Loftus & Palmer (1974)

what phenomenon was investigated?

Reconstructive memory — specifically whether leading questions and post-event information can distort eyewitnesses' memories of an event

leading question

A question that by its form or content suggests to a witness which answer is desired

reconstruive memory described by Bartlett (1932)

The idea that memory is not an exact copy of events but is reconstructed using existing schemas — unfamiliar experiences are fitted into existing knowledge frameworks, which can distort recall

aim of study 1

To investigate whether the use of different verbs in a leading question would affect participants' estimation of the speed of cars in a traffic accident

participants study 1

45 students divided into five groups of nine. Independent samples design — each participant watched all 7 films but only answered one version of the critical question

procedure study 1

Participants watched 7 films of traffic accidents (5–30 seconds each) then answered a questionnaire. The critical question asked them to estimate car speed using one of five verbs: smashed, collided, bumped, hit, or contacted

speed estimatuon results study 1

Smashed: 40.8 mph, Collided: 39.3 mph, Bumped: 38.1 mph, Hit: 34.0 mph, Contacted: 31.8 mph. Results were significant at p ≤ 0.005

2 explanations for study 1 results

1. Response bias — participants were uncertain about speed so the verb biased their estimate. 2. Schema processing — the verb changed participants' mental representation of the accident, with "smashed" activating a schema of a more severe accident

aim study 2

To investigate whether a leading question about speed in part one would cause participants to falsely report seeing broken glass (which was not present) one week later

participants study 2

150 students randomly allocated to three conditions — smashed (50), hit (50), and control/no speed question (50)

procedure study 2

Participants watched a 1-minute film of a car accident, then answered a questionnaire including a speed estimation question (smashed/hit/none). One week later they returned and answered 10 questions including the critical question: "Did you see any broken glass?" — there was no broken glass in the film

speed estimation results study 2

Smashed condition: mean estimate of 10.46 mph. Hit condition: mean estimate of 8.00 mph

broken glass results study 2

In the smashed condition 16 said yes vs. 7 in the hit condition vs. 6 in the control. A chi-square test was significant at p ≤ 0.025 — showing smashed led to significantly more false reports of broken glass

What does Study 2 demonstrate about post-event information?

That information introduced after an event (through the wording of a question) can be integrated into the original memory, making it difficult to distinguish the original experience from the post-event suggestion

How do the results of Study 2 support Bartlett's reconstructive memory theory?

Participants used their schema of a serious car accident (activated by "smashed") to reconstruct their memory — leading them to falsely "remember" broken glass that was consistent with that schema

experimental method

High internal validity — confounding variables were controlled, allowing a cause-and-effect rel

limitation

Low ecological validity — watching short films of accidents in a laboratory does not replicate the stress and emotion of witnessing a real accident, so findings may not generalise to real eyewitness situations

student sample

Students are not representative of the general population — they were likely young and inexperienced drivers, which may have affected their ability to estimate car speeds accurately

real-world implications

The research highlights serious concerns about eyewitness testimony in courtrooms — leading questions from lawyers or police could distort witnesses' memories and lead to false evidence being given