Ask everyone who was in yesterday's meeting to estimate what percentage of the talking they did. Add the numbers up. They will not total 100. They will total something closer to 140, or 180, and no one in that room will be lying.
This is not a story about egos. It's a story about memory. And once you see the mechanism, you will never again fully trust your own account of who carried a conversation — including, and especially, the account where you carried it.
The receipts you keep are only your own
In 1979, the psychologists Michael Ross and Fiore Sicoly asked married couples, separately, to estimate their share of responsibility for household tasks: cleaning, making breakfast, planning leisure activities, causing arguments. When the researchers added each couple's two answers together, the totals overshot 100 percent — over and over. Both partners believed they were doing more than half. Both were sincere.
Ross and Sicoly's explanation is the least glamorous one available, which is usually a sign it's right. It isn't self-love. It's availability — the tendency to judge how often something happens by how easily examples come to mind. You have vivid, first-person, time-stamped memories of every dish you washed. You have almost no memories of the dishes your partner washed while you were in another room. When you try to weigh the two piles, one is stacked in front of you and the other is a rumor.
A meeting is a household with more people and worse lighting.
You remember the point you made about the Q3 dependency because you remember composing it, waiting for a gap, and the small physical jolt of speaking. You remember the two comments you decided not to make — those feel like contributions too, restraint being a kind of labor. You remember thinking hard during the silences. What you do not remember, with anything like the same texture, is Priya's remark that reframed the whole problem, because when she said it you were half-listening and half-rehearsing.
So you leave with a ledger. Everyone leaves with a ledger. And every ledger is honest, and none of them reconcile.
The bias gets worse as the room gets bigger
Here is the part that should genuinely alarm anyone who runs meetings: this distortion does not stay constant as you add people. It compounds.
Research on responsibility allocation — work by Eugene Caruso, Nicholas Epley, Juliana Schroeder and colleagues — has found that the sum of claimed credit inflates as group size grows. In a group of three, the overshoot is modest. In a group of ten, the collective claim can become absurd, because each additional person adds one more full ledger of vividly remembered personal effort and one more set of contributions that everyone else can only dimly reconstruct. The math of memory is not the math of arithmetic. Each person's view is nearly complete about themselves and nearly empty about the other nine.
That same body of work identified something even less comfortable: people overclaim more readily for the good outcomes. Ask who drove a successful project and the percentages balloon. Ask who caused the delay and they shrink into the passive voice. Availability alone doesn't explain that asymmetry; motivation does its own quiet work on top. But you don't need the motivational layer to get the effect. Pure, blameless memory produces most of it.
What this actually costs you
Start with airtime. The person who dominates a meeting almost never experiences themselves as dominating it, because their internal record is full of the moments they held back — the thought they swallowed, the tangent they chose not to chase. Their restraint is vivid to them and invisible to everyone else. Meanwhile the quiet person's silence is deafening to them, and they leave convinced they said less than they did. Nobody in the room is calibrated. Telling a dominant talker to "give others space" fails not because they resist the feedback but because it contradicts a memory they can actually replay.
Then credit. Two people finish a project genuinely believing they did 70 percent of it. Both are hurt when the other is praised. Neither is being petty; each is looking at a set of memories the other person will never have access to. This is how teams that like each other quietly corrode.
And then decisions. Egocentric bias means you overweight the ideas you can most easily recall, which are the ideas you generated. You will defend your own proposal in a retrospective months later, still able to feel the shape of the thought, while the alternative someone else offered has degraded into "something about latency, I think." The evidence isn't balanced. It never was.
The fix is not humility. It's unpacking.
Telling people to be more modest doesn't work, and it shouldn't — you're asking them to distrust an accurate memory of a real contribution. The intervention that does work goes after the availability gap directly.
Savitsky, Van Boven, Epley and Wight found that when people were asked to first list each other group member's specific contributions before estimating their own share, the overclaiming shrank. Not because anyone became more generous, but because the recall task loaded the other ledgers into working memory. You cannot weigh two piles until both are in front of you. The bias is a retrieval failure, so the cure is a retrieval prompt.
This generalizes beautifully. Anything that turns other people's contributions from a rumor into a record — a written trail, a named attribution, a five-minute recall exercise — pulls estimates back toward reality. Anything that leaves each person alone with their own memory lets the ledgers drift apart until they can't be reconciled without an argument.
Your next moves
- Run the unpacking prompt before you assign credit. Before writing a project summary, performance note, or shoutout, spend three minutes listing every specific thing each other person did — by name, with the actual contribution. Do it before you write a single word about yourself. If your estimate of your own share doesn't drop, you didn't list enough.
- Do the 100-point test on your next meeting. Right after it ends, privately write your guess at what percentage of the talking you did. Then check it against the recording or transcript. Almost everyone is wrong in the same direction. One calibration is enough to permanently change how you sit in a room.
- Assign a named attribution line in your meeting notes. Not "we decided" — "Priya proposed X; Marcus raised the cost objection; we went with X." Ten extra seconds of typing creates the only durable record of who contributed what, and it's the record that will exist in six months when memories have finished editing themselves.
- In any group over five, assume the credit ledger is inflated and act accordingly. Before you feel slighted about recognition, remember that the person you feel slighted by is, statistically, feeling slighted by you. Say it out loud. It defuses more resentment than it should.
- Ask the quietest person to summarize the meeting. They heard the most, because they weren't rehearsing. Their recall of others' contributions is the least corrupted in the room, and giving them the summary is the cheapest way to correct the record.
The record you don't have to remember
The deepest thing about egocentric bias is that it's not a flaw of character — it's the price of being one person and not ten. You get a full-fidelity recording of exactly one participant, and you're asked to judge a group. Of course you get it wrong. Everyone gets it wrong, symmetrically, which is why the numbers add to 180 and why nobody can talk anyone else out of their share.
The only real answer is a record outside your head. That's the whole idea behind MeetingMortem: a fast, structured postmortem for the meeting you just left, capturing what was actually said and who actually said it, while it's still recoverable — before your memory quietly promotes you to the main character. Your ledger will always favor you. It helps to keep one that doesn't. Take a look at MeetingMortem →