Although emotional responses such as guilt, personal distress, and empathy are important determinants of altruism, it is the social situation itself—the people around us when we are deciding whether or not to help—that has perhaps the most important influence on whether and when we help.

Consider the unusual case of the killing of 28-year-old Katherine “Kitty” Genovese in New York City at about 3:00 a.m. on March 13, 1964. Her attacker, Winston Moseley, stabbed and sexually assaulted her within a few yards of her apartment building in the borough of Queens. During the struggle with her assailant, Kitty screamed, “Oh my God! He stabbed me! Please help me!” But no one responded. The struggle continued; Kitty broke free from Moseley, but he caught her again, stabbed her several more times, and eventually killed her.

The murder of Kitty Genovese shocked the nation, in large part because of the (often inaccurate) reporting of it. Stories about the killing in the New York Times and other papers indicated that as many as 38 people had overheard the struggle and killing, that none of them had bothered to intervene, and that only one person had even called the police, long after Genovese was dead.

Although these stories about the lack of concern by people in New York City proved to be false (Manning, Levine, & Collins, 2007), they nevertheless led many people to think about the variables that might lead people to help or, alternatively, to be insensitive to the needs of others. Was this an instance of the uncaring and selfish nature of human beings? Or was there something about this particular social situation that was critical? It turns out, contrary to most people’s expectations, that having many people around during an emergency can in fact be the opposite of helpful—it can reduce the likelihood that anyone at all will help.

Latané and Darley’s Model of Helping

Figure 8.8 Latané and Darley’s Stages of Helping

Two social psychologists, Bibb Latané and John Darley, found themselves particularly interested in, and concerned about, the Kitty Genovese case. As they thought about the stories that they had read about it, they considered the nature of emergency situations, such as this one. They realized that emergencies are unusual and that people frequently do not really know what to do when they encounter one. Furthermore, emergencies are potentially dangerous to the helper, and it is therefore probably pretty amazing that anyone helps at all.

To better understand the processes of helping in an emergency, Latané and Darley developed a model of helping that took into consideration the important role of the social situation. Their model, which is shown in Figure 8.8, “Latané and Darley’s Stages of Helping,” has been extensively tested in many studies, and there is substantial support for it.

Noticing

Latané and Darley thought that the first thing that had to happen in order for people to help is that they had to notice the emergency. This seems pretty obvious, but it turns out that the social situation has a big impact on noticing an emergency. Consider, for instance, people who live in a large city such as New York City, Bangkok, or Beijing. These cities are big, noisy, and crowded—it seems like there are a million things going at once. How could people living in such a city even notice, let alone respond to, the needs of all the people around them? They are simply too overloaded by the stimuli in the city (Milgram, 1970).

Many studies have found that people who live in smaller and less dense rural towns are more likely to help than those who live in large, crowded, urban cities (Amato, 1983; Levine, Martinez, Brase, & Sorenson, 1994). Although there are a lot of reasons for such differences, just noticing the emergency is critical. When there are more people around, it is less likely that the people notice the needs of others.

You may have had an experience that demonstrates the influence of the social situation on noticing. Imagine that you have lived with a family or a roommate for a while, but one night you find yourself alone in your house or apartment because your housemates are staying somewhere else that night. If you are like me, I bet you found yourself hearing sounds that you never heard before—and they might have made you pretty nervous. Of course, the sounds were always there, but when other people were around you, you were simply less alert to them. The presence of others can divert our attention from the environment—it’s as if we are unconsciously, and probably quite mistakenly, counting on the others to take care of things for us.

Latané and Darley (1968) wondered if they could examine this phenomenon experimentally. To do so, they simply asked their research participants to complete a questionnaire in a small room. Some of the participants completed the questionnaire alone, while others completed the questionnaire in small groups in which two other participants were also working on questionnaires.

A few minutes after the participants had begun the questionnaires, the experimenters started to release some white smoke into the room through a vent in the wall while they watched through a one-way mirror. The smoke got thicker as time went on, until it filled the room. The experimenters timed how long it took before the first person in the room looked up and noticed the smoke. The people who were working alone noticed the smoke in about five seconds, and within four minutes most of the participants who were working alone had taken some action. But what about the participants working in groups of three? Although we would certainly expect that having more people around would increase the likelihood that someone would notice the smoke, on average, the first person in the group conditions did not notice the smoke until over 20 seconds had elapsed. And although 75% of the participants who were working alone reported the smoke within four minutes, the smoke was reported in only 12% of the three-person groups by that time. In fact, in only three of the eight three-person groups did anyone report the smoke at all, even after it had entirely filled the room!

