Piliavin

Question 1

Aim

Answer 1

measure effect that several variables hv on helping behaviour in emergency on underground subway train

Question 2

Research method

Answer 2

• Field experiment: own natural environment but researcher still manipulates IV
  > ✓ few demand characteristics ; high internal validity also ecologically valid
  > ✕ unreliable ; difficult to control all variables, no informed consent, expensive + time consuming
• data collected by observation

Question 3

Sample

Answer 3

• 4450 travellers
• 45% black, 55% white
• average no. in carriage 43
• average critical area 8.5

Question 4

Independent variables

Answer 4

• race of victim
• type of victim (drunk, ill)
• use of role model (early at 70s or 150s)

Question 5

Dependent variables

Answer 5

• race/gender of helper
• which victim was helped more
• speed/how many helped
• ppts tried to move away
• what was said
• no. ppl helped after model

Question 6

Control variables

Answer 6

• identically dressed in bomber jacket

Question 7

Procedure

Answer 7

• 70s after train left, victim pretended to collapse
• waited for help, no help then ‘model’ helped at next stop
• at trial end, experimenters got off train + boarded train travelling opposite direction
• 6-8 trials everyday
• 1968: 103 trials, 11am-3pm, no stops, journey = 7.5 mins

Question 8

Results 1

Answer 8

• ill (cane) victim helped 95% trials
• drunk helped 50%
• victim + helper race ✕ effect except in drunk condition slight tendency same race helping
• model person hardly used ; lots helped (81 trials) but once 1 person helped so did others

Question 9

Results 2

Answer 9

• average helping time: ill = 5s, drunk = 10s
  > quantitative ~ easy to analyse/compare
• most didn’t help (women) commented on size/physical strength as reason for ✕ helping
  > human exp, insight to why they did certain things
• NO diffusion of responsibility ; more ppl ✕ reduce help received but ppl couldn’t “get away”

Question 10

Results 3

Answer 10

Longer emergency went on w out assisting:
- less impact model had on behaviour
- more likely to leave critical area
- more likely to discuss incident + justify behaviour

Question 11

Conclusions

Answer 11

• Race of victim + helper ✕ effect except drunk
• ill victim helped twice as much as drunk
• when model (hardly used) helped so did others
• ✕ diffusion of responsibility ; no. ppl in carriage had no impact on amount of help given

Question 12

Validity/Reliability

Answer 12

• high ecological validity ; more naturalistic ~ less demand characteristics
• internal validity - some controls eg. Same clothes limits impact of extraneous variables but still present
• low reliability ; difficult to control all variables

Question 13

Sampling bias

Answer 13

• males + females ✓
• large 4000+ sample ~ good for generalisability BUT can’t be sure if results can be applied to another place/setting ; ppl unable to walk away
• more likely to notice emergency due to occupied space
• outside emergency help more difficult to obtain

Question 14

Ethnocentrism

Answer 14

• ✓ mix of races 45% black + 55% white
• ✕ based only on New York

Question 15

Ethics

Answer 15

• ✕ PFH: stress/anxiety
• ✕ informed consent: didn’t know
• ✕ right to withdraw: confined, couldn’t walk away
• ✓ deception: fake model + victim

Question 16

Practical applications (shows:)

Answer 16

• diffusion of responsibility doesn’t necessarily occur
• helping behaviour = v complex w wide range of variables that may affect it
• in irl conditions, average person even in group will act when clearly sees someone in trouble
• cost/reward factors affect helping
  > maybe should be increase discussion ab costs/rewards problem + helping eg. Change laws so ppl effected - sued