Two questions run through this whole topic, and it helps to keep them apart. The first is a situation question: when someone needs help, what makes bystanders step in — or freeze? The second is a motive question: when we do help, are we ever really doing it for the other person, or is it always, deep down, for ourselves? The first question has a fairly clean answer. The second is one of social psychology's great unsettled debates, and the honest thing to say is that it's still open.
Bodhi says
Here's the sentence to hold onto: the bystander effect is not about how good people are — it's about how many of them there are. Now, the two traps exams love. Trap one: the effect is about the odds that any one bystander helps — with a big enough crowd, the victim often still gets helped, because someone steps up. Keep "probability a given person helps" and "probability anyone helps" separate, or a question will separate them for you. Trap two: the effect bites hardest when the situation is ambiguous; in clearly dangerous emergencies it weakens and can even reverse (Fischer et al., 2011) — so "crowd = guaranteed apathy" is always the wrong answer. Last: be ready to name the mechanism in a scenario. Nobody feels it's their job → diffusion of responsibility. Everyone reads everyone else's calm → pluralistic ignorance. Fear of looking foolish → audience inhibition. Scenario-to-mechanism matching is the most predictable question this unit produces.
Three words, kept apart
Before the stories, the vocabulary — because the terms nest inside one another. Prosocial behavior is the umbrella: any act intended to benefit another person — helping, comforting, sharing, donating, volunteering — whatever the motive behind it. Altruism is the strict case: helping that carries a real cost to the helper and no obvious payoff, the "pure" form everyone argues about below. Cooperation is the friendly middle: you help another person and benefit yourself in the same move, as in a study group or a shared kitchen. Most everyday kindness is prosocial and cooperative; whether any of it is truly altruistic is the fight this topic ends on.
The story that built a research program
In March 1964, a young woman named Catherine "Kitty" Genovese was attacked and murdered outside her apartment building in Queens, New York. Two weeks later, The New York Times ran a front-page story with a shattering claim: 38 witnesses watched or listened from their windows over more than half an hour, and not one of them intervened or even called the police in time. The story became instantly famous. It made "bystander apathy" a household phrase and, crucially, it goaded two young social psychologists — John Darley and Bibb Latané — into asking a scientist's version of the horror everyone felt: what is it about a crowd of witnesses that makes helping less likely?
Notice the reflex the field was already trained to distrust. The popular explanation was dispositional: New Yorkers are cold, city people don't care, modern life has made us callous. Darley and Latané suspected something more interesting and more disturbing — that the failure to help wasn't a defect of those particular people but a predictable product of the situation they were in. And so they went to the lab.
The myth
Thirty-eight witnesses coldly watched a woman be murdered over 30+ minutes and, out of apathy, did absolutely nothing while it happened.
What later reporting found
The famous "38 who did nothing" account was substantially inaccurate (Manning et al., 2007). Many supposed witnesses never saw the attack, most couldn't have grasped a murder was underway, at least some did call police, and one neighbor came out and held Genovese as she was dying. The Times narrative was a compelling story built on shaky facts.
Hold two things at once — this is the part textbooks rush
The myth being overstated does not make the science wrong. This is the trap students fall into: "Oh, the Genovese story was debunked, so the bystander effect is fake." Not at all. The journalism was flawed; the research it inspired is real, replicated, and valuable. Darley and Latané didn't rely on the newspaper — they went and built controlled experiments (smoke, seizures, staged emergencies) that produced the effect on purpose, under conditions they could measure. Learn to separate a shaky anecdote from the solid experiments it happened to motivate. That skill will serve you well.
The crowd in the lab: two classic demonstrations
Rather than argue about Queens, Darley and Latané manufactured emergencies where they could count who helped and how fast. Two studies became the backbone of the field.
The smoke-filled room (Latané & Darley, 1968). Participants filled out questionnaires in a waiting room when smoke began seeping in through a wall vent — a plausible sign of fire. Alone, most people investigated and reported it: about 75% acted within the experiment's window. But seated with two calm confederates who deliberately did nothing, only about 10% reported the smoke — even as the room hazed over. Same smoke, same danger; the only thing that changed was the social scene, and it flipped behavior almost completely.
The seizure study (Darley & Latané, 1968). Participants talked over an intercom with people they believed were other students. One of those "students" (a recording) appeared to have a seizure, choking and calling for help. When a participant thought they were the only other person on the line, nearly everyone left to get help, and quickly. When they believed four others were also listening, far fewer helped and those who did were slower. More witnesses, less helping — cleanly, in the lab, on purpose.
