On a hot July afternoon in Texas in the 1950s, a family sat on a porch in Coleman, a small town in the middle of nowhere. They were comfortable. A breeze was up. Someone mentioned, almost casually, that they could drive to Abilene for dinner. It was a 53-mile trip each way, in a car with no air conditioning, on roads the colour of chalk in the heat. Nobody especially wanted to go. But nobody wanted to be the one to say no, so they went. They got to Abilene, ate a mediocre meal in a cafeteria, and drove home four hours later, hot and exhausted.
When they got back, someone finally admitted they hadn’t wanted to go. Someone else said they hadn’t wanted to either, but had thought everyone else did. A third person said the same. In the end, it turned out that not a single person at the original porch had actually wanted to drive to Abilene. They had all gone because they each believed the others wanted to.
This story was told by the management theorist Jerry Harvey in a now-famous 1974 essay, and the phenomenon it illustrates — a group decision that none of the members actually endorsed — has since been called the Abilene paradox. Understanding why it happens is one of the more illuminating entry points into how humans actually behave in groups, and why good people, together, can make decisions none of them would have made alone.
Why groups produce decisions their members don’t want
The mechanism behind the Abilene paradox is called pluralistic ignorance, a term coined by the sociologists Floyd Allport and Daniel Katz in the 1930s. Pluralistic ignorance describes a situation where most members of a group privately disagree with a choice, but each mistakenly believes the others agree with it. Because each person thinks they’re the outlier, no one speaks up, and the choice they all privately disagree with becomes the group’s decision.
This pattern turns up in many settings. Classrooms where students don’t understand the lecture but don’t ask questions, because each student assumes they’re the only one confused. Workplaces where employees privately think a new initiative is a mistake but publicly support it, because each assumes the others believe in it. Social gatherings that continue long past when anyone wanted to leave, because no one wanted to be the first to make the move. Peer groups where young people engage in risky behaviour they privately dislike, because they believe everyone else enjoys it.
What makes pluralistic ignorance so stubborn is that it’s largely invisible from inside the group. Each member is acting on information that looks right but is systematically misleading — their reading of what others want. The decision feels unanimous because everyone is behaving as if they agree, and the appearance of consensus is itself reinforcing. Nobody’s lying. Everyone is just reading the room slightly wrong, in the same direction, at the same time.
The bigger brother: groupthink
A more alarming version of this group dynamic was documented by the social psychologist Irving Janis in his 1972 book Victims of Groupthink. Janis studied decisions that had produced major policy disasters — the American invasion of the Bay of Pigs in 1961, the escalation of the Vietnam War, the failure to anticipate Pearl Harbor — and found a consistent pattern.
High-prestige decision-making groups, he argued, often produce decisions worse than any of their members would have reached alone. The features of what he called groupthink included: pressure toward conformity, stereotyping of outsiders, illusions of unanimity, self-censorship of doubts, and the emergence of “mindguards” — members who took on the informal role of shielding the group from dissenting information.
The Bay of Pigs was Janis’s foundational example. President Kennedy’s administration, staffed with widely-admired figures including the economist Arthur Schlesinger and the defense secretary Robert McNamara, approved a plan to invade Cuba that, in hindsight, had obvious problems. The invasion force was too small. Cuban public support for the rebels had been vastly overestimated. The plan depended on unrealistic assumptions about what would happen after the initial landing.
What made the Bay of Pigs particularly diagnostic, Janis argued, was that several senior members of the group privately had serious doubts. Schlesinger wrote a memo raising concerns. But in meetings, the doubts went unvoiced, or were voiced so tentatively that the group’s apparent commitment rolled over them. Kennedy, who had his own doubts, assumed the confidence of his advisers reflected their analysis. His advisers assumed Kennedy wanted to proceed and suppressed their own doubts. The group, collectively, endorsed a plan that most of them privately questioned.
After the invasion failed catastrophically, Kennedy himself said: “How could we have been so stupid?” Janis’s answer was that the stupidity was structural, not individual. The same people, working separately, would have produced a better analysis. The group dynamic had converged on a decision its members wouldn’t have endorsed alone.
The space shuttle disaster
Janis’s framework was applied, after his death, to a second disaster with similar features. In 1986, the space shuttle Challenger exploded shortly after launch, killing all seven astronauts aboard. The immediate cause was a faulty O-ring seal that had failed in unusually cold temperatures.
