Human Performance & Leadership
Your Meeting Rewards What Everyone Already Knows
The fact that would change your decision is usually held by one person, and discussion is what buries it.
2026-08-091 min read
Meetings feel productive in direct proportion to how much agreement they produce, which is exactly why they so often reach the wrong answer. A piece of information held by everyone in the room enters the conversation almost immediately, because any one of several people can raise it and each of the others can confirm it. A piece held by one person has a single chance of being spoken, no support waiting behind it, and a speaker who must risk being the only voice on it. Discussion is therefore not a pooling process. It is a sampling process weighted toward what the group already knew before anyone sat down.
Garold Stasser and William Titus demonstrated this in 1985 by rigging the arithmetic. They gave groups a choice among candidates and split the facts so that every member, reading alone, would favor an inferior candidate — while the group's combined files clearly favored a better one. Discussion should have assembled the full picture. Instead, groups talked mostly about the facts they held in common, confirmed each other, and chose the candidate the incomplete individual views pointed to. The information that would have changed the outcome was in the room the entire time and never reached the table.
Gigone and Hastie later sharpened the finding: the more members who already know a fact, the more weight the group's final judgment gives it, regardless of how much that fact actually distinguishes the options. Repetition reads as corroboration. A leader watching a discussion converge is watching redundancy accumulate, not evidence, and the confidence produced by that convergence is the least reliable signal in the room.
Key insights
Shared information has many possible speakers and instant confirmation waiting behind it; unique information has one speaker and none.
Repeating a fact raises the weight a group gives it independently of how much that fact actually discriminates between the options.
Fast consensus is evidence about the overlap in the room, not about the quality of the decision.
Why it matters
A decision that feels unanimous is often just the group repeating its overlap, which means the confidence you carry out of the room is not evidence about the decision.
The people most likely to hold the deciding fact — the newest hire, the lone specialist, the person closest to the customer — are structurally the least likely to get it heard.
Use this tomorrow
1Before your next decision meeting, ask each attendee to send you one fact they believe nobody else in the room has, then count how many of those facts actually get spoken aloud during the discussion.
2In the meeting after that, write down each substantive fact as it is raised and mark it new or a restatement; if restatements outnumber new facts, stop the open discussion and go around the table one person at a time.
Go deeper
The 1985 paper by Garold Stasser and William Titus, “Pooling of Unshared Information in Group Decision Making,” appeared in the Journal of Personality and Social Psychology. Its design is what makes the result hard to argue with: the better option was reachable only by combining what members held separately, so a group that missed it had failed specifically at pooling rather than at reasoning. Later work found the pattern holds across group sizes and task types, and that unshared facts, when mentioned at all, tend to be mentioned once and dropped rather than picked up and built on. Reviews of the accumulated studies conclude that groups solve these “hidden profile” problems far less often than the pooling assumption behind most meetings would require.
The effect is not a failure of honesty, and treating it as one makes it worse. Nobody is withholding anything — unique facts simply lose a race that shared facts are structurally favored to win, so a leader who responds by urging people to speak up has addressed motivation while leaving the structure untouched. The interventions with evidence behind them change who speaks when: assigning explicit topic ownership before discussion, so a member knows a domain is theirs to report on, or having members record their positions privately before anyone talks. Both work by giving unique information a scheduled slot instead of making it compete for an open floor.