Three mechanisms behind the freeze
The bystander effect isn't one thing; it's at least three, all firing at once when others are present. Diffusion of responsibility — with many witnesses, each person's felt obligation to act shrinks ("surely someone else will"). Pluralistic ignorance — everyone glances around, sees other people looking calm, and privately concludes it must not be a real emergency (while they're reading the same false calm off you). And audience inhibition — the fear of stepping up, being wrong, and looking foolish in front of an audience. Notice these are all situational, and all get stronger as the crowd grows.
Two footnotes on that list. First, audience inhibition is the same thing textbooks call evaluation apprehension — the fear of acting wrongly in front of others and being judged for it. Second, the crowd does not even have to be real. Garcia and colleagues (2002) asked people simply to imagine having dinner with ten friends, versus one friend, and then asked how much time they'd volunteer for another study. Those who had pictured the bigger group offered less — an implicit bystander effect in which merely thinking about a crowd diffused responsibility and dulled generosity (Garcia et al., 2002). The presence of others works partly through your head, not only through the room.
The five-step decision model
Darley and Latané's real gift wasn't just the effect — it was a map of the decision underneath it. Helping isn't a single yes/no; it's a chain of five decisions (Latané & Darley, 1970), and a bystander who fails any one of them doesn't help. Walk the chain, and you can see exactly where a crowd jams the gears.
1 · Notice the event
You can't help what you don't perceive. A hurried, distracted, or preoccupied person may walk right past someone in trouble — a point the seminarian study below turns into its whole punchline. Failure here: you never even register that something's wrong.
2 · Interpret it as an emergency
Ambiguous situations get read against other people's reactions. If everyone looks calm, you conclude it's fine — pluralistic ignorance. The smoke-filled room dies right here: the smoke is obvious, but the social read says "no big deal." Failure here: you decide it isn't really an emergency.
3 · Assume personal responsibility
Even seeing a real emergency, you have to feel it's yours to handle. In a crowd, responsibility spreads thin across everyone — diffusion of responsibility. Alone, it's all on you, so you move. Failure here: "someone else will take care of it."
4 · Know how to help
Willing isn't enough; you need to know what to do. Not knowing CPR, not speaking the language, not seeing a safe way in — any of these can stall a person who genuinely wants to act. Failure here: you care, but you're stuck for a workable move.
5 · Decide to help
The last gate weighs costs against benefits in the moment: danger, embarrassment (audience inhibition), time, legal worry. Only if the decision comes out "yes" do you actually intervene. Failure here: the felt cost of acting wins.
How robust is the effect, really?
Because the Genovese story got debunked, students sometimes assume the whole idea collapsed with it. It didn't. A large meta-analysis pulling together decades of studies confirmed that the bystander effect is real and reliable: on average, people are less likely to help as the number of bystanders grows (Fischer et al., 2011). So the core finding stands on far firmer ground than one 1964 newspaper article.
But that same meta-analysis added a crucial wrinkle the textbooks often skip. The effect is weaker — sometimes absent — in clearly dangerous emergencies. When a situation is unambiguous and physically threatening, the presence of others can matter much less, and sometimes bystanders even help more (extra people mean extra hands and safety in numbers). The bystander effect bites hardest in ambiguous situations, where you're unsure whether help is needed and can talk yourself out of it. When the emergency is obvious and grave, people are more likely to act regardless of the crowd.
Where the field is now — and it's more hopeful than the myth
Recent research went looking at what people actually do in real public conflicts, using security-camera footage from cities across several countries. The finding cuts against the gloomy caricature: in roughly nine out of ten real public conflicts captured on CCTV, at least one bystander — on average nearly four — intervened to help (Philpot et al., 2020). Real crowds are not the frozen, apathetic mob the Genovese legend implied. The lab effect is real and the mechanisms are real, but in the wild, humans help each other far more often than the myth ever let us believe. For how the field's classic stories get re-examined this way, see Science Corrects Itself.
The situation beats the sermon: Darley & Batson (1973)
If you want the single most elegant demonstration that helping is a situation story, here it is. Darley and Batson recruited students at Princeton Theological Seminary — people literally training for a life of moral service — and, on the way to another building, sent each one past a man slumped in a doorway, coughing and groaning. The clever twist: some seminarians were told they had plenty of time; others were told they were already late and should hurry. To sharpen the irony, some were on their way to deliver a talk on the parable of the Good Samaritan — the very story about stopping for a stranger in need.
The results are almost too on-the-nose. Whether a seminarian's religiosity was intrinsic or extrinsic predicted essentially nothing. What predicted helping was the clock. About 63% of those not in a hurry stopped to help; only about 10% of those in a hurry did. Some in the rushed condition literally stepped over the groaning man on their way to give a sermon about helping a stranger in a ditch. The lesson lands with a thud: being a good person in the abstract — even a person studying to be good for a living — is a weak predictor of behavior compared with the mundane pressure of the moment. Time pressure jammed the earliest gates of the model: many hurried seminarians never fully registered the man as someone in need — and interviews suggested some noticed but hurried past in conflict, torn between two obligations.
