Testing theories of recognition memory by predicting performance across paradigms.

Area of Science:

• Cognitive Psychology
• Neuroscience
• Psychological Measurement

Background:

• Recognition judgments are often explained by signal-detection theory (SDT) assuming a single familiarity process.
• Dual-process theory (DPT), a hybrid SDT model, proposes an additional recollection-based process.
• The validity of these models is debated regarding their underlying assumptions and parameter stability.

Purpose of the Study:

• To test the predictions of SDT and DPT models of recognition.
• To compare the invariance of parameter estimates between yes/no (Y/N) and 2-alternative forced-choice (2AFC) paradigms.
• To evaluate the theoretical validity and computational flexibility of SDT and DPT.

Main Methods:

• Two experiments were conducted comparing DPT and SDT predictions.
• Parameter estimates for recollection were compared between Y/N and 2AFC testing formats.
• Model fits were assessed by the variance explained by each model's predictions.

Main Results:

• Dual-process theory (DPT) recollection estimates showed poor correlation between Y/N and 2AFC paradigms.
• Signal-detection theory (SDT) predictions explained more variance in Experiment 2 compared to DPT.
• Findings suggest limitations in the parameter invariance assumed by DPT.

Conclusions:

• The results challenge the assumption of a single, invariant recollection process in DPT.
• SDT may offer a more robust account of recognition judgments across different testing paradigms.
• Further research is needed to reconcile theoretical validity with computational flexibility in recognition models.