Memory: Reliable Or Not?
How reliable is our memory? Memory is an attribute that everyone has, although it comes in varying levels and complexities. The majority of the population would likely say, with confidence, that memory is highly reliable. Although some may agree that they don’t have the best memory personally, the concept of storing and being able to retrieve past information accurately seems well within virtually anyone’s grasp. However, numerous experiments and case studies have challenged just that.
Loftus and Palmer conducted a study in 1974 that strongly supports the understanding that our memory is not as reliable as we’d like to think. They wanted to gain an insight into the means by which memory can be influenced by post-event information. It involved two closely related laboratory experiments where participants were kept in the same controlled conditions. The first experiment involved 45 undergraduate students studying psychology at the University of Washington. They were shown several short videos of 2 cars driving and asked this question: “About how fast were the cars going when they ‘verb’ into each other?” The five verbs used were ‘smashed’, ‘collided’, ‘bumped’, ‘hit’ and ‘contacted’, all of which imply different levels of extremity regarding the contact between the cars. The choice of verb included in the question was the independent variable, while the dependent variable was how fast the participants predicted the cars were going. The second experiment involved 150 University of Washington psychology students who were shown a single video of a car driving smoothly and then getting into a traffic incident. Loftus and Palmer substituted in the question of “How fast were the cars going when they ‘verb’ each other?” The students were divided into 3 groups of 50, and each group was asked the question including the verb ‘hit’, ‘smashed’, or a control group who weren’t questioned about the cars’ speed at all. Each group of students was then asked if they remember seeing any broken glass, and to answer with either a yes or a no. Again, this was the dependent variable. There was no actual broken glass in the video.
From the initial experiment, the researchers found there to be a reasonably large range in speed depending on the verb used. In miles per hour, the average estimated for ‘smashed’ was 40.5, ‘collided’ led to 39.3, ‘bumped’ was 38.1, ‘hit’ was exactly 34 and ‘contacted’ was just 31.8. Loftus’ and Palmer’s explanations for their findings were that the diction used may have altered how each participant remembered the video, or that the participants were subject to bias and guessed based on the speed they thought appropriate for each verb, knowing that’s what the researchers wanted. Regarding experiment two, 16 of the participants who remembered broken glass being present in the video were in the ‘smashed’ group, 7 in the ‘hit’ group and ‘6’ from the control. However, the number of those who did not recall noticing broken glass was 34 in the ‘smashed’ group, 43 in the ‘hit’ one and 44 from the control. Although the independent variable didn’t necessarily affect the entirety of the students involved, there is a significant difference in both outcomes between those in each group. The difference in numbers was most obvious between the ‘smashed’ and ‘hit’ groups, which suggests a strong correlation between the information received directly from the event and that from after the event occurs (differing vocabulary).
In conclusion, what can be taken from Loftus and Palmer’s research is that event information and post-event information can combine to create a distortion of a memory or a false one altogether. Memory is not something where our brains take in all the information from a situation and are then able to recite it back accurately sometime later, but instead a process of scenarios constantly changing in response to our environment, experiences and post-event information. Despite this, a significant weakness of this study is that the researchers only considered the two types of information (during and after the event), however there are several other variables (such as pre-event info) that affect memory just as drastically. Methodologically, this study took place in a lab and therefore cannot necessarily be generalized to situations where confounding variables can’t be controlled. This also means the study lacks ecological validity, however it does have internal validity due to the researchers’ ability to control a vast range of confounding variables. The study’s results are actually reasonably applicable though as people involved in real investigations may be asked to watch videos of events and come to conclusions, just as participants in this case did. The fact that students were the subjects is also an important consideration to make, as they clearly had the incentive to participate and therefore may have displayed demand characteristics based on their understanding of how their professor (Loftus) wanted them to respond.
