Structured round-robin beats free discussion at pooling unique information

What it is

To get a group to pour out what each person knows, structured round-robin (a fixed order in which everyone takes an uninterrupted turn) reliably outperforms free discussion. The finding was established as the brainstorming "productivity loss" problem: free oral discussion yields less than the same people working alone, pooled. Of the classic losses, the hardest is production blocking — only one person can speak at a time, so everyone else is either waiting (ideas decay in the queue) or contending (interruptions and talking over). Add evaluation apprehension (holding back half-formed ideas for fear of judgment) and free-riding, and free discussion structurally compresses information output — with unique information, the most fragile kind, cut first.

Why it happens

The round-robin structure neutralizes each loss in turn. A fixed order with an uninterruptible speaking slot converts floor rights from "contend to speak" to "your turn is guaranteed," so holders no longer keep ideas alive while queued (blocking dissolves in the guarantee). An explicitly assigned duty to contribute reframes "stating my information" from self-display into fulfilling an obligation — everyone must speak, so speaking is no longer showing off (the social basis of apprehension is pulled away). As for free-riding, a named order makes absence visible at a glance. The structure's costs are equally clear: round-robin is serial, so total speech per unit time is below free-for-all discussion, and it feels ceremonial and slow. That is why it fits one specific phase — information pooling, where the goal is surfacing all unique information, not the efficiency of argument.

Studying it

• Paradigm: the classic decomposition experiments of productivity loss — comparing nominal groups (the same people working alone, outputs pooled) against interacting groups under free discussion, round-robin, and parallel writing (brainwriting), on output quantity and originality. This paradigm underlies process evaluations in groupware research.
• Variables: speaking structure (free / round-robin / parallel writing), group size, anonymity, and task type (generation / judgment) as independent variables; output quantity, unique-information entry rate, and output quality as dependent variables.
• Use in interface research: choosing structures for solicitation features (comments, feedback collection, retrospective forms) — testing the share and omission rate of unique content under different submission structures.
• Methodological caveat: output quantity is the most stable dependent variable in this literature, but its correlation with quality varies by task. The nominal-group baseline is inflated in real teams whose members' knowledge overlaps heavily ("alone, they'd write the same things anyway"), so lab gap magnitudes do not transfer directly.

Where it stops holding

Structure pays only in the pooling phase. Once the group moves to debate, negotiation, and improvised follow-up questions, the seriality of round-robin becomes pure loss — interruptions and latching-on in natural conversation are precisely what carries rapid clarification. In well-drilled small teams the three losses are already solved internally by tacit understanding and trust (floor rights flow naturally, evaluative safety is established), and imposing round-robin breaks the working rhythm; the right adaptation there is keeping only a "no-omission checkpoint." The upward-scaling answer is not longer rounds but parallel written collection (everyone writes simultaneously, then aggregate), which removes the serial bottleneck entirely.

Applying it

• Label the segments that need unique input as an explicit "pooling phase" and run them as round-robin or parallel writing, separated in time from the "discussion phase."
• For parallel written collection use independent entry fields (mutually invisible until submission), then aggregate into discussion.
• The facilitator keeps a checklist for the round and ticks off names live, making absence visible.
• Small acquainted teams can use the lightweight substitute: after open discussion, one fixed lap of "each person adds one thing nobody has mentioned" — gap-checking without reordering the whole session.
• Verification: for the same issue type, compare the number of unique information items between "discuss first, then solicit" and "solicit first, then discuss"; if the former yields significantly fewer, the discussion structure is suppressing unique information and the pooling phase needs to be isolated up front.

Related

• Same group: V9.02.1 Group discussion rehashes shared information and ignores what only individuals know · V9.02.2 Holders of unique information usually speak only when explicitly invited · V9.02.3 The first speaker's frame bounds the whole discussion · V9.02.5 Records capture only what was said; omissions stay invisible
• Nearby: V9.03 Groupthink · V1.07 Effects of Group Size
• Search terms: production blocking · nominal group technique · brainwriting