The Meeting Is The Problem

You have a problem that needs ideas, so you put the people who might have them in a room. That instinct is roughly seventy years old, it comes from a specific book, and the empirical literature has been contradicting it since the late 1950s.

Our Grade For This Claim

Applying the scheme from the first article in this series.

This is Grade A, and it is the strongest evidentiary position of anything this series has covered.

Four reasons. The comparison has been run repeatedly since 1958 and the direction is described as holding in nearly all such studies[2]. It has been meta-analysed across 20 studies with large effects[2]. The mechanism was isolated experimentally across four studies that tested and eliminated the competing explanations[3]. And the finding is counterintuitive and inconvenient, which means it survived despite nobody wanting it.

The one caution: the bulk of this evidence is laboratory work on idea-generation tasks. Whether the same magnitudes apply to a commercial meeting about a real decision is a question the literature we obtained does not answer directly.

A Note On Method

Everything here is verified to August 2026.

We obtained the abstract of the 1987 four-experiment paper and a hosted copy of parts of it[3][4], together with several peer-reviewed reviews that describe the literature[1][2][5].

We did not obtain the 1991 meta-analysis directly. Its effect sizes reach us through a peer-reviewed review that reports them[2], and its study count through a second[1].

We did not obtain Osborn (1957), and quote his claim as reproduced with a page reference in a peer-reviewed review[1].

Where we convert a reported correlation into a different effect size metric, or compute the arithmetic of meeting time, we say so and the figure is ours.

This article reviews organisational research. It is not management, human resources or legal advice.

The Claim That Started It

Where the practice comes from.

Brainstorming as a formal technique originates with Alex Osborn's Applied Imagination[6]. Osborn set out rules for the sessions, and claimed that if those rules were adhered to, the average person can think up twice as many ideas when working with a group than when working alone[1].

Three observations, ours.

The claim is specific and quantitative. Twice as many. That is a testable proposition rather than a vague endorsement of collaboration, and it is to Osborn's credit that he made it falsifiable.

It is also conditional on the rules: defer judgment, aim for quantity, welcome unusual ideas, build on others. So a defender can always argue a failed session was run badly, and that defence has been available for seventy years.

And the underlying intuition is genuinely plausible. One person's idea should spark another's. That mechanism is real; the question the evidence answers is whether it outweighs the costs of being in a room together.

What A Nominal Group Is

The experimental design, which is elegant and worth understanding precisely.

In these studies, face-to-face groups of various sizes are compared to nominal groups, which are formed by aggregating the output of a comparable number of individuals working separately[2]. A review adds the crucial detail that the separate individuals' ideas are combined into a group product by the experimenter, with ideas mentioned several times counted only once[1].

Three features make this a fair test, and this assessment is ours.

The number of people is matched. Six working together against six working separately, not six against one.

The duplicates are removed from the nominal group's output. This is the design's most important feature, because without it the separated group would win trivially by everyone independently producing the same obvious ideas.

And the nominal group receives no benefit of any kind from interaction. They do not see each other's work. So the comparison isolates exactly what being in the room adds or subtracts.

The primary outcome measure is the number of non-redundant ideas an individual or group produces[2], with quality measured separately.

The Meta-Analysis

The aggregate result.

A review reports that the outcome in nearly all such studies is that nominal groups outperform face-to-face groups in terms of the production of non-redundant ideas and idea quality[2].

Mullen, Johnson and Salas published Productivity loss in brainstorming groups: A meta-analytic integration in Basic and Applied Social Psychology, 12, 3–23[6]. A review describes it as comparing face-to-face and nominal groups from 1958 to 1990, and finding the mean effect sizes from the experimental literature to be large: r = .57 for the number of non-redundant ideas, and r = .56 for idea quality[2].

A second review, describing the same meta-analysis as covering 20 brainstorming studies, records the conclusion that nominal groups do not only produce substantially more non-redundant ideas than interactive groups, but they also produce a substantially greater number of high-quality ideas, and further that the productivity loss of interactive groups as compared to nominal groups increases rapidly with group size[1].

Two observations, ours.

The quality result matters more than the quantity result. A sceptic can dismiss idea counts as measuring noise. Idea quality is a separate measure and it moved the same way, by almost the same amount.

And the same review notes these results contradict the popular, but poorly substantiated notion that communication among individuals will result in synergistic effects[2].

How Large That Is

Putting the effect in context, with our own conversion.