What made the disaster illustrative of groupthink was that engineers at Morton Thiokol, the contractor responsible for the solid rocket boosters containing the O-rings, had raised specific concerns about the cold-weather launch the night before. Their data was clear. Their recommendation was clear. But in a teleconference with NASA, the pressure toward the launch — political, schedule-driven, organisational — produced a dynamic in which the engineers’ warnings were progressively softened, their recommendation walked back, and the launch proceeded.
The sociologist Diane Vaughan’s detailed reconstruction of the decision, in her book The Challenger Launch Decision, extended Janis’s framework by introducing the concept of normalisation of deviance — the way organisations can come, incrementally, to treat unusual risks as acceptable because they’ve been accepted before. Each small departure from safety standards was justified by reference to the previous small departures that hadn’t yet produced a disaster. The organisation wasn’t irresponsible. It was following a pattern of progressive reassurance that was rational at each step and catastrophic in aggregate.
The counter-view: sometimes groups are fine
Not every organisational researcher is persuaded that groupthink is as widespread or as dangerous as Janis’s framework implies. Some critics — including the organisational psychologists James Esser and Glen Whyte — have argued that Janis’s evidence was primarily retrospective. He studied disasters and worked backwards to identify the features that had produced them. But many groups have the same features and don’t produce disasters. The predictive power of the groupthink framework, these critics argue, is weaker than Janis claimed.
A related critique is that many good decisions also emerge from consensus processes. If you take Janis’s negative framing and apply it symmetrically to positive group outcomes — the building of successful organisations, the designing of good policy, the creative work of great teams — you’d find the same features of cohesion and mutual trust that, in his examples, produced disaster. The question, these researchers argue, is not whether groups think together, but under what conditions group cohesion becomes productive rather than corrosive.
This is probably right. The honest conclusion is that the Abilene paradox and groupthink are real dynamics, but not universal ones. They emerge in specific conditions: high external pressure, strong in-group cohesion, isolation from dissenting voices, time constraints, high-stakes decisions. Groups without those features can work excellently together. Groups with them are at specific risk, and knowing the risk exists is the start of counteracting it.
How to stop Abilene from happening to you
There’s a useful practical literature on how to prevent these dynamics, drawn from organisational research.
Ask for disagreement before endorsement. If you’re leading a meeting, the worst question is “does anyone disagree?” — which puts the burden of social friction on the dissenter. A better question is “what’s the strongest case against this?”, which invites anyone in the room to articulate the objection as an intellectual exercise, even if they don’t hold the objection themselves.
Have people write privately before speaking. One of the simplest interventions documented in organisational research. If each member of the group writes their view privately before the discussion starts, you get a much more honest distribution of views than if the discussion begins with speaking. The social dynamics that suppress dissent take hold in the act of speaking up; private writing bypasses them.
Designate a formal dissenter. The Catholic Church’s tradition of the “devil’s advocate” — a specific person assigned to argue against a proposed canonisation — is one early institutional recognition of this. Modern versions assign someone in every important meeting to play devil’s advocate, so that dissent is structurally present rather than dependent on individual courage.
Leave the room. Janis recommended that leaders absent themselves from early-stage discussions, so that junior members can form views without the social pressure of agreeing with whoever is senior. Kennedy himself, after the Bay of Pigs, changed the meeting structure for subsequent crises, sometimes literally leaving so his advisers could argue honestly with each other.
The question worth asking
Perhaps the most useful thing to take away from this body of research is a small habit. Before going along with a group decision you’re privately uncertain about, ask yourself: am I going along because I agree, or because I think everyone else does? It’s a surprisingly hard question to answer honestly, because the mechanisms of pluralistic ignorance are designed to hide themselves.
If you can, in that moment, genuinely say “I privately think this is wrong but everyone else seems to want it”, there’s a real chance that most of the other people in the room are thinking something similar. Your dissent, spoken aloud, may not be the voice that ends the group’s cohesion. It may be the voice that gives everyone else permission to say what they were already thinking.
The question to sit with, particularly if you’ve ever left a meeting unsettled by a decision you remember nodding at:
Is anyone in the room actually for this — or are you all driving to Abilene because nobody wanted to say no first?
Key research referenced: Jerry Harvey, “The Abilene Paradox” (Organisational Dynamics, 1974); Floyd Allport and Daniel Katz on pluralistic ignorance (1930s); Irving Janis, Victims of Groupthink (1972); Diane Vaughan, The Challenger Launch Decision (1996); organisational-behaviour critiques of Janis by Esser, Whyte and others.