Piliavin, Irving., Rodin, Judith., & Piliavin, Jane.  (1969). 

Good Samaritanism:  An Underground Phenomenon?




Piliavin et al. note in their study that social psychologists became particularly interested in the behaviour of bystanders following the case of the brutal murder of Kitty Genovese in New York in 1964.   The murder attracted interest from psychologists because not one person out of the 40 people, who witnessed the attack lasting over half an hour, tried to help or contacted the police. 

Many laboratory studies were carried out by social psychologists to test bystander apathy.  That is the phenomenon of when observers of an emergency situation do not intervene.  Importantly social psychologists looked for the cause of bystander behaviour not in the type of people who do or do not help but in the situational factors which influence helping behaviour

Two important concepts investigated by social psychologists were diffusion of responsibility and pluralistic ignorance.

Diffusion of responsibility is the idea that people are less likely to offer to help someone if there are others present, because they perceive responsibility as being shared between all present, and therefore see themselves as being less personally responsible.

Pluralistic ignorance is the tendency for people in a group to mislead each other about an emergency situation.  For example, a person might perceive an emergency as a non-emergency because others are remaining calm and not taking action.

The early laboratory experiments into bystander apathy consisted of  candid camera/trigger happy style scenarios where people were placed in situations such as a smoke filled room to investigate if people would sound the alarm or not.  (e.g. Latane and Darley 1968).  These experiments usually consisted of one participant and a number of confederates. One of the findings of such laboratory experiments was that people did demonstrate diffusion of responsibility.  That is they were less likely to help as the number of bystanders increased.

However Piliavin et al. recognised that these laboratory experiments lacked ecological validity in that they did not demonstrate how people would react in a realist situation.  They therefore planned to investigate helping behaviour using a field experiment where they could observe behaviour as it is in the real everyday world.



  The aim of the study was to investigate factors affecting helping behaviour.

The factors they were interested included

(i) The type of victim (drunk or ill)

(ii) The race of the victim (black or white)

(iii) The speed of helping

(iv) The frequency of helping

(v) The race of the helper. 

Importantly the field experiment also investigated the impact of the presence of a model (someone who offers help first), as well as the relationship between the size of the group and frequency of helping.



The method used was a type of field experiment using participant observation.

The participants were approximately 4450 men and women travelling on a particular stretch of the New York underground system between 11 a.m. and 3 p.m. on weekdays during the period of April 15th to June 26th, 1968. 

The average racial composition of the passengers on the train, which travelled through Harlem to the Bronx, was 45% black and 55% white.  The average number of people in the train carriage was 43, and the average number of people in the critical area where the incident was staged was 8.5.

Two particular trains were selected for the study.  The trains were chosen because they did not make any stops between 59th Street and 125th Street. This means that for about 7.5 minutes the participants were a captive audience to the emergency.

Therefore a single trial was a non-stop, 7.5-minute journey in either direction.

On each trial, a team of four students, (two males and two females), boarded the train using different doors.  Four dif­ferent teams, whose members always worked together, collected data for 103 trials.  Each team varied the location of the experimental compart­ment from trial to trial.  The female confederates sat outside the critical area and recorded data as unobtrusively as possible during the journey, while the male model and victim remained standing.  The victim always stood next to a pole in the centre of the critical area (See Figure 1).  As the train passed the first station (approximately 70 seconds after depart­ing), the victim staggered forward and collapsed.  Until receiving help, he remained motionless on the floor, looking at the ceiling.  If he received no help by the time the train slowed to a stop, the model helped him to his feet.  At the stop, the team got off and waited separately until other pas­sengers had left the station before proceeding to another platform to board a train going in the opposite direction for the next trial.  Six to eight trials were run on any given day and all trials on a given day were in the same ‘victim condition’.