Interpreting

Even if we notice an emergency, we might not interpret it as one. The problem is that events are frequently ambiguous, and we must interpret them to understand what they really mean. Furthermore, we often don’t see the whole event unfolding, so it is difficult to get a good handle on it. Is a man holding an iPhone and running away from a group of pursuers a criminal who needs to be apprehended, or is this just a harmless prank? Is someone stumbling around outside a nightclub drunk, or going into a diabetic coma? Were the cries of Kitty Genovese really calls for help, or were they simply an argument with a boyfriend? It’s hard for us to tell when we haven’t seen the whole event (Piliavin, Piliavin, & Broll, 1976). Moreover, because emergencies are rare and because we generally tend to assume that events are benign, we may be likely to treat ambiguous cases as not being emergencies.

The problem is compounded when others are present because when we are unsure how to interpret events we normally look to others to help us understand them (this is informational social influence). However, the people we are looking toward for understanding are themselves unsure how to interpret the situation, and they are looking to us for information at the same time we are looking to them.

When we look to others for information we may assume that they know something that we do not know. This is often a mistake, because all the people in the situation are doing the same thing. None of us really know what to think, but at the same time we assume that the others do know. Pluralistic ignorance occurs when people think that others in their environment have information that they do not have and when they base their judgments on what they think the others are thinking.

Pluralistic ignorance seems to have been occurring in Latané and Darley’s studies, because even when the smoke became really heavy in the room, many people in the group conditions did not react to it. Rather, they looked at each other, and because nobody else in the room seemed very concerned, they each assumed that the others thought that everything was all right. You can see the problem—each bystander thinks that other people aren’t acting because they don’t see an emergency. Of course, everyone is confused, but believing that the others know something that they don’t, each observer concludes that help is not required.

Pluralistic ignorance is not restricted to emergency situations (Miller, Turnbull, & McFarland, 1988; Suls & Green, 2003). Maybe you have had the following experience: You are in one of your classes and the instructor has just finished a complicated explanation. He is unsure whether the students are up to speed and asks, “Are there any questions?” All the class members are of course completely confused, but when they look at each other, nobody raises a hand in response. So everybody in the class (including the instructor) assumes that everyone understands the topic perfectly. This is pluralistic ignorance at its worst—we are all assuming that others know something that we don’t, and so we don’t act. The moral to instructors in this situation is clear: wait until at least one student asks a question. The moral for students is also clear: ask your question! Don’t think that you will look stupid for doing so—the other students will probably thank you.

Taking Responsibility

Even if we have noticed the emergency and interpret it as being one, this does not necessarily mean that we will come to the rescue of the other person. We still need to decide that it is our responsibility to do something. The problem is that when we see others around, it is easy to assume that they are going to do something and that we don’t need to do anything. Diffusion of responsibility occurs when we assume that others will take action and therefore we do not take action ourselves. The irony, of course, is that people are more likely to help when they are the only ones in the situation than they are when there are others around.

Darley and Latané (1968) had study participants work on a communication task in which they were sharing ideas about how to best adjust to college life with other people in different rooms using an intercom. According to random assignment to conditions, each participant believed that he or she was communicating with either one, two, or five other people, who were in either one, two, or five other rooms. Each participant had an initial chance to give his or her opinions over the intercom, and on the first round one of the other people (actually a confederate of the experimenter) indicated that he had an “epileptic-like” condition that had made the adjustment process very difficult for him. After a few minutes, the subject heard the experimental confederate say,

I-er-um-I think I-I need-er-if-if could-er-er-somebody er-er-er-er-er-er-er give me a liltle-er-give me a little help here because-er-I-er-I’m-er-er having a-a-a real problcm-er-right now and I-er-if somebody could help me out it would-it would-er-er s-s-sure be-sure be good…because there-er-er-a cause I-er-I-uh-I’ve got a-a one of the-er-sei er-er-things coming on and-and-and I could really-er-use some help so if somebody would-er-give me a little h-help-uh-er-er-er-er-er c-could somebody-er-er-help-er-uh-uh-uh (choking sounds).…I’m gonna die-er-er-I’m…gonna die-er-help-er-er-seizure-er- (chokes, then quiet). (Darley & Latané, 1968, p. 379)

As you can see in Table 8.2, “Effects of Group Size on Likelihood and Speed of Helping,” the participants who thought that they were the only ones who knew about the emergency (because they were only working with one other person) left the room quickly to try to get help. In the larger groups, however, participants were less likely to intervene and slower to respond when they did. Only 31% of the participants in the largest groups responded by the end of the six-minute session.

You can see that the social situation has a powerful influence on helping. We simply don’t help as much when other people are with us.

