Research Paper On False Memory
False Memories
When engaging in the cognitive processes of recall and recognition, we rarely focus on how accurate a memory is when retrieved. Instead, we focus on whether we are able to retrieve that memory or not. What many of us do not realize is that it is quite common for us to encode memories differently than the way they occurred. There are also instances where we remember events that never happened, and this is quite critical since many, if not all, of our cognitive processes depend on memory (Roediger & McDermott, 1995). False memories have attracted the attention of many psychologists over the years, which have conducted experiments to understand this phenomenon. One experiment in particular, known as the Deese/Roediger-McDermott …show more content…
(DRM) Paradigm, used methods against bias participants into remembering things that did not occur. This Paradigm was originally developed by Deese in 1959, and was later extended by Roediger and McDermott in 1995 (Coane, McBride, Termonen, & Cutting, 2016).
They had participants study a list of semantically related words and then later gave them a recognition task. These participants reported remembering words that were not included in the list, demonstrating false memory (Bui, Friedman, McDonough, & Castel 2013). The purpose of this experiment was to understand the limits of memory and ultimately determine the nature of the mental depictions that are most likely to provoke false memory (Roediger & McDermott, 1995). Through their experiments, Deese, Roediger and McDermott wanted to prove that confidence has nothing to do with memory accuracy; no matter how vivid a memory is there is no way deeming it correct without any objective evidence. They hypothesized that if participants were to be put under controlled conditions, they could be biased into remembering things that never happened—exhibiting false memory. I anticipated the same; if participants were presented with a word that is related to the words in the original list but not actually from that list, they should falsely report having seen the new word in the old
list. One question that arises is how exactly does false recognition occur in this paradigm? One model that attempts to explain this, the activation/monitoring theory, suggests that false memories in this paradigm result from the activation of a presented word list to the recognition items given subsequently through associative networks (Coane et al., 2016). After participants were asked to study a list of words, in many cases they falsely reported remembering a word—the lure—that is semantically associated to the previously studied list, but was never presented in that list. Under this theory, activation spread from the first list to the lure in the second list, resulting in false memory.
A similar model called the spreading activation framework, described by Collins and Loftus (1975), organizes concepts into networks. The activation of words spreads along a pathway in this network, and the more features that two words share, the closer they are to one another. Therefore, semantically related words would be proximate to one another, and thus are more likely to be activated together. This explains why participants so confidently recall seeing the lure; it was activated in the pathway due to its similarity with the other presented words (Coane et al., 2016).
Another possible interpretation that explains how false memories arise is the fuzzy-trace theory. This theory, like the others, predicts that similarity between list and lure words increases false memory. According to the fuzzy-trace theory, both verbatim and gist representations are accessed at the time of retrieval (Bui et al., 2013). While verbatim traces focus on perceptual details, gist traces focus on generalized semantic features. When participants studied a list of words in the DRM paradigm, they relied on gist traces to remember the sequence. This, in turn, produced false memory, because when they were presented with the semantically related but unstudied lure, they recalled studying it. Moreover, the thematic consistency perspective suggests that word lists are organized around a central theme. The lures share semantic features with studied words in this experiment, therefore when subjects are presented with a lure, it is familiar to them thus pushing them to believe that they have heard or seen it in the previously studied list (Zhu, Loftus, Lin, & Dong, 2013). All in all, we regularly encounter more information than we can accurately remember. We often must use relational processing to prioritize which information we would like to encode. As observed in the DRM paradigm, however, this can come at the cost of memory accuracy and potentially result in false memory (Bui et al., 2013).
