Classic Study: Loftus & Palmer (1974)

Classic Study: Loftus & Palmer (1974)

1. Generalisability

   • A large sample was used, especially in the follow up study, which makes results more generalisable.
     • Experiment 1: 45x participants were used.
     • Experiment 2: 152x participants were used.
   • As the participants used were students, the sample was not representative of the general population and therefore results may lack generalisability.
     • Participants used were all students of the researcher.

2. Reliability

   • Both experiment 1 and experiment 2 were well controlled and followed a standardised procedure. This makes the studies more replicable and therefore reliable. They both also found that leading questions did in fact influence recall, further suggesting results are reliable.
     • Experiment 1: All participants watched the same car crash and answered the same questionnaire, except question 9 was different: "How fast was the car going when they (smashed/hit/bumped/collided/contacted) one another?"
     • Experiment 2: Each group watched the same car accident and after 1 week was asked "Was there any glass?". The only difference was one group was asked "How fast was the car going when they hit?", one was asked "How fast...when they smashed?" and one group was asked nothing (each group had 50 participants in).

3. Application

   • Improvements could be made to the legal system based on these results in order to make eye witness testimony more reliable and valid.
     • The aim was to see whether leading questions would affect the recall of participants.
     • Found that leading questions did affect the recall of events in both the first experiment and the second experiment.

4. Validity

   • Participants used were her own students.
     • Results may be biased as they may have known the aim and wanted her to get the results she was looking for.
     • Students are less experienced drivers so the change in verb may have been more likely to influence their perception of the speed the cars were travelling at/they may have felt less confident in giving their estimates.
   • The results may have been influenced by the fact there was low mundane realism as participants were only watching videos, not experiencing a real crash. This means the results may lack ecological validity.