Table 8.2 Effects of Group Size on Likelihood and Speed of Helping

Perhaps you have noticed diffusion of responsibility if you have participated in an Internet users’ group where people asked questions of the other users. Did you find that it was easier to get help if you directed your request to a smaller set of users than when you directed it to a larger number of people? Consider the following: In 1998, Larry Froistad, a 29-year-old computer programmer, sent the following message to the members of an Internet self-help group that had about 200 members. “Amanda I murdered because her mother stood between us…when she was asleep, I got wickedly drunk, set the house on fire, went to bed, listened to her scream twice, climbed out the window and set about putting on a show of shock and surprise.” Despite this clear online confession to a murder, only 3 of the 200 newsgroup members reported the confession to the authorities (Markey, 2000).

To study the possibility that this lack of response was due to the presence of others, the researchers (Markey, 2000) conducted a field study in which they observed about 5,000 participants in about 400 different chat groups. The experimenters sent a message to the group, from either a male or female screen name (JakeHarmen or SuzyHarmen). Help was sought by either asking all the participants in the chat group, “Can anyone tell me how to look at someone’s profile?” or by randomly selecting one participant and asking “[name of selected participant], can you tell me how to look at someone’s profile?” The experimenters recorded the number of people present in the chat room, which ranged from 2 to 19, and then waited to see how long it took before a response was given.

It turned out that the gender of the person requesting help made no difference, but that addressing to a single person did. Assistance was received more quickly when help was asked for by specifying a participant’s name (in only about 37 seconds) than when no name was specified (51 seconds). Furthermore, a correlational analysis found that when help was requested without specifying a participant’s name, there was a significant negative correlation between the number of people currently logged on in the group and the time it took to respond to the request.

Garcia, Weaver, Moskowitz, and Darley (2002) found that the presence of others can promote diffusion of responsibility even if those other people are only imagined. In these studies, the researchers had participants read one of three possible scenarios that manipulated whether participants thought about dining out with 10 friends at a restaurant (group condition) or whether they thought about dining at a restaurant with only one other friend (one-person condition). Participants in the group condition were asked to “Imagine you won a dinner for yourself and 10 of your friends at your favorite restaurant.” Participants in the one-person condition were asked to “Imagine you won a dinner for yourself and a friend at your favorite restaurant.”

After reading one of the scenarios, the participants were then asked to help with another experiment supposedly being conducted in another room. Specifically, they were asked: “How much time are you willing to spend on this other experiment?” At this point, participants checked off one of the following minute intervals: 0 minutes, 2 minutes, 5 minutes, 10 minutes, 15 minutes, 20 minutes, 25 minutes, and 30 minutes.

Figure 8.9 Helping as a Function of Imagined Social Context

Garcia et al. (2002) found that the presence of others reduced helping, even when those others were only imagined.

As you can see in Figure 8.9, “Helping as a Function of Imagined Social Context,” simply imagining that they were in a group or alone had a significant effect on helping, such that those who imagined being with only one other person volunteered to help for more minutes than did those who imagined being in a larger group.

Implementing Action

The fourth step in the helping model is knowing how to help. Of course, for many of us the ways to best help another person in an emergency are not that clear; we are not professionals and we have little training in how to help in emergencies. People who do have training in how to act in emergencies are more likely to help, whereas the rest of us just don’t know what to do and therefore may simply walk by. On the other hand, today most people have cell phones, and we can do a lot with a quick call. In fact, a phone call made in time might have saved Kitty Genovese’s life. The moral: you might not know exactly what to do, but you may well be able to contact someone else who does.

Latané and Darley’s decision model of bystander intervention has represented an important theoretical framework for helping us understand the role of situational variables on helping. Whether or not we help depends on the outcomes of a series of decisions that involve noticing the event, interpreting the situation as one requiring assistance, deciding to take personal responsibility, and deciding how to help.

Fischer and colleagues (2011) analyzed data from over 105 studies using over 7,500 participants who had been observed helping (or not helping) in situations in which they were alone or with others. They found significant support for the idea that people helped more when fewer others were present. And supporting the important role of interpretation, they also found that the differences were smaller when the need for helping was clear and dangerous and thus required little interpretation. They also found that there were at least some situations (such as when bystanders were able to help provide the needed physical assistance) in which having other people around increased helping.

Although the Latané and Darley model was initially developed to understand how people respond in emergencies requiring immediate assistance, aspects of the model have been successfully applied to many other situations, ranging from preventing someone from driving drunk to making a decision about whether to donate a kidney to a relative (Schroeder, Penner, Dovidio, & Piliavin, 1995).