Applying the standard transformation from a correlation to a standardised mean difference, r = .57 corresponds to approximately d = 1.39, and r = .56 to approximately d = 1.35. These conversions are ours, not figures from any paper.

Set that against what this series has reported elsewhere.

Feedback interventions: d = 0.41 on average.

The disputed nudge headline: d = 0.43.

Choice overload, meta-analytic mean: d = 0.02.

Ego depletion after bias correction: approximately zero.

Two observations, ours.

The brainstorming penalty is roughly three times larger than any positive effect this series has reported, and it points in the opposite direction from what practitioners believe.

We should apply our own halving rule for good order. Even halved, this remains substantially larger than anything else covered here, and the direction is not in dispute in any source we located.

Three Suspects

Having established the loss, the field went looking for its cause.

Three candidate explanations were on the table[3][5].

Free riding. The tendency for members to free ride on the efforts of others because their contributions were less identifiable and more dispensable in interacting groups than in nominal groups[5].

Evaluation apprehension. The idea that individuals felt inhibited from sharing ideas in a group setting due to concerns about negative evaluations, despite instructions to withhold criticism[5]. Social inhibition is reported to be greater where more group members perceive others as experts[7].

Production blocking, being the individual's inability to spontaneously interject ideas without violating group etiquette or breaking the concentration of other members[2].

Two observations, ours.

The first two are motivational and social. They say people in groups try less hard or hold back. Both have obvious managerial fixes: make contributions identifiable, make the room safer.

The third is mechanical. It says nothing about motivation or courage. It says that only one person can talk at a time.

Ruling Two Of Them Out

What the four experiments found, and this is unusually clean experimental work.

Diehl and Stroebe conducted four experiments to investigate free riding, evaluation apprehension, and production blocking as explanations[3].

On free riding. In Experiment 1 they manipulated assessment expectations. Productivity was higher under personal than collective assessment instructions, but type of session still had a major impact on brainstorming productivity under conditions that eliminated the temptation to free ride[3].

On evaluation apprehension. Experiment 2 demonstrated that inducing evaluation apprehension reduced productivity in individual brainstorming. But the failure to find an interaction between evaluation apprehension and type of session in Experiment 3 raises doubts about evaluation apprehension as a major explanation of the group loss[3].

Two observations, ours, and the second is the elegant part.

Removing the free-riding incentive did not remove the group penalty. So free riding is not the cause, whatever else it does.

And the evaluation apprehension result is the sharper piece of reasoning. Apprehension does reduce productivity, but it reduces it in people working alone too. For it to explain why groups do worse than individuals, it would have to affect groups more, which is what an interaction would show. There wasn't one.

The One That Survived

The conclusion.

The paper reports that Experiment 4 showed production blocking accounted for most of the productivity loss of real brainstorming groups[4]. A later review states plainly that production blocking has been the main source of observed productivity losses in face-to-face groups[2].

Three consequences, ours, and they are uncomfortable.

The dominant cause is not motivational. Your people are not slacking, and the loss is not evidence that they lack commitment.

It is not about courage or safety either, which distinguishes it from the psychological safety material in this series. A perfectly safe room with highly motivated participants still has the problem, because the problem is structural.

Which means the standard interventions cannot fix it. Better facilitation, clearer rules, a more inclusive chair, an explicit no-criticism norm: none of these changes the fact that six people sharing sixty minutes have ten minutes of talking each.

The Delay Is The Damage

The mechanism in detail, from the follow-up work.

Diehl and Stroebe's 1991 paper, Productivity loss in idea-generating groups: Tracking down the blocking effect[6], is described as finding that the productivity loss is largely due to the delay between idea generation and verbalization, which might cause participants to suppress their ideas or forget them[7].

The same source elaborates: in brainstorming only one person may gainfully voice his or her ideas in a group at any given time; meanwhile individuals may forget or suppress their ideas because they seem less relevant or less original at a later time; and being forced to listen to ideas of other people may be distractive and interfere with an individual's own thinking[7].

Three distinct losses in that description, and separating them is ours.

Forgetting. The idea existed and is gone by the time your turn arrives.

Self-censorship on relevance. The conversation has moved on, so the idea now seems off-topic or stale, and is withheld. Note that this is not evaluation apprehension: nobody is afraid. The idea simply no longer fits the moment.

And interference. Listening is cognitively demanding, and the attention it consumes is not available for generating.

