To investigate if characteristics of a victim affects the help given in an emergency
Method
A male confederate - appearing either as drunk or disabled - collapsed on a subway. Was repeated 103 times
Results
The disabled victim was helped 95% of the time but the drunk victim was only helped 50% of the time. Help was just as likely in both empty and crowded carriages
Conclusion
Characteristics of a victim affects the help given. The number of onlookers doesn't affect help in a natural setting
Strength
Participants didn't know that their behaviour was being observed so acted more naturally
Weakness
The sample was only made up of people from a city who may be used to emergencies and therefore the conclusions may not be generalisable
