Asch’s (1956) conformity experiments did not find that people are unusually susceptible to social pressure or that they lack the cognitive capacity to identify an obvious correct answer. The participants in those studies could see perfectly well that the comparison lines were different lengths. What they could not easily do was publicly contradict a unanimous group. Seventy-five per cent of participants conformed to the group’s clearly incorrect answer on at least one trial. Thirty-two per cent of all trials produced conformity. The cost of public dissent was social, not cognitive, and for most people in most moments, the social cost won.

This finding sits at the foundation of every team decision problem. The question a leader needs to answer is not whether their team is capable of independent judgment. Almost certainly it is. The question is whether the structure of their decision-making processes allows that independent judgment to be expressed before social influence has already done its work.

Groupthink: what false consensus looks and feels like from inside

Janis’s (1972) groupthink research gave the phenomenon its name and its diagnostic framework. Working from detailed post-mortems of major political and military decisions that went catastrophically wrong despite being made by capable, experienced people, Janis identified eight symptoms that distinguish social pressure consensus from genuine agreement: illusion of invulnerability, collective rationalisation, belief in the group’s inherent moral correctness, stereotyped views of out-groups, direct pressure on dissenters, self-censorship among members, an illusion of unanimity, and the emergence of self-appointed mindguards who filter information reaching the group.

The most important feature of this list is that none of these symptoms are experienced as symptoms from inside the group. The Challenger launch decision produced all eight simultaneously. The engineers who knew the O-ring failure risk at low temperatures raised the concern through the appropriate channels. The group dynamics that Janis documented meant the concern did not survive contact with the decision-making process. The consensus that produced the launch decision was experienced by its participants as genuine agreement reached through normal deliberation. The post-mortem showed it was social pressure consensus from start to finish.

The herding problem: why the first voice matters more than any other

Lorenz et al.’s (2011) social influence research identified the specific mechanism through which group discussion reduces rather than improves collective accuracy. When people state their positions sequentially, each stated position influences the positions of those who follow. The result is convergence toward the earliest stated positions that is entirely independent of whether those positions are accurate. The group’s collective estimate after discussion is less accurate than the average of its members’ independent pre-discussion estimates would have been.

This finding has a direct structural implication. Any decision process that begins with the most senior or most confident person stating their view has already compromised the independence of every subsequent contribution before the discussion has properly started. The first voice in a hierarchical group does not just contribute a view. It anchors all subsequent views around itself, regardless of its accuracy.

Nemeth’s minority influence: why the dissenter who is wrong still helps

Nemeth’s (1986) minority influence research produced the finding that matters most for team design. Exposure to a consistent minority dissenting view improves the decision quality of the majority, not because the minority view is correct, but because its presence activates the divergent thinking and deeper information processing that genuine consideration of alternatives requires. Groups exposed to minority dissent consider more angles, generate more creative solutions, and make fewer errors than groups that reach consensus without encountering a persistent alternative view.

The dissenter who is wrong makes the majority think better. This means the value of dissent in a decision-making process is not contingent on the dissenter being right. The structural question is whether the team’s decision process creates conditions in which dissent can be expressed and sustained long enough to produce the cognitive benefits Nemeth documented.

You might also like
Psychological safety and creative risk + why people only propose genuinely original ideas when failure feels survivable
The most original ideas in any room are frequently not the ones that are voiced. They are the ones that are filtered out before the meeting begins by the implicit assessment of whether expressing them is safe. Psychological safety does not produce more good ideas. It produces more of the ideas that already exist being made available for evaluation.
The psychology of loneliness at the top and why leadership isolates in ways that are not obvious from the outside
The intuition that having more people around would reduce loneliness is one of the more reliably falsified predictions in entrepreneurial psychology. Founders who have built teams consistently report that the loneliness of the scaling phase is not less than the loneliness of building alone -- it is different, and in specific ways harder to address. The early-stage founder is lonely because they are alone. The scaling founder is lonely because they are surrounded by people to whom they cannot be fully honest, with whom genuine mutual vulnerability is no longer structurally possible, and from whom the power differential that leadership creates has transformed every genuine connection into something more managed and more performative than it was before. Understanding how leadership produces this specific isolation is the prerequisite for doing anything useful about it.
How power distance shapes the psychology of hierarchy, feedback, and accountability in international organisations
Hofstede’s power distance dimension is among the most practically consequential of the cultural dimensions for the cross-cultural entrepreneur because it shapes not the surface features of organisational life — the titles, the formalities, the seating arrangements — but the deep psychological dynamics through which information flows, decisions are made, and accountability is distributed. In high power distance cultures, the same organisational structure that produces the collaborative, challenge-rich, information-dense environment that a low power distance culture entrepreneur expects produces instead a structurally silent organisation in which the founder’s view is never genuinely tested, the team’s critical information never reliably reaches the top, and the accountability system concentrates authority in ways that compound rather than correct the errors it creates.
How reciprocity works as a behavioural economics mechanism and why giving something first is the most reliable way to receive something back
Reciprocity is not a marketing tactic that works on naive consumers. It is a universal social norm so deeply installed in human psychology that it operates below conscious awareness, produces obligation even when the gift was uninvited, and persists even when the recipient is educated about the mechanism and actively monitoring for it.