Bodhi says
Don't file the seminarian study under "religious people are hypocrites." That's the dispositional misread. The point is the opposite and far more humbling: put good, sincere people under a small situational squeeze — being a few minutes late — and their goodness quietly stops predicting what they do. The next time you walk past someone in need while rushing, remember you're not a bad person. You're a person in a hurry. That's precisely why the finding matters.
Why we help at all: the great motive debate
Shift now from whether to why. Grant that people do help — the question that has split the field for decades is whether we ever help for genuinely selfless reasons, or whether "altruism" always cashes out to self-interest underneath. This is not settled, and pretending it is would be dishonest.
On one side, C. Daniel Batson's empathy-altruism hypothesis: when we feel genuine empathy for someone — when we actually take their perspective and feel with them — we help for their sake, not ours, even when we could easily walk away at no cost. In the Toi and Batson (1982) study, participants heard about "Carol," a student who'd broken both legs and fallen behind. High-empathy participants helped whether or not they'd ever see her again; low-empathy participants helped only when it was costly to refuse (they'd face her in class). If helping were purely selfish, high-empathy people should have bailed the moment escape was easy — but they didn't. Batson reads that as fingerprints of real altruism.
On the other side sits the egoistic family of accounts, associated with social-exchange thinking generally. The arousal: cost–reward model (Piliavin and colleagues) and Robert Cialdini's negative-state-relief idea both argue that seeing someone suffer makes us feel bad, and helping is how we relieve our own distress. On this view, empathy doesn't reveal selflessness — it just raises the personal cost of walking away, so of course empathic people help more. Every "altruistic" act, the argument goes, buys the helper something: relief, social approval, self-worth, a future favor. There's no pure gift; there's just a well-disguised trade.
Spelled out, that trade is social exchange theory applied to helping: people help when the rewards outweigh the costs, and no genes or evolutionary history are needed to make the prediction. On the reward side sit the likelihood of being helped in return, relief from the distress of watching someone suffer, and the boost to social approval and self-esteem that comes from being seen as a good person. On the cost side sit physical risk, embarrassment, and plain time and effort. The uncomfortable implication is the one Batson set out to refute: if the ledger is all there is, pure altruism does not exist — people help exactly when, and only when, it pays.
Why this is genuinely hard to settle
The two camps often predict the same behavior and disagree only about the invisible motive behind it. Batson spent a career designing experiments to rule out egoistic explanations one at a time — escape being easy, mood already fixed, no one watching — and empathic people kept helping anyway. Critics keep proposing one more selfish loophole. It's a model case of how you falsify a motive you can't directly observe: not with a single knockout study, but by slowly closing off the alternatives. The current honest verdict: the evidence tilts toward at least some genuine altruism, but reasonable researchers still disagree. Hold it as an open question, not a solved one.
The deeper roots, and the everyday tilts
Step back further and evolutionary accounts offer two tidy reasons helping could be built in. Kin selection: behavior that helps close relatives passes along shared genes, so "self-sacrifice" for kin can be favored by natural selection — you're helping copies of your own genes survive. Reciprocity (reciprocal altruism): helping non-kin pays off when the favor is likely to come back around, so cooperation can evolve even among strangers who expect future contact. Neither requires anyone to consciously calculate genes or paybacks; the tendencies just outcompeted their alternatives over deep time. Ground squirrels giving alarm calls and vampire bats sharing blood meals are the classic non-human examples. Notice why altruism is an evolutionary puzzle in the first place: a trait that lowers your own survival or reproduction should be weeded out by ordinary natural selection, so kin selection and reciprocity are the two ways to explain how it survived anyway. A third, more human-specific argument adds that individuals who readily learn and internalize their group's social norms — including norms of helping — out-reproduced those who didn't, so a disposition to absorb "help when you can" became part of our inheritance. Two such norms show up in every culture: the reciprocity norm (return help you've received) and the social responsibility norm (help those who depend on you, payback or not).
Where you live matters too. Stanley Milgram (1970) proposed the urban overload hypothesis: city dwellers help less not because they care less but because they are bombarded with stimulation and cope by narrowing their attention and keeping to themselves — which jams the "notice" gate of the model above. Consistent with that, people in smaller and less dense places help strangers more. And it isn't only city versus town: people who have stayed put in one community for years help more than recent arrivals, because long residence builds social ties and a felt stake in the place (Oishi et al., 2007). Rootedness, not just population, predicts who stops.
