The foundational experiment and its organisational implication

In 1968, John Darley and Bibb Latané placed college students in a group discussion and arranged for one participant to appear to have a seizure over an intercom. When participants believed they were alone, 85% sought help. When they believed four others were also listening, only 31% did. The difference was not empathy — it was the arithmetic of felt responsibility. Each additional observer divided the perceived personal obligation to act among more people until, at a group of five, the individual’s share had fallen below the threshold that produces action.

The organisational translation is direct. When a quality problem, ethical violation, safety hazard, or strategic error is visible to an entire team, each member is implicitly dividing their perceived personal responsibility by the number of people in the room. In a ten-person team meeting, each member feels approximately one-tenth responsible for reporting. For most problems that fall short of obvious emergency, that level of felt responsibility is insufficient to overcome the personal costs of speaking up. The problem is not that nobody cares — it is that everybody assumes somebody else will handle it.

The smoke-filled room: misreading shared silence as shared confidence

In a related experiment, Latané and Darley placed participants in a room that began to fill with smoke. When participants were alone, 75% reported the smoke within six minutes. When placed with two passive confederates who ignored the smoke, only 10% reported it within the same period.

The smoke-filled room experiment is the most directly organisationally relevant finding in the bystander literature. Each observer, uncertain whether the situation is serious enough to report, looks to colleagues for social proof about its severity. Each colleague is performing the same search and observing the same apparent calm. The group converges on a shared conclusion — this must not be serious, or someone would already have acted — that no individual member actually holds with any confidence.

This is pluralistic ignorance operating at its most commercially destructive. The silence of each is misread by all as the confidence of each, producing a collective false consensus that actively suppresses the reporting any individual member, if alone, would likely have provided. This is the mechanism that produces the organisational phenomenon where everyone privately has concerns about a project, a hire, a financial decision, or a product direction — and yet the meeting ends with apparent consensus that all is well.

The five-step model: five sequential failure points

Latané and Darley’s five-step decision model — notice the event, interpret it as serious, assume personal responsibility, know what to do, decide to act — is the most operationally useful framework for understanding why problems go unreported. A person must clear every step for reporting to happen. Failure at any single stage means silence.

The five steps correspond to five distinct organisational design failures. Failure at step one is an attention problem — cognitive load, open-plan noise, and information overload prevent the problem from being noticed at all. Failure at step two is the pluralistic ignorance problem — colleagues’ apparent calm signals that the situation is not serious. Failure at step three is the diffusion problem — somebody more senior or more qualified is surely handling this. Failure at step four is a systems problem — there is no clear reporting channel and no identified person to tell. Failure at step five is the evaluation apprehension problem — the personal cost of speaking up outweighs the expected benefit.

Most organisational breakdowns occur at step three. The question “is it my job to report this?” — in the absence of any explicit answer — defaults to no, because each person assumes their partial sense of responsibility is being more than covered by the total responsibility of others.

If witnessing problems in your organisation and feeling unable to address them is significantly affecting your wellbeing, that matters beyond the organisational design question. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). International: iasp.info/resources/Crisis_Centres. Crisis Text Line — text HOME to 741741.

Why hierarchy makes the organisational bystander effect worse

In an emergency on the street, evaluation apprehension is the fear of looking foolish in front of strangers. In an organisation, it is the fear of being penalised by someone who controls performance reviews, promotions, and day-to-day work experience. The personal cost of speaking up in an organisation is measurably higher than the cost of intervening in a public emergency — which means the psychological threshold for action is higher and the proportion of observers who clear it is correspondingly lower.

The bystander effect shows up when employees witness harassment, discrimination, or unsafe practices but remain silent — particularly when the perpetrator is senior. The hierarchy amplifies both diffusion of responsibility (more senior people must surely be more responsible for this) and evaluation apprehension (the person most likely to judge my intervention negatively is the person with the most power over my career), simultaneously suppressing reporting through both mechanisms.

The interventions the research actually supports

Diffusion of responsibility is eliminated when one specific person is named as responsible for a specific action. The difference between “anyone who sees a problem should report it” and “the project lead is responsible for flagging concerns to the steering committee” is not a marginal wording difference — it is the difference between a diffusion condition and a designated-responsibility condition, with dramatically different reporting rates. Naming collapses the responsibility distribution back to one individual, removing the arithmetic that makes the bystander effect operate.

Psychological safety addresses evaluation apprehension at step five. When speaking up carries no career risk and no embarrassment, the step-five calculation produces action rather than silence. Edmondson’s finding that psychological safety is the strongest predictor of team performance operates specifically because it removes the mechanism that makes reporting personally costly even when the observer is fully aware of the problem and motivated to address it.

A book worth reading alongside this

The Lucifer Effect by Philip Zimbardo is the most comprehensive treatment of how situational factors — including diffusion of responsibility, authority hierarchies, and the gradual normalisation of inaction — produce the bystander dynamics that allow organisational problems to persist without intervention. Zimbardo’s analysis extends well beyond the Stanford prison experiment to the specific institutional structures that make good people silent in organisations where bad things are happening. For any entrepreneur who wants to understand not just the mechanisms but the systemic conditions that sustain them, and what dismantling those conditions actually requires, this book provides the most rigorous and most honest available starting point.

Have questions about this article?

If any part of this article raised questions you want to explore further, courbot.co is built for exactly that. It is courben.co’s AI assistant, designed around the psychology of entrepreneurship. Ask it anything from this article.

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 therapist finder: bacp.co.uk/search/Therapists. International: iasp.info/resources/Crisis_Centres. Crisis Text Line — text HOME to 741741.

This article is for educational and informational purposes only. Sources: Darley, J.M. & Latané, B. (1968), Journal of Personality and Social Psychology, 8(4), 377–383. Latané, B. & Darley, J.M. (1970), The Unresponsive Bystander: Why Doesn’t He Help? Fischer, P. et al. (2011), Psychological Bulletin, 137(4), 517–537.