Derived false memories using a respondent-type (ReT) procedure

The present investigation used a respondent-type (ReT) training procedure to generate derived false memories. A one-to-many ReT training procedure was implemented in order to establish two stimulus equivalence classes, each consisting of one shape and 24 random words (i.e., Class 1 and Class 2). A partial test for stimulus equivalence with a subset of stimuli from each class followed. Failing an equivalence test resulted in additional ReT training and equivalence testing on new subsets of stimuli. After passing an equivalence test, participants were presented with 12 study-list words from Class 1 for memorization, followed by a distraction task. Finally, free recall and recognition tests for the study-list words were implemented. False recall and false recognition were more frequent for nonstudied Class 1 words than for nonstudied Class 2 words. These derived false-memory effects were more pronounced among those participants exhibiting more training and testing cycles and higher accuracy on stimulus equivalence tests. Furthermore, false recall and false remembering of nonstudied Class 1 words were more frequent for words that had been equivalence-tested than for words that had not been equivalence-tested. These results show how responses to contiguous stimuli could produce derived false memories and also highlight the role played by the equivalence test in increasing their emergence.