The foundational finding

James Stoner’s original research on what he called the risky shift documented that groups asked to deliberate on choice dilemmas and reach a collective recommendation consistently endorsed riskier options than the average of their members’ individual pre-discussion positions. This was unexpected: the prevailing assumption was that group deliberation would moderate extreme positions toward the average. What Stoner found was the opposite. The group’s discussion moved the average position toward greater risk rather than toward the centre.

Serge Moscovici and Marisa Zavalloni’s subsequent research on group polarisation generalised the finding beyond the risk domain and provided the more accurate description of the mechanism: group discussion does not always produce a risky shift. It produces a shift in whichever direction the group’s initial inclination points. A group initially leaning toward caution becomes more cautious after discussion; a group initially leaning toward risk becomes more risk-endorsing. The discussion amplifies the initial tendency rather than moderating it. The risky shift is not a universal feature of group deliberation; it is the specific output of group deliberation in groups with an initial risk-taking orientation. In the founding team context, where risk-taking is a cultural identity and a valued signal of entrepreneurial seriousness, the initial orientation is almost always risk-supportive, which means the polarisation mechanism almost always pushes in the direction of greater risk.

The persuasive argument mechanism

Eugene Burnstein and Amiram Vinokur’s persuasive argument account of group polarisation identified the primary cognitive mechanism. In a group with an initial risk-leaning orientation, the arguments that members generate and exchange during discussion are predominantly risk-supporting. This is not a deliberate selection process. Members’ initial orientations shape which arguments they have available and which they find compelling, so the pool of arguments circulating in the discussion is weighted toward the initial tendency. Discussion exposes each member to risk-supporting arguments they had not previously considered, providing new reasons to hold a stronger version of their initial position. Each member’s final position is more extreme than their initial one because they have acquired additional supporting arguments, not because they have been challenged by genuinely opposing ones.

The persuasive argument mechanism explains why the standard meeting format is insufficient to counteract group polarisation. The discussion that follows an initial risk-supporting proposal in a risk-oriented founding team will predominantly generate risk-supporting arguments, because the members generating arguments are risk-oriented and the social context of the discussion selects for arguments that support the emerging consensus direction. The caution-supporting arguments that would correct the polarisation are not generated, not because members do not hold them, but because the persuasive argument dynamic does not create the conditions under which they are offered.

The social comparison mechanism

George Goethals and Mark Zanna’s social comparison account identified the complementary social mechanism. During group discussion, members discover that the average group position is more risk-supportive than they had expected. This discovery triggers the social comparison mechanism: the member who held a moderately risk-supportive position discovers that the group norm is more risk-supportive than their own position, which motivates movement toward and beyond the perceived group norm to demonstrate commitment to the group’s values. In a founding team where risk-taking is an identity signal, the discovery that the group is more risk-tolerant than one’s current position creates the competitive pressure to match and exceed the risk-taking of the group.

The social comparison mechanism is particularly potent in founding teams because the risk-taking identity is not merely a preference but a social commitment. The founding team member who expresses caution in a group that values bold action is not merely expressing a different view; they are signalling a lower commitment to the founding identity, which carries the social cost that the Moscovici minority influence research predicts for the consistent dissenter. The social comparison mechanism makes caution costly and risk-escalation costless, which predictably produces the competitive risk-escalation that the polarisation research documents.

You might also like
Why brainstorming sessions produce worse ideas than individuals working alone
Why the most popular ideation technique in business is also one of the most consistently counterproductive
The psychology of discounting and why price reductions often destroy more value than they generate
A discount increases purchase probability in the moment. What it also does — less visibly, and more consequentially — is train the customer's expectations, degrade the reference price, and communicate something about quality that the next marketing campaign cannot fully undo.
The psychology of the support network and why entrepreneurs who build alone perform differently from those who do not
The research on social connection and cognitive performance establishes that the support network is not emotional infrastructure supplementary to the real work of entrepreneurship — it is a biological resource that determines the cognitive and physiological conditions under which the real work is done.
The stress response system and how the HPA axis works + what it was designed for versus what entrepreneurship asks it to do
The hypothalamic-pituitary-adrenal axis is one of the most effective adaptive systems the body possesses — and the chronic psychosocial stress of entrepreneurship is precisely the category of demand it was not built to handle.

