Mark ở vũ trụ khácVIMark học AI
Mark ở vũ trụ khác · Another Universe Common Myths · Teamwork · Creativity

Common Myths #8: Group brainstorming produces more ideas

vs

"When the whole team thinks together, you get more ideas, and better ones."

The mytha group meeting produces more and better ideas than its members thinking on their own
Spread byAlex Osborn, the book Applied Imagination, 1953
Key evidenceTaylor et al. (1958), Diehl and Stroebe (1987), Mullen et al. (1991)
Quick verdictwrong on the number of ideas; groups still help with pooling and choosing
One-minute recap

Essentials

A promise from 1953Osborn wrote that people think up twice as many ideas in a group. He never measured it.
The 1958 test: the reverse4 people working alone, then pooled, produced nearly twice the ideas of 4 people meeting together.
The culprit: waiting to speakOnly one person can talk at a time. Ideas of those waiting get forgotten or dropped.
Why we still believe itGroup sessions are fun and feel productive. People in groups rate themselves higher than they performed.
Don't cancel the meetingGroups still help to pool, build on and pick ideas. Better: write alone first, pool later.

Sources: Osborn, Applied Imagination (1953), as quoted by Taylor, Berry and Block (1958) and Diehl and Stroebe (1987); Diehl and Stroebe (1987); Stroebe, Diehl and Abakoumkin (1992); Paulus et al. (1993).

Part 1

The question behind the myth

If four people sit in a room calling out ideas, do they come up with more than the same four people writing alone and pooling the results? Almost 70 years of experiments say no. The full answer is more interesting: groups fail at generating ideas, and help at something else.

1953The promiseOsborn: groups produce twice as much. Four rules, the first being "no criticism".
→
1958–1991The testDozens of experiments: groups that meet lose to "nominal" groups (alone, then pooled).
→
TodayThe methodWrite alone first, pool later, choose by private vote. The meeting stays, the order changes.

Sources: milestones from Diehl and Stroebe (1987) and Mullen, Johnson and Salas (1991); see Part 8.

Part 2

Where the myth came from

An advertising man, a good process, and a promise no one measured.

How the claim changed over time. 1953, Osborn: the average person thinks up twice as many ideas in a group, an unmeasured claim. 1950s: brainstorming spreads through American companies. 1958: the first test at Yale finds the opposite. Today the claim reads: a whole-team session gives more and better ideas, as if self-evident. 1953 · Osborn"twice as many ideasin a group" claimed, not measured 1950sspreads to companies Osborn, 1963: spread"too fast", often misused 1958 · Yalealone, then pooled:nearly twice the group the first test today"team sessions givemore and better ideas" sounds self-evident The promise was about quantity. As it spread, "and better" crept in and "untested" fell away.
Osborn was a practitioner, not a researcher. He wrote from experience running the BBDO advertising agency. "Twice as many" was a claim in a book; five years passed before anyone measured it.Sources: the "twice as many" line in Osborn, Applied Imagination (1953), as quoted by Taylor, Berry and Block (1958) and Diehl and Stroebe (1987) (secondary; the book page was not checked); Osborn's 1963 remark as quoted in London Review of Education (UCL Press).
1939 · BBDO hits a rough patchCo-founder Alex Osborn looks for ways to get staff to "think up" ideas faster.
1942 · The booklet How to "Think Up"The first sketch of an idea meeting.
1953 · Applied ImaginationNames "brainstorming" and its four rules: no criticism, the wilder the better, quantity wanted, combine and improve on others' ideas.
1958 · Taylor, Berry and Block (Yale)The first test: 4 people alone, then pooled, beat 4 people meeting together.
1987–1991 · Finding out whyDiehl and Stroebe identify "waiting to speak"; a meta-analysis by Mullen and colleagues confirms groups lose on both quantity and quality.
2012 · Reaching the public"Groupthink" in The New Yorker calls brainstorming a "myth"; critics say it overstates the case. Most companies still meet the old way.

Sources: 1939 and 1942 from Regent University, Journal of Transformative Innovation, "The history of brainstorming: Alex Osborn" (secondary); the four rules from Applied Imagination (1953) via review literature; Taylor, Berry and Block, Administrative Science Quarterly 3 (1958); Diehl and Stroebe (1987); Mullen, Johnson and Salas (1991); J. Lehrer, The New Yorker, January 2012, and Scott Berkun's rebuttal (2012).

