These experiments deal with the DRM paradigm. The authors make an argument in the introduction that false memories in the DRM paradigm are produced at encoding. To test this possibility they varied the context in which DRM items were presented. For each item on their DRM list they came up with an item that would tend to shift the context away from the critical lure. For instance, for the critical lure Soft, the word hard was presented at study. However, sometimes the word hard was followed by the word hat creating an overall context that is unrelated to the word soft. Very cool.
Methods
In Experiment 1 they used this process to create five kinds of lists (see Table 1 for examples):
What's kind of cool about this is that the exact same items are being used and all that's changing is the way they are arranged.Eight Related/Eight Unrelated: The eight DRM words (e.g. hard, etc.) followed by the eight context words (e.g. hat, etc.)described above. Because the context words weren't presented with their partners false recall should be unaffected. Four Related/Four Unrelated: Four DRM words are presented, followed by the eight context words, followed by the four DRM words. Notice here, the context words are still not presented close to the studied words, but the list theme is broken up in the middle. So there should be less false memory. Two Related/Two Unrelated: Two DRM words are presented, followed by two context words (unrelated to the preceding words), with this pattern repeated four times. One Related/One Unrelated: Alternate between the DRM words and the context words. However, the context words were never the one's that had to do with that particular DRM item. For instance, participants studied hard, tone, light, trade but not hard, hat, light, bulb. CU: Alternate between the DRM words and the context words. But in this case, the context words were arranged so that they followed the related word (e.g. hard, hat, light, bulb). According to the account, this condition should produce fewer false memories because the interpretation of the DRM words is being biased away from the critical lure.
Subjects studied 12 lists altogether. Two lists of each type and two practice lists. After each list subjects did multiplication problems for half a minute and then spent 90 seconds recalling the items from that list.
Results
Methods
Some participants might think of the critical lure immediately upon being presented with the DRM word, before being presented with the context word that could reduce false recall. To deal with this issue the authors presented the words in pairs simultaneously on the screen for four seconds.
Results
Experiment combined methods from Experiment 1 and 2. Half participants studied items one at a time and half studied them in pairs.
Participants thought out loud as they studied items and also during test. The researchers coded for (1) whether repetition or elaboration was used as a strategy at encoding (2) whether and how often the participant spontaneously mentioned the critical lure at study (3) whether the critical lure was mentioned during retrieval (4) whether or not the critical lure was falsely recalled.
Results
The main issue then was whether the things that happened at encoding influenced the things that happened at retrieval. Here's what they found:
They argue that spreading activation accounts should have trouble handling this. In other words, its not all mean backward associative strength although as we've seen that's part of it (Roediger et al, 2001).
There was a dissociation between true recall and
false recall. The condition that had the highest true recall also
had the lowest false recall.
|
|
|
|
|
|
|
|