That third loss connects directly to the seventh article in this series, where feedback that moved attention away from the task impaired performance. The resource is the same one.

The Arithmetic Of A Meeting

Why this is structural rather than cultural. All figures in this section are our own arithmetic on a hypothetical sixty-minute session.

If only one person can usefully speak at a time, then speaking time divides.

Two people: 30 minutes each speaking, 30 minutes waiting.

Four people: 15 minutes each, 45 waiting.

Six people: 10 minutes each, 50 waiting.

Eight people: about 8 minutes each, 53 waiting.

Twelve people: 5 minutes each, 55 waiting.

Two observations, ours.

This is a crude idealisation that assumes equal speaking time, which never happens, and it ignores that some listening is productive. We offer it to show the shape of the constraint, not as a model of a real meeting.

But the shape is the point. In a twelve-person session, each participant spends over ninety percent of the time not speaking, and that is the interval in which ideas are forgotten, judged stale, or displaced by the effort of listening.

The nominal group has none of this. Six people working separately for sixty minutes have 360 person-minutes of generation. The same six in a room have sixty.

Which Is Why Size Makes It Worse

The moderator, which follows from the arithmetic.

The meta-analysis found that the productivity loss of interactive groups as compared to nominal groups increases rapidly with group size[1].

Three consequences, ours.

That is exactly what production blocking predicts and not what free riding or evaluation apprehension straightforwardly predict. Blocking scales mechanically with headcount; motivation does not have to.

So the group-size finding is independent corroboration of the mechanism, arriving from the meta-analysis rather than from the experiments.

And it gives the single most actionable rule in this article: if you must run a group idea session, the number of people in the room is the variable with the clearest evidence behind it. Not the facilitation, not the rules, not the room. The headcount.

Why Quantity Matters For Quality

A link that justifies caring about idea counts at all.

A review notes that in line with Osborn's own proposal that quantity breeds quality, research has shown a strong correlation between the total number of ideas generated and the number of good ideas available therein, and that consequently nominal groups generate more ideas than interactive groups, and hence have more good ideas to choose from[8].

Two observations, ours.

This closes an obvious objection. If ideas were mostly worthless, generating more of them would be pointless. The reported correlation means the count is a proxy for the pool of usable options.

And it means the two meta-analytic findings are not independent. More ideas produces more good ideas, which is part of why the quality effect is nearly as large as the quantity one.

The Illusion Of Group Effectivity

The finding that explains why none of this has changed anything.

Stroebe, Diehl and Abakoumkin published work on what they termed the illusion of group effectivity, described as the belief, persistent despite contradictory empirical evidence, that groups can stimulate creativity[6].

We did not obtain this paper and report only its title and that characterisation.

Our own reading of why the illusion would arise, offered as reasoning rather than as findings from that paper.

The group session feels productive. Ideas are audible, energy is visible, and the output is collectively witnessed.

The nominal group's advantage is never observable to the participants. Nobody in the room experiences the ideas that were forgotten while waiting, because forgetting leaves no trace.

And there is a misattribution available: a participant who hears a good idea in the room may credit the room for it, when the person who said it had it on the way in.

This is the fifth time this series has landed on the same structure. The failure produces no artefact, the success is visible, and the practice validates itself.

Why This Has Not Changed Anything

An honest accounting. This section is our own.

The finding is nearly seventy years old, large, replicated, meta-analysed and mechanistically explained. Group brainstorming remains standard practice in essentially every organisation.

Four reasons we would suggest, none of them stupid.

The illusion above. Participants believe it worked.

Meetings serve purposes other than idea generation: alignment, buy-in, shared understanding, visible inclusion. Those may be worth the productivity loss, and the research measures only the idea output.

The alternative requires organisation. Getting six people to work separately and then pooling the results takes more coordination than booking a room, and the cost is paid up front.

And nobody is accountable for the ideas that were not generated, which is the recurring theme.

The Second Failure: What Gets Discussed

A related problem in group decision-making, reported here briefly because our sourcing is thin.

A review notes that there are several reasons why groups focus on shared information. First, shared information is more likely to be mentioned because more members have this information, and therefore it is more likely to be sampled than is unshared information, citing Stasser and Titus (1985, 1987) among others. Second, discussing shared information is socially rewarded[5].

We obtained no primary source on this literature and report it only as this review describes it.

Two observations, ours, offered cautiously.

The sampling argument is purely statistical. If four of six people know something and one person knows something else, the first fact has four chances to enter the conversation and the second has one. No bias is required.