Figure 1:  Layout of adjacent and critical areas of subway car




The four victims (one from each team) were males, aged between 26 and 35, three white, one black, all identically dressed in Eisenhower jackets, old slacks and no tie.  On 38 trials the victims smelled of alcohol and car­ried a bottle of alcohol wrapped in a brown bag (drunk condition), while on the remaining 65 trials they appeared sober and carried a black cane (cane condition).  In all other aspects, victims behaved identically in the two conditions, and each victim participated in drunk and cane trials.  (There were more cane than drunk trials because one of the teams of students ‘didn’t like’ playing the drunk victim.

The models (white males aged 24 to 29) were all casually, but not identically, dressed.  There were four different model conditions used across both drunk and cane victim conditions:

  1. Critical area - early:  the model would stand in the critical area and wait until passing the fourth station before he helped the victim (approximately 70 seconds after the collapse).

  2. Critical area - late:  the model would stand in the critical area and wait until passing the sixth station before he helped the victim (approximately 150 seconds after the collapse).

  3. Adjacent area - early:  the model would stand in the adjacent area and waited until passing the fourth station before he helped the victim.

  4. Adjacent area - late:  the model would stand in the adjacent area and waited until passing the sixth station before he helped the victim.


When the model intervened, he helped the victim to a sitting position and stayed with him for the remainder of the trial. 

 A number of observations were recorded.

They observers noted the total number of passengers who came to the victim’s assistance, plus their race, sex and location. 

The observers also noted the race, sex and location of every pas­senger, seated or standing, in the critical and adjacent areas.

A second observer also noted how long it took for help to arrive

The observers also recorded comments made by nearby passengers and also tried to elicit comments from a passenger sitting next to them.



Helping behaviour was very high and much higher than earlier laboratory studies.  Therefore it was not possible to investigate the effects of the model’s helping because on the majority of the trials the victims were helped before the model acted.

The cane victim received spontaneous help on 62 out of the 65 trials, and the drunk victim received spontaneous help on 19 out of 38 trials. 

On 60% of the 81 trials where spontaneous help was given, more than one person offered help.  Once one person had started to help, there were no differences for different victim conditions (black/white, cane/drunk) on the number of extra helpers that appeared.  The race of the victims made no significant difference to helping behaviour, but there was a slight tendency for same-race helping in the drunken condition.

It was found that 90% of helpers were male.  Although there were more men present, this percentage was statistically significant;

It was also found that 64% of the helpers were white; this was what would be expected based on the racial distribution of the carriage.

Diffusion of responsibility was not evident.  The diffusion of responsibility hypothesis predicts that helping behaviour would decrease as the number of bystanders increases.   In fact the field experiment found that the quickest help came from the largest groups.  However, in the earlier laboratory experiments there was only one participant and the other bystanders were confederates. 

Nobody left the carriage during the incident (mainly because the train was moving), on 21 of the 103 trials a total of 34 people left the critical area, particularly when the victim appeared to be drunk.

More comments were obtained on drunk than cane trials and most of these were obtained when no one helped until after 70 seconds; this could be due to the discomfort passengers felt in sitting inactive in the presence of the victim, perhaps hoping that others would confirm that inaction was appropriate.

The following comments came from women passengers:  “It’s for men to help him”; “I wish I could help him - I’m not strong enough”; “I never saw this kind of thing before - I don’t know where to look”; “You feel so bad that you don’t know what to do”.



Piliavin et al. developed a model to explain their results. Called the Arousal: Cost – Reward Model

They argue that firstly, observation of an emergency situation creates an emotional arousal in bystanders.  This arousal may be perceived as fear, disgust or sympathy, depending on aspects of the situation.  (Note the similarity with Schachter and Singer’s two factor theory of emotion) 

This state of arousal can be increased by a number of factors including:

empathy with the victim (i.e. whether you can perceive yourself in the victims situation);
being close to the emergency;
the length of time the emergency continues for.

This state of arousal can be reduced by a number of factors including: 

seeking help from another source;
leaving the scene;
deciding the person doesn’t need or deserve help.

Therefore according to this model we are motivated to help people not by altruism (acting in the interest of others) but as a way of reducing unpleasant feelings of arousal.

