Eye Witness Testimony

• Gabbert et al - Study in post event discussion.
• Gabbert et al:

  • 60 students and 60 adults from Aberdeen
  • Split into pairs, one watched a video of a girl stealing a wallet and the other didn’t
  • They were encouraged to discuss what they’d seen
  • 71% in the experimental group recalled information they had not actually seen
  • 60% said the girl was guilty
• Loftus and Palmer - Car crash study.
• Yullie and Cutshall - Stress improves memory.

  Volenteers that rated themselves higher on the stress scale were 85% accurate while those who rated themselves lower were 75%
• When did Loftus and Palmer‘s study take place?

  1974
• Loftus and Palmer (1974)

  • Aim - to investigate how information supplied after an event influences a witnesses memory.
  • Method - laboratory experiment using independent measures.
  • Independent variable - the verb used.
  • Dependant variable - the participants speed estimate.
• Loftus and Palmer (1974)

  • Method - 45 students (9 in each group) of the University of Washington were chosen through opportunity sampling.
  • They were each shown seven film-clips of traffic accidents.
  • Each student was asked to writer an account on the video and estimate the speed.
• Loftus and Palmer (1974)

  Condition 1: Smashed into each other

  Condition 2: Collided into each other

  Condition 3: Bumped into each other

  Condition 4: Hit each other

  Condition 5: Contacted each other

  The adjectives went down in severity for each group.
• Loftus and Palmer (1974)

  The average estimation of speed decreased with the severity of the adjective. Between “smashed“ and “contacted” was approximately a 10 mph difference.
• Loftus and palmer (1974)

  Conclusions

  • They argue the difference in speed estimates could be due to distortion in the memory.
  • They also acknowledge the possible presence of a response bias factor , in which the participant is not sure of the correct speed and adjusts their estimate to fit in with the questioners thinking.
  • They argue two kinds of information go into a persons memory after an event: the information obtained from perceiving the event, and information supplied to us after the event.
• Loftus and Palmer (1974)

  • Ecological validity - Lowered because students were shown videos instead of experiencing events in real life. However being questioned would be similar to an investigation in real life.
  • Lab experiment - Fine control of variables, reliable and internal validity increased. Loftus actually wants demand characteristics.
  • Practical applications - Gives insight to EWT in court and suggests leading questions should not be used. EWT should be less relied upon.
• Loftus and Palmer (1974)

  • Generalisability - Sample is not generalisable because it is ethnocentric and was gathered through opportunity sampling.
• Loftus et al - Weapon Focus
• When did Loftus et al complete the weapon focus study?

  1987
• Loftus et al - Weapon Focus

  • Procedure - 36 students from the University of Washington aged 18-31 were recruited through advertisements. They were offered money.
  • They were shown 18 slides of a series of events in a taco restaurant. In the control group the cashier is handed a check and in the experimental group the cashier was threatened with a gun.
  • They were then asked to answer a questionnaire which asked them to identify the person with either the cheque or gun.
• Loftus et al - Weapon Focus

  • Findings - In the control condition 38.9% made correct identification, while in the experimental condition only 11.1% made the correct identification.
  • Eye fixation data showed the participants focussed longer on the gun than the cheque.
  • Tunnel effect - anxiety and weapon focus narrow attention.
• Loftus et al - Weapon Focus

  • Lacks ecological validity - uses slides rather than real life events.
  • Reliable - lab experiment.
  • Practical applications - it allows courts to take into account the fact EWT may not be useful and valid in every scenario.
  • Generalisability - Loftus used an ethnocentric sample from University of Washington without a representative age range.