Which means a group discussion is structurally biased toward what the group already collectively knows, and structurally unlikely to surface the thing only one person knows. That is frequently the thing worth having.

We flag firmly that this second literature deserves its own treatment with proper sources, which this article does not provide.

The Fix Has A Name

The practical remedy, which is not a new idea.

A source records that the Nominal Group Technique was developed as a structural intervention to counter these failures by enforcing independent generation before group discussion, attributed to Delbecq and colleagues[9].

Two observations, ours.

The critical word is enforcing. The technique does not ask people to think independently; it separates the generation phase from the discussion phase as a matter of procedure.

And note what it does not do: it does not abolish the meeting. It changes the order. Generate alone, then convene to evaluate, combine and choose.

Our own view of why the ordering is right. The evidence says groups are worse at generating. It does not say they are worse at evaluating, and the sixth article in this series noted the same distinction between generating a judgment and checking one. Putting the group where it is not handicapped is the whole design.

What Groups Are Actually For

The constructive half. This section is our own reasoning.

Nothing in this literature says convening people is a mistake. It says convening them to generate ideas is.

Four things a group plausibly does better than the same people separately, and we mark these as reasoning rather than findings.

Eliminating. Killing a bad idea benefits from several people noticing different objections, and blocking does not impede noticing.

Combining. Two partial ideas becoming one whole one requires both to be present, which is an argument for pooling after generation.

Committing. A decision people helped make is one they are more likely to execute, which is a real benefit the idea-count literature does not measure.

And catching the error, which the second article in this series described as the specific value of a team that will speak up.

None of those requires the group to be the place where ideas first appear.

In A Small Firm

Where this bites differently, in our assessment.

Three observations.

Small firms have naturally small groups, and the loss scales with size. A three-person discussion is much closer to the nominal ideal than a twelve-person workshop, so a small business is already partly protected.

But small firms also have one person whose ideas dominate the airtime, usually the owner. In the arithmetic above, unequal speaking time makes the blocking worse for everyone else, not better.

And the coordination cost of the alternative is lower. Asking three people to spend twenty minutes writing before a meeting is trivially easy in a small firm and a logistical exercise in a large one.

Our own view: the nominal group technique is cheaper to adopt in a small business than anywhere else, and the small business is the least likely to have heard of it.

What To Do

Separate generation from discussion. Have people produce ideas alone first, then convene. That is the Nominal Group Technique and it exists precisely to counter this failure.

Reduce the headcount in idea sessions. The loss increases rapidly with group size, and headcount is the variable with the clearest evidence behind it.

Do not try to fix it with facilitation. The dominant cause is mechanical rather than motivational, so better chairing, clearer rules and a safer room address the wrong variable.

Collect in writing before the meeting. Written submission removes the turn-taking constraint entirely, which is the specific thing causing the loss.

Count what arrives before versus during. If almost every idea in your process first appears in a room, your generation phase is happening in the worst available setting.

Use the group to evaluate, combine and commit. The evidence indicts group generation. It does not indict group judgment, and the commitment benefit is real and unmeasured here.

Ask what the quiet person wrote down. The structural argument on shared information suggests the thing only one person knows is the least likely to be said aloud.

Distrust the feeling that the session went well. The illusion of group effectivity is a named finding, and the ideas lost to waiting leave no trace for anyone to notice.

The Limits Of This Analysis

Several caveats matter. This article reviews organisational research and is not management, human resources or legal advice. Everything is verified to August 2026. We did not obtain the 1991 meta-analysis directly; its effect sizes reach us through one peer-reviewed review and its study count through another. We did not obtain Osborn (1957) and quote his claim as reproduced with a page reference in a review. We did not obtain the 1991 blocking follow-up study or the illusion of group effectivity paper, and report the latter only by title and characterisation. We obtained no primary source on the shared information literature and report it solely as one review describes it; that material deserves separate treatment with proper sourcing. Our conversion of r = .57 and r = .56 into approximate d values is our own standard transformation and appears in no paper. Our meeting-time arithmetic is our own crude idealisation, assumes equal speaking time which never occurs, ignores that some listening is productive, and is offered to show the shape of a constraint rather than to model a real meeting. The bulk of this evidence is laboratory work on idea-generation tasks, and whether the same magnitudes apply to commercial meetings about real decisions is not answered by the literature we obtained. The four things groups plausibly do better, the three-part decomposition of the blocking loss, the account of why the illusion arises, the four reasons the practice persists and the small-firm analysis are our own reasoning, not findings. Roughly fifteen years of literature after our most recent source was not reviewed.

