Loftus and Palmer (1974)


Eyewitness testimony can be crucial in criminal trials, indeed guilt is often dependent on the evidence of bystanders.This reliance on eyewitness testimony is, however, predicated on the assumption that our memories are trustworthy. Loftus & Palmer tested the reliability of eye witness testimony. They investigated whether a memory of an event can be changed by the way witnesses are questioned (the effect of leading questions).

Memory is not like a camera – we always reconstruct our memories. Previous research has shown that memories for everyday objects are poorly remembered and illustrate how our existing schemas influence our everyday recollection of events. A mental schema is a pocket of information about some aspect of the world, stored in our long term memory, that helps us process information.

slide1 (1)

Adjusting memories for events and objects to fit with our expectations, beliefs and or stereotypes (basically our schemata) is known asCONFABULATION. Cognitive Psychologists suggest that memory recall is dependent on a person’s interpretation of the world.  In other words, we see a situation according to what we expect and assume is normal.

Leading questions introduce new information, which may therefore activate wrong schemas in witness’ mind. Consequently, witnesses may recall events incorrectly.


To investigate the affect of leading questions on eye witness memories when reporting numeric details (e.g. time, speeds etc).

Experiment One

Method & Design

Independent variable: Critical verb used (smashed, collided, bumped, hit & contacted) when questioning the participants about the accident.

Dependent variable: Speed estimates (mph)

Laboratory experiment with an independent measures design


An opportunity sample of 45 students from the University of Washington. These participants were all the same age and were divided into one of 5 conditions (9 participants per condition).


Participant’s were shown 7 film clips of car accidents, raining from 4 to 30 seconds long  in random order (originally part of a driver safety film). After watching each clip, they were asked to:

  1. Write an account of what they had just seen.
  2. Answer some questions.  One was a critical question – ‘How fast were the cars going when they **** each other?’

There were 5 conditions in the experiment, and participant’s were assigned to one condition only. In each condition participants were asked a different critical question where the word ‘hit’ was given as either: smashed, collided, bumped, hit or contacted.


Verb Mean speed estimate
Smashed 40.8
Collided 39.3
Bumped 38.1
Hit 34.0
Contacted 31.8

The phrasing of the question had a considerable effect on the estimate of speed.


Loftus and Palmer developed two possible explanations to explain the findings.

  1. Response bias factors: participants were uncertain of the correct answer and therefore the wording of the question influenced the answer that they gave (demand characteristics).
  2. The form of the question actually caused a change in the participants memory representation of the accident.

In the first experiment, there is a degree of uncertainty over whether prompts given by the researcher had led the participants to alter their memories, or whether it had simply led them to give certain answers. In the second experiment, the researchers aimed to show that information provided after an event is capable of distorting memories.

Experiment Two

Method & Design

Independent variable: Critical verb used (smashed or hit) when questioning the participants about the accident and broken glass.

Dependent variable: Answer given to the critical question ‘Did you see any broken glass?’.

Laboratory experiment with an independent measures design


An opportunity sample of 150 students from the University of Washington. These participants were all the same age and were divided into one of 3 conditions (50 participants per condition).


Participant’s were shown a one minute film which contained a 4 second  car accident.

Participants were divided into 3 experimental groups:

Group 1 were asked: How fast were the cars going when they hit each other?

Group 2 were asked: How fast were the cars going when they smashedeach other?

Group 3 were asked: no question about the speed of the vehicles.

A week later the participants were asked to return to the psychology lab and were asked some further questions including the critical question ‘Did you see any broken glass?’. This question appeared in a randomly within the other questions.

There was in fact no broken glass in the film.


Response Smashed Hit Control
Yes 16 7 6
No 34 43 44


Once again the wording of the question had a considerable effect on the estimate of speed and on the number of ‘Yes’ responses when asked about broken glass although the differences between the control and the ‘hit’ condition was not significant.

The idea that the cars had “smashed” into each other had led participants to incorporate the notion of broken glass into their memories (as “smashed” implies that glass was broken).

Participants took in information from the original scene and then merged this with information given after the event. This caused them to produce a memory which contained some of the original information but then would be swayed by the next piece of information that they were exposed to.

Controls in the study: The age of the participants, the use of the video and the location of the experiment, all the participants were asked the same questions (apart from changes in the critical verb), the position of the key question in the second study was randomized.



+ A Lab experiment was used which had high control and thus allowed for the accurate measurement of variables, thus more objectivity. Control over the IV allows for cause and effect and also grate ability to replicate the study.

+ Independent measures design prevents order effects (such as practice and fatigue) as participants only participate in one condition.

+ A large sample size was used thus individual differences were controlled for. The sample is therefore representative of the general population and findings can be generalised.

+ Quantitative data was obtained, giving objective, statistical information which can be easily analysed and used to draw comparisons between participants.

+ The study is very useful and has practical applications that can be applied to the real world, for instance to avoid using leaden questions within police questionning of witnesses and suspects.


– Artificial lab conditions may produce unnatural behaviour (e.g. demand characteristics) that lacks ecological validity. Also, total control over variables is not possible.

– As participants only participated in one condition, there may have been individual differences affecting the dependent variable and thus reducing the validity of the study.

– The sample included students from same area (Washington) which is ethnocentric. Participants were all of the same age. It may be the case that they were less experienced drivers and thus less experienced to estimate speeds. Also, students are known for having good memories for situations like this, thus not representative of normal memory.

– Participants were aware that they were participating in an experiment. They are therefore more likely to perform to demand characteristics and act unnaturally to either work out the aims of the study or to please the experimenter.

 Lacked ecological validity as it was carried out in a laboratory and the real life experience of witnessing an accident would be far more arousing than simply watching the events of a television. Furthermore, the participants had been cued to watch the video, whilst crashes in real life a largely unexpected.

– Quantitative data reduces behaviour to numbers. It doesn’t tell us why behaviour occurs and thus lacks in-depth insight and fails to consider the thoughts, feelings and experiences of participants.

– Participant’s memory may not have been altered at all. They may simply have been using the information provided by the experimenter (smashed, hit etc.) to take a guess as to the speed of the vehicle. Therefore the study may lack validity because the researchers did not measure was they set out to measure.

Click here for model answers and exam practice

Useful resources:

Section A questions

Section B questions

Multiple Choice Quiz

One thought on “Loftus and Palmer (1974)

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>