When polarisation serves the team

Group polarisation is not uniformly harmful. In groups with an initial caution-leaning orientation facing decisions where boldness is required, the polarisation mechanism can push the collective position toward the necessary action that individual caution would prevent. The founding team that is risk-averse in a domain where genuine risk-taking is commercially required, entering a competitive market, abandoning a failing approach, making a bold hire, benefits from the polarisation mechanism’s amplification of the initial caution-departing tendency. The research on group polarisation documents a mechanism, not a pathology. The pathology arises specifically when the mechanism amplifies risk-taking in domains where caution is warranted and when the cohesion of the group suppresses the dissent that would correct it.

Irving Janis’s groupthink connection

Irving Janis’s groupthink research identified the conditions under which group polarisation becomes its most extreme and most damaging form. The combination of high group cohesion, insulation from external information, directive leadership, and high stakes produces the illusion of invulnerability: the shared belief, generated by the mutual reinforcement of the polarisation mechanism in a closed system, that the group’s judgment is superior and that caution is unnecessary. The Bay of Pigs planning group, which Janis analysed in detail, showed each of the polarisation mechanisms operating within the specific groupthink conditions that prevented their correction. The risk-supporting arguments circulated freely; the caution-supporting arguments were socially suppressed; the social comparison mechanism produced competitive endorsement of increasingly bold action; and the cohesion of the group made dissent sufficiently costly that the one member who held a dissenting view did not voice it in the meeting where it would have mattered.

What the research shows about correction

Cass Sunstein and Reid Hastie’s deliberation research identified the most reliably effective available polarisation correction: the pre-deliberation independent judgement. Requiring each member to commit to an explicit position and reasoning before the group discussion begins prevents both the persuasive argument mechanism and the social comparison mechanism from proceeding from an undeclared and therefore malleable starting point. The member who has committed to a specific caution-supporting position before discussion begins has a declared starting point that the discussion’s persuasive arguments must overcome rather than an undeclared orientation that those arguments can amplify without resistance. The independent position prevents the cascade.

Gary Klein’s pre-mortem technique provides the most directly applicable structural intervention for the specific risk domain that group polarisation most affects. The deliberate generation of failure scenarios before the commitment is made introduces the caution-supporting arguments that the risk-leaning group discussion systematically fails to generate. The prospective hindsight frame, imagining the decision has already been made and the outcome was catastrophic, activates the failure-oriented thinking that the group’s risk-supportive orientation suppresses, producing a more balanced distribution of arguments that the subsequent discussion can actually engage with. The pre-mortem does not prevent risk-taking. It prevents the specific form of risk-taking that results from a group discussion that has only considered one side of the available evidence.

If the experiences described in this article 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: Stoner, J.A.F. (1961). A comparison of individual and group decisions involving risk. Unpublished master’s thesis, MIT. Moscovici, S. and Zavalloni, M. (1969). The group as a polarizer of attitudes. Journal of Personality and Social Psychology, 12(2), 125-135. Burnstein, E. and Vinokur, A. (1977). Persuasive argumentation and social comparison as determinants of attitude polarization. Journal of Experimental Social Psychology, 13(4), 315-332. Goethals, G.R. and Zanna, M.P. (1979). The role of social comparison in choice shifts. Journal of Personality and Social Psychology, 37(9), 1469-1476. Janis, I.L. (1972). Victims of Groupthink. Houghton Mifflin. Sunstein, C.R. and Hastie, R. (2015). Wiser. Harvard Business Review Press. Klein, G. (2007). Performing a project premortem. Harvard Business Review, 85(9), 18-19. Brafman, O. and Brafman, R. (2008). Sway. Doubleday. Taleb, N.N. (2007). The Black Swan. Random House.