In 1984 in North Carolina, Jennifer Thompson was sexually assaulted. While being raped, she took it upon herself to take note of every possible detail she could about her perpetrator. That way, in the future, she’d be able to aid the police in bringing justice to the man who did this to her. She eventually managed to escape, and so her rapist left and went on to rape another woman that same night. She later went to the police and worked with them to create the exact picture of what she believed he looked like. Once the picture had been released, one suspect who seemed to match the photo pretty accurately was Ronald Cotton. His photo was then added to a photo lineup which Thompson was asked to examine. Although she was told that there was a chance her rapist wasn’t included in the lineup, she studied the photos for several minutes and then pointed positively at Cotton. When he was brought in for questioning, Cotton’s alibi proved inaccurate, giving the police even more reason to suspect he was guilty. Thompson was then brought in again for a real-life lineup, and once deciding on Cotton again, an officer confirmed that she had chosen the same man as in the previous photo lineup. This is an important detail as it was likely part of what helped to solidify her memory of that event. Needless to say, Cotton was soon sentenced to prison as guilty of all charges, much due to Thompson’s convincing testimony. However, after spending several years in prison, Cotton came across Thompson’s real rapist himself – Bobby Poole. He was given a court case against Poole in the hopes of proving his innocence, but largely due to how adamant Thompson was on Cotton being guilty, his case was unsuccessful. Even more years passed before Cotton learned about using DNA as evidence in trials. He hired another attorney to carry out the testing despite the extreme threat it would pose to his life (should it come out positive), but sure enough his DNA was not found to match that of the real rapist – Poole. Despite the completely undeniable evidence, Thompson still struggled to wrap her head around the fact that her rapist was not at all who she’d thought it was for the past 11 years. She eventually asked for Cotton’s forgiveness, and they are now close friends.
This study has very strong ecological validity because it was a completely real situation as opposed to a controlled experiment. Since then, it has become more and more clear that eyewitness testimony is certainly not something that should be looked upon by the court as a concrete form of evidence. Further research conducted suggests between 20 and 25 percent of witnesses point out someone in a lineup who the police are positive is not guilty. Several elements of the justice process play a role in how Thompson and so many others are able to wrongfully convict innocent people despite being so confident. For starters, photo lineups are often deceiving because although people are told their perpetrator may not be included, they’re usually eager to find the person to blame. Additionally, the fact that the pictures in Thompson’s lineup were all shown together meant she was likely subconsciously searching for the person who most closely represented the man in her memory, even though he didn’t fit the image perfectly. The validation Thompson received once told she chose the same man in both the photo and real-life lineups only solidified her false memory further, which is another clear weakness in the justice system that could easily be improved. Many suggest that people completely uninvolved in the case should be the ones who conduct lineups, therefore eliminating the purposeful or unconscious bias that officers may bring.
Despite there being substantial evidence that our memory is not at all as strong as we like to assume, there are a few significant situations where that actually hasn’t been the case. The first case where eyewitness testimony was conducted in a real-life situation was by Yuille and Cutshall in 1986. The main aims of their research were to record and evaluate eyewitness accounts, to look into conflicts related to research in laboratories and most importantly, to examine how accurate witness accounts can be or what mistakes are made within them. The incident occurred in Vancouver, Canada. A thief walked into a gun shop, tied up the owner and proceeded to steal the store’s guns and money. However, the owner managed to untie himself and quickly rushed outside in an attempt to catch the registration number on the culprit’s car. As he hadn’t yet driven away, the thief shot at the owner twice. The owner then went on to shoot at him six times, ultimately killing him. Luckily, the owner recovered from his severe injuries. There were also numerous witnesses who saw the incident from passing cars, side streets, buildings, etc.