Psychological safety and the Asch ally effect

Edmondson’s (1999) psychological safety research and Asch’s conformity research point to the same structural conclusion from different directions. Asch found that a single ally who expressed the correct view reduced conformity from 32% of trials to 5.5%. One person willing to state a different view publicly was sufficient to dramatically change the social dynamics for everyone else in the room. Edmondson’s research established that psychological safety is the primary determinant of whether that ally exists: whether the person who holds a different view believes it is safe to say so.

The psychologically safe team has not eliminated the social cost of dissent. What it has done is reduce that cost sufficiently that the 5.5% dissent rate becomes the norm rather than the exception. The structural interventions that produce psychological safety, consistent leader behaviour that rewards challenge and does not punish honest assessment, are the same interventions that allow the Asch ally to function.

The structural fixes: separating individual judgment from social expression

The research-supported structural interventions for genuine consensus share a common design logic: they separate the generation of individual positions from the social dynamics of group expression. Anonymous pre-meeting input collects individual views before any social anchoring has occurred. The structured devil’s advocate assigns the dissenting role formally, removing the personal social cost from the individual who would otherwise bear it. The pre-mortem, in which the group is asked to assume the decision has already failed and to identify why, activates the divergent thinking that Nemeth’s research documents as the primary benefit of minority dissent.

Amazon’s six-pager format applies this logic at scale. All participants read the briefing document in silence at the start of the meeting and form individual views before any verbal discussion begins. The social influence mechanism that Lorenz documented, where early stated positions anchor all subsequent positions, cannot operate on positions that have not yet been stated. The discussion that follows engages genuinely independent assessments rather than positions already shaped by whoever spoke first.

Books worth reading on this

The Wisest One in the Room by Thomas Gilovich and Lee Ross draws on decades of social psychological research to examine why groups and individuals so consistently fail to make the most of their own collective intelligence, covering the specific cognitive and social mechanisms through which good judgment gets overridden by conformity pressure, motivated reasoning, and the structural features of group decision-making that allow social influence to do more work than independent assessment. Gilovich and Ross are two of the most distinguished social psychologists of their generation, and this book represents an unusually accessible account of research that is usually confined to academic journals. Their specific treatment of the conditions under which group discussion improves versus degrades decision quality maps directly onto the Asch and Lorenz findings this article describes, and their practical account of what it takes to create the conditions for genuine independent judgment in group settings gives the entrepreneur or team leader a concrete framework rather than a list of abstract principles. For anyone responsible for a team that makes consequential decisions, this is the most practically grounded available account of the psychology behind why those decisions so often reflect the room’s social dynamics more than its members’ actual views.

If the dynamics described here are significantly affecting your wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP: bacp.co.uk/search/Therapists. Crisis Text Line — text HOME to 741741 (US, UK, Canada, Ireland). International: internationaltherapistdirectory.com.

This article is for educational and informational purposes only. Sources: Asch, S.E. (1956), Studies of Independence and Conformity, Psychological Monographs, 70(9). Janis, I.L. (1972), Victims of Groupthink, Houghton Mifflin. Lorenz, J. et al. (2011), How Social Influence Can Undermine the Wisdom of Crowds, PNAS, 108(22), 9020-9025. Nemeth, C.J. (1986), Differential Contributions of Majority and Minority Influence, Psychological Review, 93(1), 23-32. Edmondson, A.C. (1999), Psychological Safety and Learning Behaviour in Work Teams, Administrative Science Quarterly, 44(2), 350-383. Klein, G.A. (2007), Performing a Project Premortem, Harvard Business Review, 85(9), 18-19. Gilovich, T. & Ross, L. (2015), The Wisest One in the Room, Free Press. Tetlock, P.E. & Gardner, D. (2015), Superforecasting: The Art and Science of Prediction, Crown.