Related Reading

Behavioural Finance Debate The Idea, Not The Person The other half of what goes wrong when a group convenes to think. Behavioural Finance A Third Of It Made Things Worse The same finite resource: attention spent listening is attention not spent generating. No. 11 Not A Medical Condition The WHO announced burnout was in the ICD, the press reported a new disease, and the WHO issued a correction the next day. The classification it did...

Frequently Asked Questions

Do brainstorming meetings actually produce fewer ideas?

On the evidence, yes. Studies comparing interacting groups with the same number of people working separately find the separated groups produce more non-redundant ideas and more high-quality ideas. A meta-analysis of 20 studies reported the effects as large, and the direction has held in nearly all such studies since 1958.

Is it because people hold back, or slack off?

Neither, mostly. Four experiments tested free riding and evaluation apprehension and found neither explained the loss: removing the free-riding incentive did not remove the penalty, and apprehension reduced productivity for people working alone too. Production blocking accounted for most of it.

What is production blocking?

Only one person can usefully speak at a time, so everyone else waits. During the wait, ideas are forgotten, judged stale because the conversation moved on, or crowded out by the attention that listening consumes. It is mechanical rather than motivational, which is why facilitation does not fix it.

Does group size matter?

It is the clearest lever available. The meta-analysis reports the productivity loss increasing rapidly with group size, which is exactly what a turn-taking constraint predicts and is not what a motivational explanation straightforwardly predicts.

So should we stop having meetings?

No. The evidence indicts group generation, not group judgment. The Nominal Group Technique exists precisely for this: enforce independent generation first, then convene to evaluate, combine and decide. That puts the group where it is not handicapped.

Why has nobody acted on a seventy-year-old finding?

Partly the illusion of group effectivity, a named finding describing the persistent belief that groups stimulate creativity despite contrary evidence. The session feels productive, and the ideas lost while waiting leave no trace for anyone to observe or account for.