Closer to daily life, two smaller but reliable tilts are worth knowing. Mood matters: people in a good mood help more (the "feel-good, do-good" effect), and, oddly, people in a bad mood sometimes help more too — because a kind act can relieve their gloom (negative-state relief again). Guilt is the sharpest case: shoppers led to believe they had broken a stranger's camera were far more willing, minutes later, to help a different person pick up a spilled bag (Regan et al., 1972) — helping as a way to settle an internal account. And who gets helped isn't random. We help more readily those we see as similar, likable, and — pointedly — as deserving. In a striking field study, a man asking passersby for money received more than twice as much when dressed in a suit than in jeans (Callaghan et al., 2022): the very same request, read as more "deserving" from someone who looked high-status. Helping, in other words, is shot through with our snap judgments about who has earned it.
Check yourself
The one thing to carry out of this topic
Two truths, held together. First, whether a good person helps in a given moment depends far less on how good they are than on the situation they're standing in — how many others are watching, how ambiguous the trouble looks, whether they're in a hurry. That's the bystander effect and the seminarian study, and it's genuinely humbling. But the second truth is the hopeful correction the myth buried: real people, in real emergencies, help each other far more than the "38 apathetic witnesses" legend ever let on (Philpot et al., 2020). Know the mechanisms that can freeze you — diffusion, pluralistic ignorance, audience inhibition — and you're better equipped to override them. Naming the trap is the first step out of it.
References
Callaghan, B., Delgadillo, Q., & Kraus, M. W. (2022). The influence of signs of social class on compassionate responses to people in need. Frontiers in Psychology, 13, 936170. https://doi.org/10.3389/fpsyg.2022.936170
Darley, J. M., & Batson, C. D. (1973). "From Jerusalem to Jericho": A study of situational and dispositional variables in helping behavior. Journal of Personality and Social Psychology, 27(1), 100–108. https://doi.org/10.1037/h0034449
Darley, J. M., & Latané, B. (1968). Bystander intervention in emergencies: Diffusion of responsibility. Journal of Personality and Social Psychology, 8(4, Pt. 1), 377–383. https://doi.org/10.1037/h0025589
Fischer, P., Krueger, J. I., Greitemeyer, T., Vogrincic, C., Kastenmüller, A., Frey, D., Heene, M., Wicher, M., & Kainbacher, M. (2011). The bystander-effect: A meta-analytic review on bystander intervention in dangerous and non-dangerous emergencies. Psychological Bulletin, 137(4), 517–537. https://doi.org/10.1037/a0023304
Garcia, S. M., Weaver, K., Moskowitz, G. B., & Darley, J. M. (2002). Crowded minds: The implicit bystander effect. Journal of Personality and Social Psychology, 83(4), 843–853. https://doi.org/10.1037/0022-3514.83.4.843
Latané, B., & Darley, J. M. (1968). Group inhibition of bystander intervention in emergencies. Journal of Personality and Social Psychology, 10(3), 215–221. https://doi.org/10.1037/h0026570
Latané, B., & Darley, J. M. (1970). The unresponsive bystander: Why doesn't he help? Appleton-Century-Crofts.
Manning, R., Levine, M., & Collins, A. (2007). The Kitty Genovese murder and the social psychology of helping: The parable of the 38 witnesses. American Psychologist, 62(6), 555–562. https://doi.org/10.1037/0003-066X.62.6.555
Milgram, S. (1970). The experience of living in cities. Science, 167(3924), 1461–1468. https://doi.org/10.1126/science.167.3924.1461
Oishi, S., Rothman, A. J., Snyder, M., Su, J., Zehm, K., Hertel, A. W., Gonzales, M. H., & Sherman, G. D. (2007). The socioecological model of procommunity action: The benefits of residential stability. Journal of Personality and Social Psychology, 93(5), 831–844. https://doi.org/10.1037/0022-3514.93.5.831
Philpot, R., Liebst, L. S., Levine, M., Bernasco, W., & Lindegaard, M. R. (2020). Would I be helped? Cross-national CCTV footage shows that intervention is the norm in public conflicts. American Psychologist, 75(1), 66–75. https://doi.org/10.1037/amp0000469
Regan, D. T., Williams, M., & Sparling, S. (1972). Voluntary expiation of guilt: A field experiment. Journal of Personality and Social Psychology, 24(1), 42–45. https://doi.org/10.1037/h0033553
Toi, M., & Batson, C. D. (1982). More evidence that empathy is a source of altruistic motivation. Journal of Personality and Social Psychology, 43(2), 281–292. https://doi.org/10.1037/0022-3514.43.2.281