|
Roediger, H.L. & McDermott, K.B. (1995). Creating False Memories: Remembering Words Not Presented in Lists. Journal of Experimental Psychology, 21, 4, 803-814. Henkel, L.A. & Franklin, N. (1998). Reality Monitoring of Physically Similar and Conceptually Related Objects. Memory & Cognition, 26(4), 659-673. Concepts and Memory In week 13, our class focused on how categories and concepts affect memory. Two papers were discussed; Roediger & McDermott (1995), which investigated intrusions of words never presented in a list and Henkel & Franklin (1998), which examined reality and source monitoring of physically and conceptually similar objects. In experiment 1, subjects were presented with a verbal list of fifteen words. The lists were related to each other by a key, non-presented word. For instance, one group of words had a variety of relations to the key word “sleep”, the non-presented word. Then, subjects were asked to write down the words that were presented in the list. Free recall followed the presentation of each list. After all six lists, subjects were asked to judge whether each word had been previously presented by means of a recognition test. The probability of recalling a key word was the same as the probability of recalling a word presented in the middle of the list. The probability of recognizing a key word that had not been presented was 84%. In experiment 2, the authors used a much larger set of lists. There were three main reasons for conducting Experiment 2, to examine whether or not prior recall affected recognition, to investigate the rate of false-alarms for critical items not presented in lists, and to elicit remember/know judgments. False recall enhanced false recognition and increased the number of remember judgments for the critical items. A great deal of time in class was spent dissecting the lists. The class tried to isolate how the words in the list were ‘related’ to the critical item. The group wondered what makes each list cohere. Is it the association of the words to the critical item that leads to false recall? One list that the class spent a great deal of time looking at was the ‘Sweet’ list. Intuitively, it made sense that ‘sour’, ‘candy’, and ‘sugar’ were presented items. However, why would ‘heart’, ‘tart’, or ‘soda’ be part of the list? They do not immediately bring to mind ‘sweet’ nor do they invoke any association with the word other than through a chain of other words. For instance, ‘heart’ would first have to invoke an idea of a heart-shaped cake or a heart-shaped box of chocolates before it could invoke ‘sweet’. Why wouldn’t ‘heart’ bring to mind the words love or beats? Perhaps these seemingly loosely related words serve as primes. The class also discussed how the first few words in the list were the strongest associates of the target word. It seemed that this was to set up the concept first. If the correct concept was not immediately elicited, then people would be garden pathed toward a different key word. Would the list work if the strongest associations were not given at the very beginning? Also, would the lists work if we did not segregate them? In other words, if the experimenters gave the subjects all of the lists at once and then asked them to free recall, would the key words intrude? The second paper by Henkel & Franklin (1998) addressed the concerns of reality monitoring. Reality monitoring is similar to source monitoring, which is the process of remembering where any particular piece of information was presented. Marcia Johnson, who founded the theory of source monitoring, argues that the source of originally presented information can be misattributed to another source of presentation. For example, something that we may have only thought of, we can misattribute to something that we believe we heard. This is because memories do not get overwritten by new information and can be accessed individually. Three source monitoring experiments were conducted. The first experiment found evidence that people made more source monitoring errors for items (both imagined and perceived) paired with physically similar items as opposed to objects paired with conceptually related objects. The second experiment found evidence for higher error rates when the perceived items were both physically and conceptually similar to viewed items. The last experiment showed that participants were more apt to make errors when the items were physically similar to perceived items and when the number of physically similar items increased. The class questioned if the errors could be due to tracking the labels of the perceived and imagined items. On page 662, the perceived magnifying glass and the imagined lollipop are nearly identical. If the two items are identical, then perhaps it is not a question of monitoring where the picture was presented, but instead, of monitoring the labels of the items. The experiment would have been more conclusive if there was a defining element of the magnifying glass, which would have made it slightly different than the lollipop. Hence, perhaps there needed to be a distinctively different part of one of the pictures to reduce the source monitoring effect. One person brought up the point of ecological validity. Are line drawings ecologically valid? Can one assume that any findings of the experiment could be replicated in the real world? Perhaps with more robust items, people are better at source monitoring. Pictures or three-dimensional objects (instead of line drawings) may have brought about different results. The class posed the question of whether or not concepts actually cause or prevent source monitoring errors. This is one place where the Roediger & McDermott (1995) paper ties into the Henkel & Franklin (1998) paper. Roediger & McDermott (1995) would argue that concepts directly affect the ability to accurately monitor the source of words. Both papers were concerned with the creation of false memories and how these false memories are related to concepts. In the Roediger & McDermott paper, the source of the false memory may have been due to imagination of the key word or to the assumption that the key word was presented (perhaps primed by other words conceptually related to that particular word). This false memory for the key word was then confused with the memory for presented words, leading to false recall and recognition. If salient information was tied to the presentation of the words, such as remembering the sound of the words or the presenter’s tone of voice, it is less likely that a false memory for the key word would have formed. In essence, salient contextual cues act as tags for the source (imagined or presented) of the memory. Likewise, if the pictures in Henkel & Franklin’s experiments came with salient tags, participants would have been less likely to misattribute the sources of their memories. This series of experiments showed that concepts can work for and against accurate memories in a number of ways. For instance, Roediger & McDermott showed that concepts are likely structured with varying degrees of association. One’s concept of ‘sweet’ , for example, has several strong associates, and a few weak associates that all prime the concept. This can be an aid for memory because individuals do not necessarily have only one way of accessing a word. On the other hand, as Henkel & Franklin have shown, concepts can cause intrusions in memory and cause inaccuracies. |
|
This page was last updated: |