Miller, M.B. and Wolford, G.L. (1999). The role of Criterion Shift in False Memory. Psychological Review, 106, 398-405.

Reason for this study: Many studies have looked at memory distortions. Work in such areas as eyewitness testimony, recovered memories, and false memories are examples of such studies. Roediger and McDermott reintroduce a paradigm designed by Deese to study false memories. In the present study the DRM paradigm is reconstructed to test if the results could be explained in terms of criterion shifts.

Experiment 1:

• 24 undergraduates participated for extra credit in an introductory psychology course. Participants were tested in six groups ranging from three to five participants.
• Within subject design. Each participant heard twelve of the original list from the Roediger and McDermott study plus two unrelated lists.
• In four lists, the critical lure was placed in the first position of the list. The displaced word was used as a non-presented related list item during recognition. In four lists the critical lure was placed in the tenth position of the test. The displaced word was used as a non-presented related list item during recognition.
• Unrelated lists consisted of six words from the thematic lists that were not presented to the participants. The other nine words were chosen from categories not related to the thematic list.
• Design closely followed the Roediger and McDermott study.
• After each list was presented, half the participants did a filler task while half did free recall for two minutes.
• After all lists were presented, the participants were given the recognition test. The test asks them to make old/new distinctions and remember/know distinctions.

• Participants responded old to critical lures as often as they did to presented items. This does not mean that critical lures behave as presented items since the presented critical lures raised the old responses from .81 to .97. (table 1)
• When critical lures were presented, 80% of participants gave remembered judgements to these items versus 17% know judgements. When critical lures were not presented, 44% gave remembered judgements versus 37% know judgements.
• On free recall, 27% of the critical lures that were not presented were remembered versus 96% of the presented critical lures.
• d(a) was used to measure sensitivity and bias. (Table 2)

Experiment 2

• Identical to Experiment 1, the participants gave confidence ratings instead of r/k judgements.

• Results were very similar to Experiment 1.
• d(a) used to measure sensitivity and bias for the confidence rating. (Table 2)

• Results show that performance is more consistent with criterion shift than with change in sensitivity.
• Researchers disputed the idea that the results could be explained by shifts in distortion rather than shifts in criterion.
• Research proposes principles to distinguish between evidence of false memories and simple errors of memories. (1) Condition must product as many false memories as a similar condition produces true memories. (2) The participants indicate on r/k or confidence judgements that they believe that the memory is real.
• Evidence that at least some false memories aren't produce at encoding, but due to a criterion shift during retreival.