Response time and encoding strength: Moderating the choice similarity effect

The choice similarity effect (CSE), first demonstrated by Tulving, shows that forced-choice recognition memory decisions are more accurate but made with lower confidence when the distractor (A') is similar to the target (A) on that test trial, relative to when the target (A) is paired with a distractor, B', that is similar to a studied but untested item, B. Following Tulving, Experiments 1a to 1c examined variation in the CSE as a function of response times and deadline versus self-paced responding. Results showed that the accuracy advantage for A-A' test trials was most pronounced for the fastest responses and disappeared or reversed for slower responses, providing evidence for Tulving's prediction that A-A' pairs facilitate access to memory. These results also suggest that the A-A' advantage may be moderated by the use of response deadlines. Experiment 2 evaluated a prediction of matching models of recognition memory-specifically that the magnitude of the A-A' advantage increases with stronger encoding of studied items. Consistent with those models, Experiment 2 showed a larger A-A' advantage for items studied three times than for items studied once.