The grain of truth

A group of 4 meeting together does produce more ideas than one person alone (Taylor 1958 found that too). The mistake is the comparison: it should be 4 people working alone.

Feeling productive

The room is loud, ideas fly. It is hard to remember which ideas were yours, so you think you contributed more than you did.

Fun and bonding

A group session is nicer than facing a blank page alone. Enjoyable is easily mistaken for effective.

Sources: Taylor, Berry and Block (1958); Stroebe, Diehl and Abakoumkin, Personality and Social Psychology Bulletin 18 (1992); Paulus, Dzindolet, Poletes and Camacho, PSPB 19 (1993).

Part 1 post

Where the myth came from

Osborn 1953"twice as many"4 ruleswhy it's easy to believe
Part 2 post

What the evidence says

nominal groupswaiting to speakbigger groupsillusion
Part 3 post

Work lesson

write alone first30-minute meetingprivate voteMark's score
Part 3

What the evidence says

Five points, one diagram each.

The test: real groups vs "nominal" groups

real vs nominal groupsTaylor, Berry and Block, 1958
Two set-ups. Left, the real group: 4 people together, calling out ideas under Osborn's four rules. Right, the nominal group: 4 people apart, writing ideas, which are then pooled with duplicates counted once. Result: the nominal group produced nearly twice as many ideas; the real group only about half. real group: 4 people, 1 room ……… out loud, in turn, no criticism nominal group: 4 people, 4 desks write alone, then pool, drop duplicates ideas (nominal = 100%) real group ≈ half 100%
A fair test compares the same number of people for the same time. 12 four-person groups following Osborn's rules, against 48 individuals randomly combined into 12 nominal groups, on the same 3 problems, 12 minutes each. The nominal groups did better on total ideas, unique ideas and quality scores.Source: Taylor, Berry and Block, "Does group participation when using brainstorming facilitate or inhibit creative thinking?", Administrative Science Quarterly 3, pp. 23–47 (1958), via the abstract and the description in Diehl and Stroebe (1987). "About half" is how reviews summarise it; the original means were not available.

The main culprit: waiting to speak

production blockingDiehl and Stroebe, 1987, 1991
Four timelines, one per person. At any moment only one line has a "speaking" bar. On the other three, ideas that come up (dots) have to wait; several are crossed out because they were forgotten or dropped after hearing a similar idea. AnnBenCaraDan ▬ speaking (one at a time) ● idea while waiting✕ forgotten or dropped
Not laziness, a traffic jam. Diehl and Stroebe tested three explanations: free riding on others, fear of being judged, and waiting to speak. The first two had effects, but small ones. When they manipulated waiting to speak directly, it accounted for most of the loss.Sources: Diehl and Stroebe, "Productivity loss in brainstorming groups: toward the solution of a riddle", Journal of Personality and Social Psychology 53(3), pp. 497–509 (1987), Experiment 4, per the abstract; Diehl and Stroebe, JPSP 61, pp. 392–403 (1991). The figure is an illustration, not data.

Not a one-off study

replication and meta-analysisDiehl and Stroebe 1987, table 1; Mullen et al. 1991
22 squares for 22 experiments comparing real and nominal groups. 18 are filled: the real groups produced fewer ideas. The other 4 are left empty. 18 of 22 experiments:meeting groups produced fewer ideas 1991 meta-analysis: worse on bothquantity and quality each square = one experiment in the summary table of Diehl and Stroebe (1987)
The result replicates very consistently. The meta-analysis by Mullen, Johnson and Salas also found larger losses in bigger groups, when a facilitator or experimenter was present, and when ideas had to be spoken aloud.Sources: 18/22 from the summary table in Diehl and Stroebe (1987), as restated by Nijstad, Stroebe and Lodewijkx, European Journal of Social Psychology (2006) (secondary); Mullen, Johnson and Salas, "Productivity loss in brainstorming groups: a meta-analytic integration", Basic and Applied Social Psychology 12(1), pp. 3–23 (1991).

The illusion of group productivity

illusion of group productivityStroebe et al. 1992; Paulus et al. 1993
Two boxes. Feeling: people in groups are more satisfied with their performance. Reality: people working alone produce more ideas. In between: in a group it is hard to separate your ideas from others', and you notice less when you are stuck. how the meeting felt "we had loads of ideas"more satisfied than solo workers what the count shows people working alone producemore ideas once pooled ≠ because: hard to tell your ideas from others' · you notice less when stuck while others talk
This is why the myth survives. Decision makers sit in the meeting too, and they also feel it went well.Sources: Stroebe, Diehl and Abakoumkin, "The illusion of group effectivity", PSPB 18, pp. 643–650 (1992); Paulus, Dzindolet, Poletes and Camacho, PSPB 19, pp. 78–89 (1993); the "noticing when stuck" explanation from Nijstad et al. (2006), via BPS Research Digest (February 2006).

The way out: think alone first, pool later

hybrid process · brainwritingGirotra et al. 2010; Rietzschel et al. 2006
Two processes. Team from the start: the group thinks together, then chooses. Hybrid: each person thinks alone, then the group pools, then chooses. In an experiment with 44 students the hybrid process produced about 3 times as many ideas in the same time, with higher average quality and a better pick of the best idea. Caveat: generating many ideas is not the same as picking the right one; every group is weak at choosing. team from start whole group thinks together choose hybrid think alone pool, build choose ≈ 3× the ideasin the same time but: more ideas ≠ picking the right one Rietzschel 2006: chosen ideas no better; choice not significantly above chance
Groups are not useless, just used in the wrong order. The hybrid process was also better on the best idea, which is what companies actually need. But choosing is a shared weakness, so Part 6 uses a private vote and criteria written in advance.Sources: Girotra, Terwiesch and Ulrich, "Idea generation and the quality of the best idea", Management Science 56(4), pp. 591–605 (2010) (44 students, per Knowledge at Wharton; "about 3 times" per the 2009 INSEAD working paper, via summaries); Rietzschel, Nijstad and Stroebe, Journal of Experimental Social Psychology 42, pp. 244–251 (2006), abstract.
Part 4

Fact check: Solid, Caution, False

12 things people say about brainstorming, checked against the research. Click each line for the reasoning.

SOLIDGroups that meet produce fewer ideas than the same number of people working alone, then pooled

From Taylor (1958) to the 22-experiment table of Diehl and Stroebe (1987, meeting groups lost in 18 of 22) and the meta-analysis by Mullen et al. (1991). One of the most consistently replicated results in group psychology.

SOLIDThe bigger the group, the bigger the loss

Bouchard and Hare (1970) compared groups of 5, 7 and 9: the gap to nominal groups widened. Mullen et al. (1991) also found group size increased the loss.

SOLIDThe main reason is waiting your turn to speak

Diehl and Stroebe (1987, Experiment 4; 1991) manipulated turn-taking directly and found it explained most of the loss. Nijstad et al. (2003) found long waits disrupt the flow of thinking.

SOLIDPeople in groups think they performed better than they did

Stroebe, Diehl and Abakoumkin (1992); Paulus et al. (1993): group members are more satisfied with their performance, even though those working alone have more reason to be.

CAUTION"More people produce more ideas than one person"

True, and this is the grain of truth in the myth: a group of 4 meeting together still beats one person alone (Taylor 1958). But it is the wrong comparison. The right question: with the same 4 people, which method yields more?

CAUTION"So thinking alone is always best"

Don't fix one myth with another. Working alone yields more ideas, but Rietzschel et al. (2006) found nominal groups choose no better. The hybrid process (Girotra 2010) and writing-then-passing slips (Paulus and Yang 2000) sometimes beat pure solo work, though not every study finds this.

CAUTION"Brainstorming is useless, cancel the meetings"

Sutton and Hargadon (1996) followed the design firm IDEO for over a year: its brainstorms also kept a shared memory of solutions, built skills and impressed clients. Lab studies only count ideas in a short session (12 minutes per problem in Taylor 1958); companies need more than that.

CAUTION"Brainstorm online, type ideas in, problem solved"

Typing in parallel means no one waits their turn. A 2007 meta-analysis of electronic brainstorming (usually cited as DeRosa, Smith and Hantula) found large electronic groups beat nominal groups, but for small groups working alone still won. Dennis and Valacich (1999) called it a new "illusion of productivity".

CAUTION"Banning criticism is what produces more ideas"

Osborn's first rule. Nemeth et al. (2004), in the US and France, found that telling groups "debate, even criticise" worked better than "don't criticise". The evidence is thin; treat it as an open question, not a new rule.

FALSE"Osborn proved groups produce twice as many ideas"

It was a claim in a 1953 book, based on advertising experience, with no comparative experiment. The first test (1958) found the opposite. Osborn himself later wrote that the method spread "too fast" and was often misused.

