Classic Study: Loftus & Palmer 1974

Classic Study: Loftus & Palmer 1974

1. Exp 1

   • Aim: To explore how info provided after an event, in form of leading questions, might affect people's memories
   • Participants: 45 students from university of washington
   • Procedure: P's shown 7 videos of car crashes. Following each of videos, p's given standardised questionnaire first asking them to give account of what they had seen, then to answer number of q's (most were filler q's). Critical question: "How fast were the cars going when they _________ into each other?". Conditions: smashed, collided, bumped, hit, contacted.
   • Results: speeds estimated incorrectly; mean estimates for 'lowest verb' (contacted) and 'highest verb' (smashes) differed by 9mph
   • Conclusions: Distortion of memory occurred because it has been reconstructed, the verbal label attached to event by question could have led to p cognitively amend their memory AND/OR p's may not have been sure of speed to start with, and therefore may adjust their answers to fit with the expectations of researcher (response-bias)

2. Exp 2

   • Aims: to show more conclusively that info provided after an event is capable of distorting memories
   • Participants: 150 students from university of washington
   • Procedure: all p's shown short film of multiple car crash lasting 1 minute. They were split into conditions, given questionnaire: smashed, hit, control group (not asked about speed). Ps were recalled 1 week later and given another series of q's. Critical question: "Did you see any broken glass?" (there was no glass in film)
   • Results: mean speed estimate in 'smashed' condition was 10.5mph, which is sig higher than 'hit' condition (8mph). P's in smashed condition were more likely to think they saw broken glass (16) than p's in hit condition (7) and control (6). Majority of ps said no, they did not see broken glass
   • Conclusions: labels attached to car crash by q's affected memory of ps altering their perception of event a week later. The idea that the cars had 'smashed' into each other had led ps to incorporate idea of broken glass into their memories ('smashed' implies that glass was broken)

3. Overall Conclusions: Reconstructive memory is supported, EW's to car crashes are not very good at estimating the speeds of the vehicles involved, leading q's can influence memory of EW's, EWT may be unreliable

4. Issues & Debates

   • Ethics: minor deception - minimal concerns
   • practical issues: real EWs cannot be studied, difficult to obtain representative samples
   • Reductionist: Quantitative data (speed estimates) - lacks meaning, we dont know WHY
   • comparing explanations: witness/system variables - leading qs = SYSTEM
   • Science: high reliability, low ecological validity
   • Knowledge over time: based on Bartlett's reconstructive memory, triggered huge amounts of further research into EWT
   • Use in society: EWT may be unreliable, improve criminal justice system
   • Socially Sensitive: Positive - changes in criminal justice system to make it more ethical, Devlin Report 1976 recommend trial judges be required to instruct juries that it is not safe to convict on a single EWT alone