land. The first could have been sent to anyone, and the recipient knows it.
Where this shows up
New jobs, especially the first fortnight. Parties where you knew one person. Any group you joined mid-way. And most reliably, the friendship that faded for no reason either party can name, which is very often two people each assuming the other let it go.
Where this tips into something worse
Two limits worth being straight about.
The liking gap is a finding about the average case, not a guarantee about every conversation. Some interactions genuinely go badly, and using the research to override a specific, well-founded read is its own error. The useful version is treating your post-conversation pessimism as unreliable evidence, not as automatically false.
And this is a finding about ordinary social self-doubt. Persistent, distressing social anxiety that limits your life is a different thing, and it responds to treatment that a communication book is not a substitute for. If that is closer to where you are, that is worth taking to a professional rather than to a technique.
Two things to try
Write the prediction down
Before you send a message, write what you expect to happen. Then compare. The gap between prediction and outcome is the thing that actually shifts the belief.
Stop auditing on the way home
The replay is not analysis, it is rehearsal of a conclusion you reached before the evidence. Notice it starting, and take the specific thing you enjoyed instead.
Where to start
Message one person this week who you assume has forgotten you. Reference one specific thing. Ask for nothing.
You will be wrong about how it lands, in a predictable direction. That is the entire finding, and it is one of the few in this book you can test in an afternoon.
Common questions
Why do I always think conversations went badly?
Because you have access to your own internal commentary and the other person does not. Research on the liking gap finds people systematically underestimate how much a conversation partner liked them and enjoyed the exchange, and the error runs one way for almost everyone (Boothby et al., 2018).
Does the liking gap go away over time?
Not quickly. It has been found to persist beyond first meetings rather than resolving after people get to know each other, which is part of why it is worth knowing about rather than waiting out.
How do I stop replaying conversations afterwards?
Treat the replay as rehearsal rather than analysis, since you reached the conclusion before you reviewed the evidence. The more effective intervention is collecting contrary data: send a message, predict the response in advance, and compare.
What if the conversation genuinely did go badly?
Sometimes it did, and the research describes the average case rather than every case. The useful version is treating your post-conversation pessimism as unreliable evidence, not as automatically wrong.
Sources
Boothby, E. J., Cooney, G., Sandstrom, G. M., & Clark, M. S. (2018). The liking gap in conversations: Do people like us more than we think? Psychological Science, 29(11), 1742–1756. · Levin, D. Z., Walter, J., & Murnighan, J. K. (2011). Dormant ties: The value of reconnecting. Organization Science, 22(4), 923–939. · Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social Psychology, 9(2, Pt. 2), 1–27.