FALSE"Fewer ideas, but the group's ideas are better"

Mullen et al. (1991): meeting groups lost on quality too. Taylor (1958) also found nominal groups ahead on quality scores; once total ideas were accounted for, that edge disappeared, so real groups do not make up for it in quality.

FALSE"The more people you invite, the more ideas"

For groups talking aloud, the opposite: every extra person is one more person waiting to speak (Bouchard and Hare 1970; Mullen et al. 1991). The one notable exception is large groups typing ideas on computers.

Sources: as given in each line; full list in Part 8. Several studies could only be read via abstracts or reviews, as noted in Part 8.

Part 5

Verdict on the myth

Scores are Mark's assessment based on the sources in Part 8, not data from any study.

How wrong
4/5
How persistent
5/5
How harmful
2/5
What the myth gets right
  • A group beats one person alone
  • Hearing others' ideas can spark new ones
  • Meetings help pool, build on ideas and share ownership
What it gets wrong
  • Loses to the same people working alone, on quantity and quality
  • Bigger groups lose more
  • Feeling productive is not being productive

Wrong exactly where it is proudest, generating ideas; right where few people look, pooling and choosing. The harm is modest: wasted time and lost ideas from the quieter people.

More: why "persistent" scores 5/5

The contrary evidence dates from 1958, has been replicated dozens of times, was meta-analysed in 1991 and reached major media in 2012. Yet "let's all brainstorm" is still how countless meetings begin.

One reason is the illusion of productivity: the person who called the meeting also feels it went well. Another: group sessions have real benefits (bonding, shared understanding, shared ownership), just not the one the myth promises.

Part 6

Work lesson: write alone first, pool later

A script for a 30-minute idea meeting for 4–8 people. You need: sticky notes (or one shared document), pens, a timer.

The 30-minute script. Before: send one single question. Minutes 0 to 3: state the question and rules. Minutes 3 to 11: write alone, in silence, one idea per note. Minutes 11 to 17: pass the notes, read others' ideas, add building ideas. Minutes 17 to 23: stick them on the wall in clusters, merge duplicates. Minutes 23 to 27: private dot vote, 3 dots each. Minutes 27 to 30: pick 1 to 3 ideas, assign an owner and a deadline. before 0–3' 3–11' 11–17' 17–23' 23–27' 27–30' 1 questionsent ahead rules"silence" write aloneone idea per noteno talking pass notesread, addbuilding ideas stick, mergedrop duplicatesnow talk private vote3 dotseach decideowner,deadline nobody waits to speak what groups do well: pool, build, choose Change the order, don't cancel the meeting: think alone in the first half, talk in the second.
Each block patches a weakness the research found. Silent writing avoids waiting to speak; passing notes keeps the "others' ideas spark mine" benefit (in the style of Bernd Rohrbach's 6-3-5 method, 1969); reading ideas one by one and voting privately are steps of the Nominal Group Technique (Delbecq and Van de Ven, 1971); criteria are written in advance because every group is weak at choosing.Source: the timings are Mark's suggestion, not research data. Basis: Diehl and Stroebe (1987); Paulus and Yang (2000); Rietzschel et al. (2006); Girotra et al. (2010).
One question, sent a day ahead"How can we halve our response time to customers?" beats "ideas for customer service".
The boss writes last, speaks lastThe most senior person's ideas easily anchor the room. When reading ideas aloud, go round the table, boss last.
Write the criteria before votingFor example: testable within 2 weeks, no extra headcount. Vote on the criteria, not on who spoke best.
Online: use a shared documentEveryone types in their own column for 8 minutes, mics off. Typing in parallel means nobody waits their turn.
Invitation template (copy and use)

"Thursday 10:00, 30 minutes. Question: [one sentence]. For the first 8 minutes everyone writes alone, no discussion; then we pool and choose. If you have ideas already, jot them down and bring them."

Afterwards: send a photo of the sticky-note wall, the 1–3 chosen ideas, the owner and the deadline. Keep the ideas that weren't chosen in a shared list.