This was chosen as a case study for various reasons including the fact that there were multiple eyewitnesses, as well as sufficient evidence left at the crime scene that lined up well with witnesses’ testimonies. Overall 13 out of the 21 witnesses agreed to be involved in the research. Once ensuring they understood the police’s report, Yuille and Cutshall then asked participants to describe what happened. Following that they were recorded answering a series of questions from an officer. After approximately 4 to 5 months, the same 13 witnesses were given the same questions as they were initially, using the same method too. The participants were also asked two misleading questions: one about a busted headlight (which didn’t exist) and the other about a yellow quarter panel (despite it actually being blue). Additionally, they were asked to rate their level of stress during the incident on a scale of 1-7, as well as elaborate on their emotion levels before the event and any possible issues experienced after it (like sleeplessness). Yuille and Cutshall used a specific procedure for comparing the police interviews with those of the witnesses to uncover an accurate account of what happened. The specifics provided by both officers and witnesses were split into either ‘action’ or ‘description’ details, and within description details were those based around either objects or people.
Overall, Yuille and Cutshall were able to determine a more detailed account of the incident than the police were. This was largely related to them asking questions that were of little to no relevance to the police. But despite this, the police were found to have received more action and person details. Out of all 13 witnesses, 10 remained unaffected by the purposely misleading information. This led the researchers to believe that the accuracy of witnesses’ memories wasn’t as poor as previous studies in labs made it seem and may have even been the perfect example of flashbulb memory. The event’s accuracy likely depended a lot on the memorability or the situation. The fact that all witnesses included in the investigation experienced it from a third-person perspective led to them being able to provide lots of objective details, even months later, as the memory may not have been traumatic or stressful enough to be suppressed.
Regarding methodology, the results from this case study cannot be generalized to all cases involving eyewitnesses, as the research was only tested on a sample group of 13 individuals. Yuille and Cutshall’s predictions about the role of flashbulb memory in this situation also can’t be accurately generalized to other laboratory studies on memory, as those in labs tend to target different memory processes. The study’s findings are reliable in the sense that the researchers were cautious in ensuring witnesses’ memories were not altered by their testimonies. As well, Yuille and Cutshall converted the details they received into numerical data, therefore opening it up to methodological biases regarding their personal interpretation of the procedure. On the other hand, this research does have high ecological validity as it’s based around a completely real event. This field study is very ethically sound, as all witnesses involved had given consent and their identities remained confidential. They also experienced no undue stress or harm from the researchers, and although minor deception was involved (with the misleading questions), it was insignificant in affecting the emotional state of the participants. As far as their right to withdraw, 5 witnesses did chose to not be included in the study, and the store owner (the only living witness to experience the incident firsthand) was not obliged to participate either. A slight weakness is that the 13 witnesses were not explicitly given a debrief regarding the misleading questions included in the questions. But overall, both Yuille and Cutshall were completely able researchers and the officers conducting the questions were also fully qualified.
In conclusion, the majority of psychological research points to memory not being nearly as reliable as we’d like to think it is. Although Yuille and Cutshall’s field study suggested that some aspects of memory can be somewhat reliable, most other research conducted (both in labs and through case studies) proves quite the opposite. Our memories are subject to change from all sorts of variables, from post-event information to biases in the justice system and subconsciously developing false memories. Ultimately, going forward we need to be aware of just how susceptible human memory is to change, even though we might not notice it ourselves.
- ‘Eyewitness Testimony Part 1.’ Youtube, uploaded by CBS News, 9 Mar. 2009, www.youtube.com/watch?v=u-SBTRLoPuo&t=55s. Accessed 10 Sept. 2019.
- ‘Loftus and Palmer (1974) – Eyewitness Testimony.’ Psych Yogi, 21 Mar. 2016, psychyogi.org/loftus-and-palmer-1974-eyewitness-testimony/. Accessed 10 Sept. 2019.
- Nooknotes12. ‘The Way We Remember.’ Total Perspective, WordPress, 21 Mar. 2013, totalperspectiveblog.wordpress.com/tag/the-psychology-of-memory/. Accessed 11 Sept. 2019.
- ‘Yuille and Cutshall 1986.’ A2 Psychology, a2edexcelpsychology.weebly.com/yuille-and-cutshall-1986.html. Accessed 10 Sept. 2019.