Methods
In the original DRM paradigm, participants studied a list composed of associated words (e.g. tired, rest, bed, nap, yawn) that were related to a lure (e.g. sleep) not presented on the initial list. Originally, Deese (1959) used a single-trial method to examine the role of word relatedness in free recall (Roediger & McDermott, 1995). He developed a single critical lure and 12 associated words. After studying one trial of word associates, Deese had subjects engage in an immediate free recall test. Roediger and McDermott, on the other hand, extended Deese’s free recall paradigm to include recognition tests (Roediger & McDermott, 1995). Subjects under Roediger and McDermott’s paradigm studied six different lists of 12 associated words. After each list, subjects were given a different sequence and were asked to determine which words were old (recognized from the previous list) or new (not in the previous sequence). The “False Memory” CogLab experiment was fairly similar to the DRM study. Global data on the website was collected from 41,172 participants—20 of those participants from our cognitive psychology class. After reading a background of this study and the necessary instructions to complete it, participants were shown a sequence of words, with each word presented one at a time for about a second. Once the sequence was complete, a set of buttons appeared at the bottom of the screen that included words that were viewed on the list, as well as new “distractor” words that were not. Participants had to indicate which words they remember were on the list; they did this for six trials. Like the DRM paradigm, this experiment was specially designed to bias subjects into recognizing a word that was not presented on the previous list. The six word lists, like the DRM paradigm, were comprised of words that were semantically related in one way or another. For example, one list included the words: Legs, sofa, seat, sit, recliner, bench, desk, stool, and so on. These words all seem to be under the central theme of “furniture”. Once participants were presented with a list of new and old words in which they had to choose from, that central theme served as a guide for indicating the words they remember viewing on the previous list. As participants click on the words they recognize, they come across a word that is under the “furniture” theme, yet was never shown on the original list. In this specific CogLab trial, this word, which was coined by Deese, Roediger, and McDermott as the lure, was chair, which fits the theme. Participants continued this procedure for five additional trials—each trial with a different theme—and then were given their results after completing the experiment. The independent variable in this experiment is the types of words presented on each trial. This includes the original word list, as well as the unrelated distractor words and related lure that were not presented in the original list. The dependent variable, on the other hand, was the percentage of each word type reported by the participants.
Results
After completing the experiment, I found that the results corresponded well with my hypothesis. Participants recognized the related lures at almost the same level as original list items. Looking at my individual data on the CogLab website, I identified about 92% of the original list items when asked to select which words I recognized. For all six trials, 6% of the items I identified were lure, not related to or presented in the original list. 66% of the time, I recognized related lures that were also not presented to me on the original list, which is a pretty significant number. Looking at my trial-by-trial data, in the original list item category, I identified about six to seven words for each trial. Under unrelated lure (not on the list) there were three different trials where I falsely recognized one word that I thought was on the original list but I did not identify any in the other three trials. Under related lure, there were only two trials where I did not choose the lure. The other four trials I so confidently reported remembering them from the original list. Comparing my data to the global data (based on 41,178 participants worldwide), they are somewhat similar. Related lure was often reported, almost as much as the original list item. The unrelated lure, on the other hand, was reported a number of times as well but not enough to be considered significant (6%). The percentage of times an original list item was reported is about 79%. Interestingly, the related lure was reported 73% of the time, which is relatively similar to percent recognized of original list item. In the original DRM paradigm, Deese (1959), Roediger and McDermott’s (1995) results were comparable to my data and the global data obtained. Individuals in their study recognized related lure at percentages close to original list item recognition, thus accurately demonstrating false memory (Roediger & McDermott, 1995).
Discussion
It is safe to say that almost all aspects of cognition depend on memory. It is also safe to say that not all the memory we process is accurate. It is quite common for us to encode memories differently than the way they actually took place and practically speaking; there are many crucial events that rely heavily on the memory of human observers. Eyewitness testimony in murder trials is an example (Bui et al., 2013). Based on the results obtained through the experiment conducted on the CogLab website, as well as those carried out by Deese (1959), Roediger and McDermott (1995), individuals can so confidently recall an event, but the vividness of their memory does not accurately depict how true that memory is. In reality, there is no way of assessing the accuracy of a memory without objective evidence such as actual footage, a voice recording, or photograph of the event.
The experiments explained previously verified that we are susceptible victims of false memory. Something as simple as remembering words of a list can result in false recognition, let alone real world situations where accuracy of detail is crucial. In simplified terms, the results of this data can be stated as follows: individuals can be biased into remembering things that never occurred to them (Roediger &McDermott, 1995). How so? In the actual study, the words that participants falsely identified as an old list item seemed familiar to them because they were related to those old list words. They made associations between these words and concluded that they both were, in fact, studied on the previous list. Their resulting false memories for these words were quite robust, thus emphasizing the reconstructive nature of memory (Bui et al., 2013).
The particular effect used in this experiment was quite robust as well. Even though participants were introduced to the idea of false memory first hand, when asked to read a background on the false memory phenomenon, most still exhibited false memory—myself included. One thing I did notice however was how quickly he original list items appeared on the screen. Each word appeared for one second and moved on to the next word, almost giving the reader no chance to fully process the words on the list. This could be so that the participant encodes only gist representations of the list, which in turn increases false memories when presented with new words that fit that gist. If I were to redo the experiment, I would perhaps show the words for longer intervals of time, or maybe the whole list at once. This would give participants the chance to actually study them, because realistically speaking; the memories we encode on a day-to-day basis are not always rushed. If doing so produces different results, this could reveal a factor that can potentially have an influence on human memory accuracy: time. All in all, through studying these experiments we learn that memory, no matter how obstinate the recaller is, cannot be trusted (Roediger & McDermott, 1995).