Piliavin et al. go on to argue that the chosen response depends on a cost-reward analysis by the individual.  These include:

Costs of helping, such as effort, embarrassment and possible physical harm;

Cost of not helping, such as self-blame and perceived censure from others;

Rewards of helping, such as praise from self, onlookers and the victim;

Rewards of not helping, such as getting on with one’s own business and not incurring the possible costs of helping. 

According to Piliavin et al. the results of their filed experiment can be explained using their Arousal: Cost – Reward Model.   For example:

The drunk is helped less often because the perceived cost is greater - helping a drunk is likely to cause disgust, embarrassment or harm.  The cost of not helping is less because nobody will blame another for not helping a drunk because he is perceived as partly responsible for his own victimisation.

Women help less often than men because the cost to them in terms of effort and danger is greater and, since it may not be seen as a woman’s role to offer assistance under these circumstances, the cost of not helping is less.

Diffusion of responsibility is not found in the cane-carrying situation because the cost of not helping is high and the cost of helping is low.

As time without help increases, so does the arousal level of the bystanders.  A late model is not copied because people have already chosen an alternative way of reducing arousal; they leave the area or engage in conversation with others in order to justify their lack of help.



Evaluation of Procedure/Method


 The study can be criticised on ethical grounds.  A problem with the field experiment is that the participants cannot give their consent, because they do not know that they are participants in an experiment.  Similarly the participants are being deceived because they are unaware that it is not a genuine emergency.  Participants were also not debriefed as this would have been almost impossible. Following from this it is possible that participants had feelings of guilt, distress, and anxiety. 

A further problem with field experiments is that they are more difficult to control than laboratory experiments.  For example we could question whether travellers on the trains saw more than one trial.  Field experiments are also more difficult to replicate and more time consuming and expensive.


A main strength of the study has to be its high level of ecological validity.  The study was done in a true to life environment and consisted of an incident, which could and does happen.  However, some of the participants were very close to the victim and were in a situation where they could not escape.  This is often unlike many other situations where we come across emergency situations and this may be one of the reasons why diffusion of responsibility did not occur. 

The sample size was also very large and we would assume a fairly representative sample of New Yorkers.   The researchers should therefore be able to generalise their findings with much more certainty than if they had carried out a study on say 40 students. 


Evaluation of Explanation

A criticism of the Arousal: Cost – Reward Model is that it takes a very negative view of people.  It denies that people act altruistically and assumes that behaviour is always measured in some form of cost or benefit.  Altruism refers to behaviours which are unselfish and motivated by another person’s needs.  Piliavin’s Arousal: Cost Benefit model assumes that we therefore never behave altruistically

Furthermore Piliavin et al. and a number of other psychologists chose to investigate the inactivity of the bystanders, but not why women are violently and sexually attacked by men regardless of the presence or absence of bystanders.  According to Banyard & Grayson (2000), “they appear to have gone to the theatre and described the audience without ever looking at the play.”  Kitty Genovese’s murderer was Winston Moseley, who had murdered three other women, raped at least four more, and attempted rape on another woman.  Apparently he had a ‘taste’ for raping dying women.  Banyard and Grayson (2000) argue that “surely the central problem that needs to be addressed is not the behaviour of the bystanders, but the behaviour of the murderer, and the construction of male sexuality that encourages grotesque acts of violence against women”.

Finally it is interesting to note that 40 years on from the murder of Kitty Genovese there is still considerable debate in New York about the accuracy of the original reporting of the events in the New York Times.  Perhaps there were not as many witnesses and some people may have called the police.



Piliavin, I.M., Rodin, J.A. & Piliavin, J. (1969) Good Samaritanism: An underground phenomenon? Journal of Personality and Social Psychology, 13, 289 -99)





GROSS, R. (1999) Key Studies in Psychology, 3rd Edition. London: Hodder and Stoughton

BANYARD, P. AND GRAYSON, A. (2000) Introducing Psychological Research; Seventy Studies that Shape Psychology, 2nd Edition. London: Macmillan







More Stuff

More pages.

And A Bit More Stuff

Some more pages.