Groups are bad at thinking at the same time. Groups are good at listening, building on ideas and choosing together. Give each job its own slot.Mark, summarising the studies in Part 3.
Part 7

Further reading

Diehl and Stroebe, "Productivity loss in brainstorming groups: toward the solution of a riddle" (1987)The original paper that found the "waiting to speak" culprit. Four experiments that read like a detective story.
Stroebe, Nijstad and Rietzschel, "Beyond productivity loss in brainstorming groups: the evolution of a question", Advances in Experimental Social Psychology 43 (2010)A review of 50 years of research, from Osborn to current practice.
Sutton and Hargadon, "Brainstorming groups in context" (1996)The other side of the story: how brainstorming at IDEO pays off beyond idea counts.
Common Myths #09: Memory works like a video cameraNext up: why everyone remembers a meeting differently, and why to write decisions down right away.
Part 8

Sources

Original papers first. Note: full texts could not be downloaded while writing this piece; facts come from abstracts and from reviews restating the original studies (secondary sources), as marked next to each number.

  1. A. F. Osborn, Applied Imagination, Scribner, 1953 (revised 1963). The "twice as many" line via secondary quotation.
  2. D. W. Taylor, P. C. Berry and C. H. Block, "Does group participation when using brainstorming facilitate or inhibit creative thinking?", Administrative Science Quarterly 3, pp. 23–47, 1958. semanticscholar.org
  3. T. J. Bouchard and M. Hare, "Size, performance, and potential in brainstorming groups", Journal of Applied Psychology 54, 1970.
  4. M. Diehl and W. Stroebe, "Productivity loss in brainstorming groups: toward the solution of a riddle", JPSP 53(3), pp. 497–509, 1987; and "Productivity loss in idea-generating groups: tracking down the blocking effect", JPSP 61, pp. 392–403, 1991.
  5. B. Mullen, C. Johnson and E. Salas, "Productivity loss in brainstorming groups: a meta-analytic integration", Basic and Applied Social Psychology 12(1), pp. 3–23, 1991. doi:10.1207/s15324834basp1201_1
  6. W. Stroebe, M. Diehl and G. Abakoumkin, "The illusion of group effectivity", PSPB 18, pp. 643–650, 1992. sagepub.com
  7. P. B. Paulus, M. T. Dzindolet, G. Poletes and L. M. Camacho, "Perception of performance in group brainstorming: the illusion of group productivity", PSPB 19, pp. 78–89, 1993.
  8. R. I. Sutton and A. Hargadon, "Brainstorming groups in context: effectiveness in a product design firm", Administrative Science Quarterly 41(4), pp. 685–718, 1996.
  9. A. R. Dennis and J. S. Valacich, "Electronic brainstorming: the illusion of productivity", Information Systems Research 10(2), pp. 110–133, 1999.
  10. P. B. Paulus and H.-C. Yang, "Idea generation in groups: a basis for creativity in organizations", OBHDP 82(1), pp. 76–87, 2000 (via secondary summary).
  11. C. Nemeth, B. Personnaz, M. Personnaz and J. Goncalo, "The liberating role of conflict in group creativity", European Journal of Social Psychology 34, pp. 365–374, 2004.
  12. B. A. Nijstad, W. Stroebe and H. F. M. Lodewijkx, "Production blocking and idea generation: does blocking interfere with cognitive processes?", Journal of Experimental Social Psychology 39, 2003.
  13. E. F. Rietzschel, B. A. Nijstad and W. Stroebe, "Productivity is not enough", Journal of Experimental Social Psychology 42, pp. 244–251, 2006.
  14. B. A. Nijstad, W. Stroebe and H. F. M. Lodewijkx, "The illusion of group productivity: a reduction of failures explanation", European Journal of Social Psychology, 2006 (source of the 18/22 figure).
  15. D. M. DeRosa, C. L. Smith and D. A. Hantula, "The medium matters: mining the long-promised merit of group interaction in creative idea generation tasks in a meta-analysis of the electronic group brainstorming literature", Computers in Human Behavior 23, 2007 (authors as usually cited, not verified).
  16. K. Girotra, C. Terwiesch and K. T. Ulrich, "Idea generation and the quality of the best idea", Management Science 56(4), pp. 591–605, 2010. Knowledge at Wharton
  17. R. Korde and P. B. Paulus, "Alternating individual and group idea generation: finding the elusive synergy", JESP 70, pp. 177–190, 2017.
  18. A. L. Delbecq and A. H. Van de Ven, "A group process model for problem identification and program planning", Journal of Applied Behavioral Science 7, 1971 (Nominal Group Technique); B. Rohrbach, the 6-3-5 method, 1969.
  19. History: Regent University, Journal of Transformative Innovation, "The history of brainstorming: Alex Osborn"; Buffalo State, "Alex F. Osborn" page; J. Lehrer, "Groupthink", The New Yorker, 30 January 2012; C. Jarrett, BPS Research Digest, February 2006.