Loftus and Palmer 1974: Memory & Eyewitness Testimony
Retrieved April 2017
How can language distort the effects of memory in eye witness testimony?
One vitally important issue in psychology is memory.
- How do we remember?
- How can we improve our memory?
- What affects our ability to recall information?
- How reliable is eyewitness testimony given during court cases?
Some believe that information after the event could affect eyewitness testimony, and that, unless certain things are taken into account, eyewitness testimony has little reliability.
Subsequent research by Loftus and Palmer (1974) in the paper Reconstruction of Automobile Destruction showed that the language used when questioning witnesses to an event could actually influence their memories of that event.
So, the researchers believed that if a certain wording was used in a question, respondents would provide different accounts of an event. So, were they right about this, and how did they come to this conclusion?
What were the aims and the hypotheses of the Loftus and Palmer study?
Loftus and Palmer believed that leading questions could affect recall in those asked to provide eyewitness testimony, and their particular aim was to test whether leading questions would affect recall of the speed of a car and cause people to misremember other details (particularly the presence of broken glass) during a traffic accident.
How did Loftus and Palmer design their study?
Loftus and Palmer tested their hypothesis by setting up two lab experiments. The first experiment involved asking an opportunity sample of 45 students, each allocated to one of five groups. Each group was asked a particular question utilizing a verb (smashed, collided, bumped, hit, contacted) after having watched a video of a car accident.
In their question, they were asked what speed cars were travelling at when they collided during an accident. For example, one group was asked ‘How fast were the cars travelling when they smashed?’.
The second experiment conducted was relatively similar to the first. 150 students, split into three groups of fifty, were each shown a clip of a multiple car accident. The first group of 50 were asked the question ‘how fast were the cars travelling when they hit?’ the second the same question but with the verb ‘smashed’ and the third were the control group, and were not asked a question.
A few days later, without watching the video again, they were asked ten questions, with one placed randomly on the list: ‘Did you see any broken glass? Yes or No?’. There was no broken glass on the original.
What were their findings?
In the first experiment, if questions were phrased using more emotive words like ‘smashed’, people overestimated the speed that the cars were travelling at during an accident.
For example, if people were asked ‘how fast were the cars travelling when they smashed’ they estimated the cars were travelling approximately 41mph, compared a lower estimate of 32mph with questions using the word ‘contacted’. In the second experiment, respondents who were asked the question with the verb ‘smashed’ were more likely to report seeing broken glass.
Loftus and Palmer have two explanations for this. Firstly, what they called the ‘response bias factor’. In other words, the way the question was phrased influenced the person’s answer, making them overestimate the speed of the car as a result of the verb used. However, their memories of the event were not affected.
The second explanation is that a person’s memory and perception of the event would actually change as a result of the question, and this false memory would be stored in their memory. This seems to have been confirmed by the second experiment, as the participants ‘remembered’ seeing broken glass, thus illustrating that leading questions can change the way an eyewitness remembers an event.
What do the critics say?
Because the experiment was a lab experiment, it may not have had ecological validity, meaning that it may not have been representative of the way memories are formed in a natural environment.
Likewise, another way in which the study lacks ecological validity is because the respondents merely watched a video of an accident, and this is very different from being an eyewitness to an accident in real life.
In fact, a study by Yuille and Cutshall (1986) showed that misleading information and leading questions did not change the perception of those who had witnessed a real life bank robbery. Thus these language changes may only have an impact in the lab.
Another problem with the study is the sample used. The participants in the research were all students, and students are not representative of the general population, which may make the data questionable and affects its validity.
However, on the plus side, the study was conducted in a controlled environment and so it as able to show a cause and effect relationship between the independent variable (the phrasing of the questions) and the dependent variables (the estimation of speed and the memory of broken glass). Another strength of the study is its replicability; is it easy to set up another experiment like that of Loftus and Palmer in order to test their findings.
What implications does this study have?
The data garnered by this study may seem relatively banal and inconsequential, but the findings of Loftus and Palmer’s study could actually have profound consequences for the judiciary, the police and the criminal justice system.
An eyewitnesses reporting of an event, and in fact their memory of this event, could actually be changed by the way in which an interviewer phrase the questions, which could have a massive bearing on any criminal case.
As a result Loftus and Palmer advise against the use of leading questions during investigations. In fact, Elizabeth Loftus has appeared as an expert witness in countless trials, and her research and the research of others has been used to develop the Cognitive Interview, a way to question eyewitnesses that allows them to recall information more accurately.
This study also has implications for the way we communicate with others; if we want to get a truthful answer, we need to be wary of how we phrase a question.