Groupthink is the pursuit of concurrence overriding the testing of options

What it is

Groupthink is what happens in a cohesive group when the drive to maintain consensus overpowers the testing of options — members start optimizing for "we are united" rather than "we are right." The symptoms arrive as a set: illusion of invulnerability, direct pressure on dissenters ("how can you doubt this"), self-censorship (doubts held but unvoiced), the illusion of unanimity (no objections equals everyone agrees), and self-appointed mindguards filtering information to fit the leader's preference. It differs from ordinary conformity: conformity is an individual's choice under group pressure; groupthink is a pathology of the whole decision process — the verification stage is bypassed systematically while the group feels good about its decision quality.

Why it happens

The driving force is the group treating "staying unified" as an asset to protect. In a highly cohesive group, solidarity itself is capital — belonging, smooth operation, shared identity — and any public dissent spends that capital, so group norms raise the social cost of voicing doubts until it is not worth paying. Self-censorship moves first: each member who had a reservation sees everyone else calmly in favor and attributes the doubt to their own oversensitivity (informational and normative conformity working simultaneously), and stays silent; that silence then becomes evidence for the next person — "unanimity" is illusory, yet every silent member personally confirms it. External threat and time pressure intensify the loop: the harsher the environment, the more the group contracts and the more it treats internal dissent as an additional risk. The end state is a hollowed-out decision process: no alternatives generated, no risks assessed, no outside information searched, no contingency plans drafted — not from incompetence, but because every one of those steps requires someone willing to ask "could we be wrong," and nobody asks anymore.

Studying it

• Paradigm: case induction is the original route — archival analysis of decision disasters (the Bay of Pigs and its kin) extracting the symptom-and-antecedent cluster; experimental induction manipulates cohesiveness and leader preference in scenario tasks and observes the decay of information search, alternative generation, and dissent.
• Variables: cohesiveness, timing of leader statements, procedural safeguards (presence of a devil's advocate), and external threat / time pressure as independent variables; number of alternatives, depth of risk discussion, frequency of dissent, and decision quality as dependent variables.
• Use in interface research: evaluating whether decision-process tools can detect groupthink precursors in behavioral data — collapsing dissent rates, stalled alternative generation, narrowing information search.
• Methodological caveat: case studies carry outcome bias by construction (only failed decisions are studied; successful cohesive decision-making may be survivorship bias at work). In experiments, groupthink is a multi-symptom construct — measuring one symptom (reduced dissent) is not equivalent to the construct; and coding the symptom checklist is subjective, demanding inter-coder reliability controls.

Where it stops holding

Groupthink is not a synonym for "groups decide badly." It is a failure mode under a specific conjunction — high cohesiveness plus structural defects plus situational pressure all present — and is therefore blockable at specific joints (the topics of the following entries: antecedent conditions, leader-statement timing, institutionalized dissent, independent proposals). Loose, low-cohesion groups have other diseases (free-riding, coordination failure), not this one. Guard against concept abuse too: not every "we agreed quickly" is groupthink — if alternatives were genuinely tested and dissent genuinely solicited, fast consensus is efficiency, not pathology.

Applying it

• Write "verification happened" into the decision template as a requirement: alternatives list, risk assessment, external information search — three mandatory sections, and an empty section needs a stated reason.
• Collect reservations anonymously before the meeting as a fixed input to discussion — the anonymous channel bypasses self-censorship.
• Watch for three signals co-occurring: the meeting goes too smoothly, nobody states conditions, and all information comes from the inner circle — co-occurrence triggers a procedural review, not further progress.
• Verification: retrospectively code past decisions on the symptom profile (alternative count, dissent record, information-source distribution) and compare with contemporaneous "smooth but later wrong" decisions; if the wrong ones score systematically higher, the checks above need to become standing procedure.

Related

• Same group: V9.03.2 Risk peaks when high cohesiveness meets external pressure · V9.03.3 A leader who states a preference early suppresses later dissent · V9.03.4 A designated devil's advocate institutionalizes dissent · V9.03.5 Independent drafting before pooling prevents premature convergence
• Nearby: V9.02 Information Pooling and Missed Perspectives · V9.01 Voting and Consensus Mechanisms
• Search terms: groupthink · concurrence-seeking · mindguards