Psychology: Social psychology: Case study: 3: Good Samaritan

Psychology: Social psychology: Case study: 3: Good Samaritan

Authors: Piliavin et al. (1969)

Key term:  Subway Samaritans

Background: In 1964, Kitty Genovese was brutally stabbed and murdered in a neighborhood with apparently approximately 38 witnesses, where no-one called the police or went out to help, causing psychologists to conduct research into this behaviour which later become known as the bystander effect - when individuals do not offer help in the presence of other bystanders; and diffusion of responsibility - the bigger the group, the less likely people are to help.

Aim/Hypothesis: There are four aims to this experiment:

1. Test diffusion of responsibility in a real-life setting.
2. The effect of the type and race of victim on the speed, frequency and race of helper. Would someone who was ill more likely to get help than someone's who was drunk (at their own fault)? Are people more likely to help someone of their own race?
3. Effects of modelling - Are people more likely to help when they see someone else helping?
4. Examine relationship between size of group, frequency and time of helping response.

Method: Field experiment and non-participant, naturalistic observation (conductor of experiment not in the experiment).

Variables: 

• Independent variables:

1. Type of victim (ill/drunk)
2. Race of victim (black/white)
3. Model conditions
   -Critical-early: Model in same carriage as victim, helps 70 seconds after victim falls.
   -Critical-late: Model in same carriage as victim, helps 150 seconds after victim falls.
   -Adjacent-early: Model in adjacent carriage as victim, helps 70 seconds after victim falls.
   -Adjacent-late: Model in adjacent carriage as victim, helps 150 seconds after victim falls. 

• Dependent variables:

1. Frequency of helping
2. Speed of helper
3. Race of helper
4. Gender of helper
5. Comments made by passengers

Design: Independent groups - people on the subway only experienced an ill black trial, ill white trial, drunk black trial or drunk white trial. 

Setting: 

• New York City 
• 7 1/2 minute train journey 
• IND (independent line) from 59th street station to 125th street station.

Experimenters: Students from Columbia University. 

4 teams of 4; 1 male victim, 1 male model, 2 female observers. In total; 

• 4 male victims (1 black and 3 white, aged 26-35)
• 4 male models (all white, aged 24-29) 
• 8 female observers. 

Participants/ Sampling technique: 

• 4450 men and women traveling between 11am and 3pm. 
• Unaware they were in an experiment. 
• 45% black and 55% white. 
• Sample was self-selecting because it consisted of participants who were on the train at that time.

Apparatus: 

• Subway train (old model which observers could sit in the same place each time) which only had 13 seats. 
• All victims dressed in Eisenhower jackets and old slacks (no tie). 
• Victim in ill condition: black cane
• Victim in drunk condition: smelled of alcohol, had a liquor bottle in a brown bag. 

Controls:

• Same 7 1/2 train journey for all trials
• Victims wore same clothes and fell over at the same time (after 70 seconds) in the same place and the same way.
• Started the journey in the same place (observer 1 in adjacent carriage near exit, observer 2 in adjacent carriage in far corner.

Procedure

1. Members of the team of four position themselves in the specific locations on the train.
2. Subway leaves the station, 70 seconds later, victim (black/white and ill/drunk) staggers forward, collapses and remains on the floor, staring at the ceiling of the carriage.
3. If no one helps, the model intervenes. One of the four model conditions used (above in independent variables).
4. Observer 1 records: Gender, race and location of passengers (seated or standing) in critical carriage, total number of passengers, total number who went to help.
   Observer 2 records: Gender, race and location of passengers in adjacent area, time taken for first passenger to help, time taken for someone to help the model.
   Both observers record comments by passengers sitting next to them. 
5. If no one helps, the model helps the victim to his feet. At the next station the team of four get off the train, cross over and repeat procedure on the train going in the opposite direction. 6-8 trials completed in a day.

Data:

• Quantitative data: Demographic characteristics (gender, race), frequency of helping, speed of helping, etc.
• Qualitative data: Comments made by passengers

Findings:`

• 43 people present in each carriage
• Total 103 trials
• 78% of victims received spontaneous help
• Ill victim: Spontaneous help on 62 out of 63 trials - model only helped 3 times. Median helping time was 5 seconds.
• Drunk victim: Spontaneous help on 19 out of 38 trials. Median helping time was 109 seconds.
• When spontaneous help was given, on 60% of the 81 trials two, three or more helpers joined. No difference between black or white, ill or white victim. If one person helped others join.
• 60% of first helpers in critical area were male; 90% of people helping were male 
• No significant difference in race of helpers - 64% of first helpers white.
• Same race helping:
  When victim was white, 68% of first helpers were white; when victim was black, only 50% of first helpers were white - Tendency for same race helping
  Ill condition - no difference in race of helpers.
  Drunken condition - mainly members of the same race
• People left the critical area only 20% of the time - 34 people; more people left when the victim was drunk than ill.
• Most comments happened during drunk trials - especially when no one helped after 70 seconds.
  - May be due to discomfort of not helping, needing to justify inaction eg. "It's for men to help", "I'm not strong enough".

No diffusion of responsibility - In this study, helping was faster when there were more people.

• Victim and witnesses were face to face; unlike laboratory experiments. 

Conclusion

• Individuals who appear to be ill receive more help than a drunk person.
• Men are more likely to help than women - however, this could be a weakness of generalization, because in 1969, gender roles were different; women were still not seen as completely equal to men.
• Same-race helping is more likely, especially when victim is drunk.
• No strong relationship between number of bystanders and speed of helping - expected diffusion of responsibility not observed.
• The longer the emergency continues without help, the less impact a model has, and people are more likely to leave the immediate area.

Study results

Model of response in emergency situations

• Observing any emergency situation causes arousal in a bystander.
• Arousal: Unpleasant feeling (eg. sympathy, fear) that a bystander feels a need to reduce. Arousal can be higher if witness can relate to victim (eg. same race), the bigger the emergency, and the longer the situation continues.
• Arousal can be reduced by helping, going to get help, leaving the scene or concluding the victim doesn't deserve help.

Cost-reward matrix

• Determines response - People weigh up the costs and benefits before making a decision to help.
• Costs of helping - eg. possible physical harm
• Benefits of helping - eg. social approval 
• Benefits of not helping - eg. getting to work on time

Strengths

• Ecological validity - Could happen in real life.
• Behaviour is valid - Passengers did not know they were in an experiment, so behaviour is natural, although this is unethical, because passengers are deceived and there is no informed consent.

Weaknesses

• IV may not be affecting the DV - Situational variables are difficult to control in a field experiment, and it could be an EV affecting the DV - Eg. Positioning of people in carriage could not be controlled, so some people may not have noticed the incident, which could affect the level of helping.
• Unethical - Participants did not know they were in a study - no informed consent and deception.