Background:

The DRM paradigm developed by Roediger and McDermott, adapted from Deese, shows that when given a list of 12-15 closely related words, subjects will often falsely recall a non-present critical lure. For example, when given a list featuring the words bed, rest, & awake, a subject might report hearing sleep. Because this effect has become so commonplace, researchers are now looking for ways to manipulate the experiment in order to lower false recall and recognition. In this particular study, Gallo et al. are focusing on why some lists yield stronger false recall responses than others, especially considering the lists are all composed in the same manner (the first 15 associates of a critical word that is not presented).

To explain this phenomenon, Gallo et al. will look to several different theories. One theory, the Gist-based theory, suggests the critical item is remembered because it is consistent with this gist representation. The critical item need not be activated prior to retrieval, rather it is remembered because of the semantic overlap it shares with the list words. Similarly, the feature theory claims that semantic features taken from list items overlap with those of the critical word, which in turn leads to false remembering. However, Gallo et al. have developed their own theory to account for false memory, the activation/monitoring theory. The theory suggests that when processing the list items (both at study or test) the critical non-presented associate is activated, and false remembering results from a failure to correctly monitor the source of activation. The difference between their activation theory and that of the semantic overlap theories is what you attribute the false memory to, be it falsely remembering internally generated events as having been externally perceived (activation) or remembering events that were not previously encountered, but were consistent with one’s understanding (semantic overlap).

Experiment 1

Deese (1959) found that a list’s backward associative strength is predictive of it’s likelihood to produce a false recall of the critical item (MBAS). Taking this information, Gallo et al. wanted to compare the amount of false recall in lists with high MBAS against list with low MBAS. Using 28 study lists of 15 words each, they chose 20 lists that should yield relatively low scores of false recall, and 8 which should yield relatively high scores. After the lists had been studied and recalled, a final recognition test was given consisting of 168 items ( half of which were “old” and half of which were “new”). They were also asked to rate their confidence on a scale of one to four.

Results

Gallo et al. found low levels of false recall for most of the lists with low MBAS. They also found that lists with low MBAS resulted in relatively lower false recognition. What’s more, the average confidence level accompanying false recognition from each list was positively correlated with the probability of false recognition from that list, as well as the MBAS.

Experiment 2

In this experiment, Gallo et al. manipulated presentation rate to see if it would have any further effects on false recall. For example, would slowing down the presentation rate increase false remembering because subjects would have more time to encode information that could support false remembering? Or, would it lower false remembering because subjects would have more time to engage in item specific processing , which would allow for more distinctive recollection of individual items? In terms of the activation/monitoring theory, slowing presentation rate should enhance the editing process between memories for list items and critical items. In this experiment, 90 students were broken up into several groups where they were presented with 16 lists of 15 words for study and immediate recall. A 3x2 mixed factorial design was used with presentation rate ( .5, 1, or 3s) manipulated between subjects, as well as list type manipulated within subjects.

Results

The findings of Experiment 2 demonstrate that both false recall and false recognition decreased when presentation rate was slowed down. Moreover, it showed that presentation rate had the same effects on both strong and weak word lists. It also reinforced the finding of experiment 1 that strong lists yielded a higher rate of false remembering. Confidence levels increased for recognized list items with the slower presentation rates as well.

Experiment 3

With this experiment, Gallo et al. attempted to explore the extent to which testing recall effected the recognition tests. The method used was the same as in experiment 2 with two exceptions. First, between the lists, instead of doing recall, subjects were rewarded with math problems. Second, when given the recognition test, half of the subjects were to give confidence judgments, whereas the other half made “remember/know” judgments.

Results

One finding the experiment suggested was that the effect of rate on false recognition in experiment 2 was driven by prior recall. Also, they found that some critical items elicit more high confidence and remember false alarms than others, suggesting that false recognition in strong lists is more compelling than that from weak lists. There were several differences were found between experiments 2 & 3. For example, list items were not recognized as often in experiment 3 as in experiment 2. This suggests that the act of prior recall can boost recognition performance for list items. Furthermore, the rate of false alarms was considerably higher in experiment 3 than in experiment 2.

General Discussion

Here are the major results: