Theoretical note: Exploring Luce's (1963) low-threshold model applied to recognition memory

Area of Science:

• Cognitive Psychology
• Neuroscience
• Memory Research

Background:

• The mediation of recognition memory by discrete versus continuous processes remains a significant debate in cognitive psychology.
• Luce's low-threshold model, a discrete process model, has been historically overlooked but quantitatively matches data previously thought to support continuous models like signal detection theory.
• Previous research indicated quantitative equivalence between discrete and continuous models, prompting further investigation into their distinct explanatory roles.

Purpose of the Study:

• To replicate and extend the finding that Luce's low-threshold model quantitatively matches recognition memory data as well as signal detection theory.
• To investigate the differential fit of discrete (low-threshold) and continuous (signal detection theory) models to recognition memory data based on stimulus encoding strength.
• To propose an alternative framework for recognition memory that integrates both discrete and continuous mediation through strategic control processes.

Main Methods:

• Replication of previous findings across 8 distinct experiments.
• Quantitative comparison of the predictive accuracy of Luce's low-threshold model and signal detection theory.
• Analysis of model performance in relation to the encoding strength of stimuli presented during recognition tasks.

Main Results:

• The low-threshold model and signal detection theory were found to quantitatively describe recognition memory data similarly across 8 experiments.
• A key finding revealed that the low-threshold model better explains data from strongly encoded stimuli.
• Conversely, signal detection theory demonstrated a superior fit for data associated with weakly encoded stimuli.

Conclusions:

• The equivalence between discrete and continuous models in recognition memory is explained by their differential fit to stimuli based on encoding strength.
• A novel framework is proposed where a control process, potentially driven by metacognitive judgments of item strength, can shift memory mediation between discrete and continuous processes.
• This suggests that recognition memory may not be exclusively discrete or continuous but dynamically regulated by strategic control mechanisms.