When engaging in the cognitive processes of recall and recognition, we rarely focus on how accurate a memory is when retrieved. Instead, we focus on whether we are able to retrieve that memory or not. What many of us do not realize is that it is quite common for us to encode memories differently than the way they occurred. There are also instances where we remember events that never happened, and this is quite critical since many, if not all, of our cognitive processes depend on memory (Roediger & McDermott, 1995). False memories have attracted the attention of many psychologists over the years, which have conducted experiments to understand this phenomenon. One experiment in particular, known as the Deese/Roediger-McDermott …show more content…
(DRM) Paradigm, used methods against bias participants into remembering things that did not occur. This Paradigm was originally developed by Deese in 1959, and was later extended by Roediger and McDermott in 1995 (Coane, McBride, Termonen, & Cutting, 2016).
They had participants study a list of semantically related words and then later gave them a recognition task. These participants reported remembering words that were not included in the list, demonstrating false memory (Bui, Friedman, McDonough, & Castel 2013). The purpose of this experiment was to understand the limits of memory and ultimately determine the nature of the mental depictions that are most likely to provoke false memory (Roediger & McDermott, 1995). Through their experiments, Deese, Roediger and McDermott wanted to prove that confidence has nothing to do with memory accuracy; no matter how vivid a memory is there is no way deeming it correct without any objective evidence. They hypothesized that if participants were to be put under controlled conditions, they could be biased into remembering things that never happened—exhibiting false memory. I anticipated the same; if participants were presented with a word that is related to the words in the original list but not actually from that list, they should falsely report having seen the new word in the old
list. One question that arises is how exactly does false recognition occur in this paradigm? One model that attempts to explain this, the activation/monitoring theory, suggests that false memories in this paradigm result from the activation of a presented word list to the recognition items given subsequently through associative networks (Coane et al., 2016). After participants were asked to study a list of words, in many cases they falsely reported remembering a word—the lure—that is semantically associated to the previously studied list, but was never presented in that list. Under this theory, activation spread from the first list to the lure in the second list, resulting in false memory.
A similar model called the spreading activation framework, described by Collins and Loftus (1975), organizes concepts into networks. The activation of words spreads along a pathway in this network, and the more features that two words share, the closer they are to one another. Therefore, semantically related words would be proximate to one another, and thus are more likely to be activated together. This explains why participants so confidently recall seeing the lure; it was activated in the pathway due to its similarity with the other presented words (Coane et al., 2016).
Another possible interpretation that explains how false memories arise is the fuzzy-trace theory. This theory, like the others, predicts that similarity between list and lure words increases false memory. According to the fuzzy-trace theory, both verbatim and gist representations are accessed at the time of retrieval (Bui et al., 2013). While verbatim traces focus on perceptual details, gist traces focus on generalized semantic features. When participants studied a list of words in the DRM paradigm, they relied on gist traces to remember the sequence. This, in turn, produced false memory, because when they were presented with the semantically related but unstudied lure, they recalled studying it. Moreover, the thematic consistency perspective suggests that word lists are organized around a central theme. The lures share semantic features with studied words in this experiment, therefore when subjects are presented with a lure, it is familiar to them thus pushing them to believe that they have heard or seen it in the previously studied list (Zhu, Loftus, Lin, & Dong, 2013). All in all, we regularly encounter more information than we can accurately remember. We often must use relational processing to prioritize which information we would like to encode. As observed in the DRM paradigm, however, this can come at the cost of memory accuracy and potentially result in false memory (Bui et al., 2013).
Methods
In the original DRM paradigm, participants studied a list composed of associated words (e.g. tired, rest, bed, nap, yawn) that were related to a lure (e.g. sleep) not presented on the initial list. Originally, Deese (1959) used a single-trial method to examine the role of word relatedness in free recall (Roediger & McDermott, 1995). He developed a single critical lure and 12 associated words. After studying one trial of word associates, Deese had subjects engage in an immediate free recall test. Roediger and McDermott, on the other hand, extended Deese’s free recall paradigm to include recognition tests (Roediger & McDermott, 1995). Subjects under Roediger and McDermott’s paradigm studied six different lists of 12 associated words. After each list, subjects were given a different sequence and were asked to determine which words were old (recognized from the previous list) or new (not in the previous sequence). The “False Memory” CogLab experiment was fairly similar to the DRM study. Global data on the website was collected from 41,172 participants—20 of those participants from our cognitive psychology class. After reading a background of this study and the necessary instructions to complete it, participants were shown a sequence of words, with each word presented one at a time for about a second. Once the sequence was complete, a set of buttons appeared at the bottom of the screen that included words that were viewed on the list, as well as new “distractor” words that were not. Participants had to indicate which words they remember were on the list; they did this for six trials. Like the DRM paradigm, this experiment was specially designed to bias subjects into recognizing a word that was not presented on the previous list. The six word lists, like the DRM paradigm, were comprised of words that were semantically related in one way or another. For example, one list included the words: Legs, sofa, seat, sit, recliner, bench, desk, stool, and so on. These words all seem to be under the central theme of “furniture”. Once participants were presented with a list of new and old words in which they had to choose from, that central theme served as a guide for indicating the words they remember viewing on the previous list. As participants click on the words they recognize, they come across a word that is under the “furniture” theme, yet was never shown on the original list. In this specific CogLab trial, this word, which was coined by Deese, Roediger, and McDermott as the lure, was chair, which fits the theme. Participants continued this procedure for five additional trials—each trial with a different theme—and then were given their results after completing the experiment. The independent variable in this experiment is the types of words presented on each trial. This includes the original word list, as well as the unrelated distractor words and related lure that were not presented in the original list. The dependent variable, on the other hand, was the percentage of each word type reported by the participants.
Results
After completing the experiment, I found that the results corresponded well with my hypothesis. Participants recognized the related lures at almost the same level as original list items. Looking at my individual data on the CogLab website, I identified about 92% of the original list items when asked to select which words I recognized. For all six trials, 6% of the items I identified were lure, not related to or presented in the original list. 66% of the time, I recognized related lures that were also not presented to me on the original list, which is a pretty significant number. Looking at my trial-by-trial data, in the original list item category, I identified about six to seven words for each trial. Under unrelated lure (not on the list) there were three different trials where I falsely recognized one word that I thought was on the original list but I did not identify any in the other three trials. Under related lure, there were only two trials where I did not choose the lure. The other four trials I so confidently reported remembering them from the original list. Comparing my data to the global data (based on 41,178 participants worldwide), they are somewhat similar. Related lure was often reported, almost as much as the original list item. The unrelated lure, on the other hand, was reported a number of times as well but not enough to be considered significant (6%). The percentage of times an original list item was reported is about 79%. Interestingly, the related lure was reported 73% of the time, which is relatively similar to percent recognized of original list item. In the original DRM paradigm, Deese (1959), Roediger and McDermott’s (1995) results were comparable to my data and the global data obtained. Individuals in their study recognized related lure at percentages close to original list item recognition, thus accurately demonstrating false memory (Roediger & McDermott, 1995).
Discussion
It is safe to say that almost all aspects of cognition depend on memory. It is also safe to say that not all the memory we process is accurate. It is quite common for us to encode memories differently than the way they actually took place and practically speaking; there are many crucial events that rely heavily on the memory of human observers. Eyewitness testimony in murder trials is an example (Bui et al., 2013). Based on the results obtained through the experiment conducted on the CogLab website, as well as those carried out by Deese (1959), Roediger and McDermott (1995), individuals can so confidently recall an event, but the vividness of their memory does not accurately depict how true that memory is. In reality, there is no way of assessing the accuracy of a memory without objective evidence such as actual footage, a voice recording, or photograph of the event.
The experiments explained previously verified that we are susceptible victims of false memory. Something as simple as remembering words of a list can result in false recognition, let alone real world situations where accuracy of detail is crucial. In simplified terms, the results of this data can be stated as follows: individuals can be biased into remembering things that never occurred to them (Roediger &McDermott, 1995). How so? In the actual study, the words that participants falsely identified as an old list item seemed familiar to them because they were related to those old list words. They made associations between these words and concluded that they both were, in fact, studied on the previous list. Their resulting false memories for these words were quite robust, thus emphasizing the reconstructive nature of memory (Bui et al., 2013).
The particular effect used in this experiment was quite robust as well. Even though participants were introduced to the idea of false memory first hand, when asked to read a background on the false memory phenomenon, most still exhibited false memory—myself included. One thing I did notice however was how quickly he original list items appeared on the screen. Each word appeared for one second and moved on to the next word, almost giving the reader no chance to fully process the words on the list. This could be so that the participant encodes only gist representations of the list, which in turn increases false memories when presented with new words that fit that gist. If I were to redo the experiment, I would perhaps show the words for longer intervals of time, or maybe the whole list at once. This would give participants the chance to actually study them, because realistically speaking; the memories we encode on a day-to-day basis are not always rushed. If doing so produces different results, this could reveal a factor that can potentially have an influence on human memory accuracy: time. All in all, through studying these experiments we learn that memory, no matter how obstinate the recaller is, cannot be trusted (Roediger & McDermott, 1995).