This article looks at implicit memory for critical lures in the DRM paradigm. There have been a couple of prior attempts to find implicit false memories.

• Bennett (1996) found evidence for an automatic component using process dissociation
• McDermott (1996) used stem completion and fragment completion measures and found significant priming for the critical lure when those two measures were combined (neither was significant by itself
• Smith, et. al (1998) failed to find evidence for significant priming in stem completion when responses had to be made in under 2 seconds

Yet McKone and Murphy raise certain objections to McDermott’s (1996) study. These objections include:

• Possible contamination by explicit memory
• No instructions to avoid explicit recall
• No self report to determine if explicit recall was used
• 12-20 seconds allowed for response
• Similar pattern of results for explicit and implicit false memories which furthers the contamination account according to McKone and Murphy

Method

Subjects were presented with 8 DRM type lists. Lists were shown visually at a rate of one word every 1.5 seconds. After all 8 lists were presented there was a five minute filled delay followed by one of two types of tests.

Subjects tested with either a stem completion test or stem cued recall test.

• The stem completion test involved presenting three letter word stems (e.g. vac-) which the subject was to complete as quickly as possible with the first word that came to mind.
• The stem cued recall test made use of the same three letter word stems but the subject was supposed to use the stem as a cue for explicit recall and to leave blank those where there had not been a corresponding word.

They included a questionnaire at the end to see if subjects in the implicit condition used explicit memory and concluded that they did not.

Results

Not surprisingly they found evidence for explicit recall of both critical lures (corrected recall= 0.38) and targets (corrected recall = 0.34).

They also found evidence for priming in stem completion for both critical lures (corrected stem completion = .18) and targets (corrected stem completion = .17)

This time they used an auditory presentation of words. The idea here is that usually stem completion is influenced by the match between the study modality and test modality.

One prediction might be that one would find reduced priming for targets but not for CLs because the CLs rely on activation of a semantic representation that can be accessed through multiple modalities.

Results

Again they found evidence for explicit recall of both critical lures (corrected recall= 0.41) and targets (corrected recall = 0.43).

There was now no significant priming for either critical lures (corrected stem completion = .06) or targets (corrected stem completion = .06).

This is really really cool but kind of odd. "[O]ur results show that priming for lures depends on matching the modality of the test phase to the perceptual modality in which the associated items were studied."

The point in Experiment 3 is to look at the effects of multiple presentations. Prior research has shown that multiple presentations of targets tends to increase true memories while deceasing false memories.

But Schacter, et al (1998) found that in amnesics showed no decrease in false memories with repeated presentations. Implicit memory should work the same way. Increased verbatim retrieval should not limit reliance on gist.

So in Experiment 3 they went back to visual presentations and presented items 5 times for study.

Results

Compared to Experiment 1, Experiment 3 produced higher levels of stem cued recall for targets (CR = 0.59) and lower levels of stem cued recall for critical lures (CR = 0.21).

They also found evidence for priming in stem completion for both critical lures (corrected stem completion = .15) and targets (corrected stem completion = .25)

Compared to Experiment 1, Experiment 3 produced no significant change in levels of priming for targets or critical lures. (Jim’s note: although the pattern of results seems to be for targets to be non-significantly higher but no such trend for CLs)

The finding of modality specific priming of critical lures in stem completion is pretty amazing! McKone and Murphy make two suggestions about how this could occur:

• Our old buddies IARs: One possibility is that the presentation of targets leads to the production of implicit associative responses and that these IARs involve activation of surface forms that match the modality of presentation of the targets.
• Organization of Perceptual Representation System: Argument here is that within a perceptual representation system there are direct links between the surface forms of related items or (more strongly) that there are no modality independent representations at all.

Taking FTT to task: McKone and Murphy provide three arguments against a FTT interpretation:

• True and false memories are similar in a variety of ways.
• McKone and Murphy argue that under most circumstances true memories reflect a mixture of gist and verbatim and that gist leads to false memories only when the test requires verbatim level specificity.
• Level of specificity is a continuous not dichotomous variable