Groupthink
Groupthink is conformity pressure inside a cohesive group that suppresses dissent, so alternatives go unvoiced and the group converges early on a poor option. It is the most famous idea in this topic and the most thinly supported.
1. Where it comes from
Irving Janis named it in 1972 from foreign-policy fiascos — the Bay of Pigs, Pearl Harbor, Korea, Vietnam, Watergate — against two successes he judged it avoided, the Marshall Plan and the Cuban missile crisis. It is not settled: a 1991 review found the replications “modest, their quality mixed and their findings only partially conclusive” [1].
2. What survived testing
Each component of Janis’s model fared differently by 1991 [1].
| Component | Evidence as of 1991 |
|---|---|
| Cohesiveness, Janis’s one group-level cause | Repeatedly unsupported. Flowers, Fodor & Smith and Leana all found leadership mattered and cohesiveness did not; Raven found the Nixon group conflict-laden — cohesion absent in one of Janis’s own cases |
| Directive leadership | Supported across four independent studies |
| Stress, which Janis ranked first | Largely untested: experiments “rarely succeeded in creating the kind of decisional stress” the theory requires |
| Stereotyped out-group images | Disconfirmed by Tetlock — no more negative references to adversary states |
| The founding cases | Mixed: Barrett argued groupthink played no role at all in Vietnam |
3. Circularity, and a better mechanism
The central charge against it is circularity [1]: policy failures were picked first and groupthink then looked for, so “groupthink is inferred from policy failure and failure is explained in terms of groupthink.” The alternative mechanism offered is anticipatory compliance — low-status members conforming to what they take the leader to want. Power and hierarchy, not friendliness: a better fit to a software team than cohesion.
Example. At a release go/no-go meeting the principal engineer opens with “I think we ship.” The two juniors holding the flaky-test data say nothing and the release goes out. Nobody pressured them and the team is not close-knit — they read the room and complied in advance. Cohesion explains none of it; the reporting line explains it.
4. The Abilene paradox
The Abilene paradox is a different failure: a group unanimously agrees on a course no member privately favours, each believing the others want it. Jerry B. Harvey described it in 1974. Groupthink is dissent suppressed by pressure; the Abilene paradox is dissent nobody realises exists. Pressure needs cover for the dissenter; false consensus needs only that someone ask.
5. Judge the process, not the outcome
A bad process can still end well: “bad procedures need not always produce bad results; decision-makers may get lucky” [1]. A shipped release is therefore no evidence that the meeting which approved it worked — the same convergence reached from three directions on Cognitive bias.
6. A warning about pages like this one
The review essay closes on a charge that lands on handbooks, this one included [1]:
“the very popularity of groupthink may, in fact, act as an impediment to careful interdisciplinary integration. This emerges clearly from the uncritical adoption of sections on ‘the dangers of groupthink’ in many policy analysis and management handbooks.”
Read the sections above as an account of a contested construct, not as a hazard list to check your team against.
How solid is this?
- Scope, stated in the source’s own words. Groupthink analysis is held to be meaningful only for high-level groups facing consequential, non-routine choices: “it is not really interesting to perform groupthink analyses of regular problem-solving groups at some lower level of management or policy-making.” Applying it to a sprint team is an extrapolation nobody has tested.
- Vintage. That record stops at 1991, so do not call the evidence weak full stop; say that as of 1991 its own reviewer described the replications as modest, mixed and only partially conclusive.
- Not among the measured biases. Groupthink is not one of the 37 cognitive biases mapped in software engineering, and the Abilene paradox does not appear there at all [2].
- Rebori’s groupthink is not Janis’s. The gloss used in the group-decision leaflets — dissent withheld “because they believe no one will agree with them” — is closer to the Abilene paradox, dropping cohesion, stress and structural faults.
- Janis is read here at second hand. His own books are Victims of Groupthink (1972) and Groupthink (1982), the second carrying the systematic statement — so 1972 dates the naming, not the model. The eight symptoms reach this page through the 1991 review, which is also why no remedy list appears: that review reproduces none. The Abilene paradox arrives the same way and is the least directly sourced claim on the page.
Acknowledgments
This content is heavily inspired by and adapted from lectures by Eduardo Miranda and David Root [3] on software project management. The structure, examples, and pedagogical approach reflect their teaching materials and frameworks.
References
- P. ’t Hart,“Irving L. Janis’ Victims of Groupthink,” Political Psychology, vol. 12, no. 2, pp. 247–278, 1991, doi: 10.2307/3791464.
- R. Mohanani, I. Salman, B. Turhan, P. Rodríguez, and P. Ralph,“Cognitive Biases in Software Engineering: A Systematic Mapping Study,” IEEE Transactions on Software Engineering, vol. 46, no. 12, pp. 1318–1339, 2020, doi: 10.1109/TSE.2018.2877759.
- D. Root,Managing Software Development. Carnegie Mellon University / Innopolis University lecture series, 2014.
Disclaimer: AI is used for text polishing and explaining. Authors have verified all facts and claims. In case of an error, feel free to file an issue.