References

1. Nijstad, B. A., and colleagues. Beyond Productivity Loss in Brainstorming Groups: The Evolution of a Question, Advances in Experimental Social Psychology, on Osborn having claimed that if his rules were adhered to the average person can think up twice as many ideas working with a group than alone (Osborn, 1957, p. 229); on empirical studies comparing interactive groups with the same number of individuals working alone, whose ideas were combined into a group product by the experimenter with ideas mentioned several times counted only once, having consistently failed to support this assumption; on Mullen and colleagues (1991) having concluded from a meta-analysis of 20 brainstorming studies that nominal groups produce substantially more non-redundant ideas and a substantially greater number of high-quality ideas; and on the productivity loss of interactive groups increasing rapidly with group size. Note: a peer-reviewed review chapter; we obtained portions. sciencedirect.com
2. Computers in Human Behavior. 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, on face-to-face groups being compared to nominal groups formed by aggregating the output of a comparable number of individuals working separately; on the outcome in nearly all such studies being that nominal groups outperform face-to-face groups in non-redundant ideas and idea quality; on Mullen, Johnson and Salas (1991) having conducted a meta-analysis comparing face-to-face and nominal groups from 1958 to 1990 and found mean effect sizes to be large, at r = .57 for the number of non-redundant ideas and r = .56 for idea quality; on these results contradicting the popular but poorly substantiated notion that communication among individuals results in synergistic effects; on production blocking being based on the individual's inability to spontaneously interject ideas without violating group etiquette or breaking the concentration of other members; on production blocking having been the main source of observed productivity losses in face-to-face groups; and on the primary dependent variable being the number of non-redundant ideas. Note: a peer-reviewed article; our sole source for the meta-analytic effect sizes, which we did not obtain from the meta-analysis itself. sciencedirect.com
3. Diehl, M., & Stroebe, W. (1987). Productivity loss in brainstorming groups: Toward the solution of a riddle. Journal of Personality and Social Psychology, 53, 497–509, published abstract, on four experiments investigating free riding, evaluation apprehension and production blocking as explanations of the difference typically observed between real and nominal groups; on Experiment 1 manipulating assessment expectations, with productivity higher under personal than collective assessment instructions but type of session still having a major impact under conditions that eliminated the temptation to free ride; on Experiment 2 demonstrating that inducing evaluation apprehension reduced productivity in individual brainstorming; and on the failure to find an interaction between evaluation apprehension and type of session in Experiment 3 raising doubts about evaluation apprehension as a major explanation. Note: we obtained the published abstract. semanticscholar.org
4. Diehl and Stroebe (1987), hosted copy, on Experiment 4 showing that production blocking accounted for most of the productivity loss of real brainstorming groups, and on the quality measures used being total quality, average quality, number of original or unique ideas, and number of good ideas. Note: a hosted PDF of the paper; we obtained portions. homepages.se.edu
5. RAND Corporation working paper. A Review of the Effects of Group Interaction on Processes and Outcomes in Analytic Teams, on early attributions of productivity loss to evaluation apprehension, being individuals feeling inhibited from sharing ideas in a group setting due to concerns about negative evaluations despite instructions to withhold criticism, or to the tendency for members to free ride because their contributions were less identifiable and more dispensable in interacting groups; on Diehl and Stroebe (1987) having demonstrated in their seminal research that the productivity loss is caused by production blocking, that is, listening to others and waiting for one's turn to speak; and on several reasons why groups focus on shared information, being that shared information is more likely to be sampled because more members hold it, citing Stasser and Titus (1985, 1987), and that discussing shared information is socially rewarded. Note: a working paper; it is our only source on the shared information literature and we obtained no primary source for it. rand.org
6. Stroebe, W., Diehl, M., & Abakoumkin, G. (1992). The Illusion of Group Effectivity. Personality and Social Psychology Bulletin, publisher record, describing the illusion of group effectivity as the belief, persistent despite contradictory empirical evidence, that groups can stimulate creativity; and giving full citations for Diehl and Stroebe (1987), Journal of Personality and Social Psychology 53, 497–509; Diehl and Stroebe (1991), Productivity loss in idea-generating groups: Tracking down the blocking effect, Journal of Personality and Social Psychology 61, 392–403; Mullen, Johnson and Salas (1991), Productivity loss in brainstorming groups: A meta-analytic integration, Basic and Applied Social Psychology 12, 3–23; and Osborn, A. F. (1957), Applied Imagination (rev. ed.), New York: Scribner's. Note: we did not obtain this paper and report only its title and the quoted characterisation; used principally to verify the citations of the other works. dx.doi.org
7. Working paper on brainstorming practice, on production blocking meaning only one person may gainfully voice ideas at any given time while individuals may forget or suppress theirs because they seem less relevant or less original at a later time; on being forced to listen to others' ideas being distractive and interfering with an individual's own thinking; on the productivity loss being largely due to the delay between idea generation and verbalization, citing Diehl and Stroebe (1991); and on evaluation apprehension, being the fear of negative evaluations preventing people from presenting their more original ideas, with social inhibition greater where more group members are perceived as experts. Note: a working paper, used for its summary of the two Diehl and Stroebe papers, neither of which we obtained in full. arxiv.org
8. Journal of Experimental Social Psychology. Productivity is not enough: A comparison of interactive and nominal brainstorming groups on idea generation and selection, on Osborn's proposal that quantity breeds quality; on research having shown a strong correlation between the total number of ideas generated and the number of good ideas available therein, citing Diehl and Stroebe (1987); on nominal groups generating more ideas and hence having more good ideas to choose from; and on interactive groups suffering from production blocking because members have to take turns expressing their ideas, citing Diehl and Stroebe (1991). Note: a peer-reviewed article; we obtained portions. sciencedirect.com
9. Working paper on multi-agent idea generation, on the brainstorming hypothesis that groups generate more ideas than individuals having been refuted by subsequent research, citing Mullen and colleagues (1991); on Diehl and Stroebe (1987) having identified production blocking, the inability to generate ideas while listening to others, as a primary cause; and on the Nominal Group Technique, attributed to Delbecq and colleagues (1986), having been developed as a structural intervention to counter these failures by enforcing independent generation before group discussion. Note: a working paper, used only for its description of the Nominal Group Technique's purpose and attribution. arxiv.org

This article reviews organisational research and is not management, human resources or legal advice. The 1991 meta-analysis was not obtained directly; its effect sizes reach this article through a peer-reviewed review. Osborn (1957), the 1991 blocking follow-up and the illusion of group effectivity paper were not obtained. The conversion of correlations into standardised mean differences, and all meeting-time arithmetic, are the authors' own and appear in no paper. The evidence base is predominantly laboratory work on idea-generation tasks.