Memory is an important cognitive process that guides our behaviours; it is often relied heavily upon to solve small matters in everyday life and huge issues in legal systems. But is this process actually reliable? A study by Bartlett suggests that this may not be so – reconstructive memory, the theory that memory is not exact or precise but must be pieced together by our experiences, can be distorted by the culture we are brought up in. Additionally, Loftus and Palmer have concluded with findings from their experiment that post-event information can easily distort memory of the actual event.
However, Loftus’ study was criticized by Yuille and Cutshall for being too lacking in ecological validity, and has in fact obtained findings that contrasted it. A factor that influences the reliability of our memory are schemas – “packets” of information obtained from previous experiences that subsequently affect our perceptions to specific things. Bartlett was interested in the role of these schemas in recall. He asked European and Native American participants to read the Native American folk story, War of the Ghost, twice. Fifteen minutes after reading, both groups were asked to reproduce the story.
This story was chosen because of its cultural significance towards Native Americans – European participants will adapt their own cultural schemas to their memory of the story. As a result, Native Americans found it easier to reproduce the story, and European participants did indeed use their own cultural schemas to fill in gaps of the story which they could not remember accurately, such as filling in “boat” for “canoe”. This shows that memory is indeed susceptible to distortion caused by the culture we are brought up in, hence not necessarily as reliable as we think.
The way this study was performed was very straightforward and highly replicable, meaning this experiment can be repeated to see if results are the same for different cultures and different stories. Gender bias was eliminated as participants were selected at random. However, as participants did not receive specific instructions and were only asked to reproduce the story, nor were they asked to be as accurate as possible, there could be distortions as they could have been guessing the aim of the experiment, hence causing demand characteristics.
The methodology in general was not sophisticated – there were no independent, dependent and control variables, meaning there were no control groups, hence providing one-sided results. This all adds up to low internal validity. Alternatively, the form in which Europeans remember better may be different to the form in which Native Americans remember better – for example, Europeans may remember images better than words – this is demonstrated in another study by Cole and Scribner, where two groups outperformed each other in two different conditions.
Another factor that influences the reliability of memory is the use of leading questions, as shown in a study by Loftus and Palmer. The aim of the study was to see if varying leading questions, and adding post-event information, would affect the memory of an incident. 150 students were shown a video of a car accident. Afterwards, they were asked to describe what had happened in three conditions, firstly “how fast were the cars going when they smashed into each other”, secondly replacing “smashed into” with “hit”, and thirdly not a question about car speed (control group).
Smashed” averaged an estimate of 41km/h, whereas “hit” averaged an estimate of 34km/h. One week later, participants were asked whether they had seen broken glass; 32% said smashed glasses was involved in the “smashed” group, but only 14% of the “hit” group and 12% of the control group – in actual fact, however, there was actually no broken glass at all. This suggests how the word implicitly affected the way the situation was remembered by triggering the activation of certain schemas, for example using the word “smashed” implies somewhat of a severe accident, hence associating it with broken glass.
Loftus and Palmer concluded that our memory is fragile and susceptible to distortion with post-event information, hence again suggesting that our memory is not necessarily reliable. The use of a control group (who were not asked about car speed) as well as clear independent (the verbs) and dependent (speed of the car) variables allows this experiment to be high in internal validity and because of the simple procedure, it can be easily replicable and therefore be repeated for different cultures and ages.
Gender bias was also eliminated here, like from Bartlett’s study, as students of both genders were used. Additionally, this study cannot, arguably, be generalised – students from one school cannot be represent the entire world population, similar to Bartlett’s study, where only two cultures were involved in the experiment. This experiment also lacks ecological validity as they were essentially not real-life witnesses of the event, meaning their recall would not be affected by emotions felt if participants witnessed first-hand.
Also, in real-life situations, participants would often be taken by surprise and would not be paying full attention to it, whereas in this highly artificial environment they know they have to be focussing on the screen, meaning this study lacks mundane realism. Yuille and Cutshall have criticized Loftus and Palmers’ experiment for being far too lacking in ecological validity – they argue that an experiment within a laboratory does not reflect what would happen in real life.
Yuille and Cutshall carried out a very similar experiment but in a real-life setting – they interviewed people who had witnessed a real robbery and discovered that leading questions did not in fact distort people’s memory – the memory of these details were instead quite amazing. The degree to which details was remembered seemed to increase with witnesses that were closer to the event. Yuille and Cutshall concluded that leading questions had no effect on recall, and the closer to the event a person was, the more accurately they would remember.
This suggests that memory is still quite reliable. Being a field experiment, this study would have higher ecological validity and mundane realism as was one of the aims of the experiment – it was an interview conducted in a natural environment, unlike the previous studies such as Bartlett who made participants reproduce a seemingly random story, or Loftus who made participants estimate the speed of a car in a video.
Internal validity, however, was not maintained at a high level like in Loftus’ study – there was no control over what had happened before the interview, and it is possible that distortions to their memory of the event had already happened. Participants were also not selected by the researchers, and instead were voluntary; meaning one gender of participants may outnumber the other gender.
There is also the issue of generalising this study to a wider population, as the witnesses were very limited and the sample was only a small fraction of that number, meaning such a small number of people cannot possibly represent an entire populace. However, all in all, this experiment does indeed reflect with high ecological validity and mundane realism that leading questions do not affect memory. In conclusion, Bartlett’s study suggests how memory is reconstructive and is therefore influenced by cultural schemas, hence rendering memory an unreliable cognitive process.
Loftus and Palmer support this argument of unreliability and had obtained findings that suggest leading questions and post-event information do indeed distort memory. However, this is contrasted by Yuille and Cutshall, who have proven in a real-life situation that memory is not affected by leading questions – however, they did not ask for further details after one week like Loftus had, therefore allowing Loftus’ conclusion of memory being easily distortable to still be true.