Response time models separate single- and dual-process accounts of memory-based decisions

Area of Science:

• Cognitive psychology
• Neuroscience
• Behavioral economics

Background:

• Human decision-making frequently deviates from economic rationality due to cognitive biases.
• The memory bias, where individuals favor options they recall better regardless of utility, is empirically supported but lacks a clear cognitive basis.
• Existing models offer competing explanations for the memory bias, including single-process and dual-process accounts.

Purpose of the Study:

• To rigorously compare single-process and dual-process computational models of the memory bias.
• To determine the underlying cognitive mechanisms driving memory-based decision-making.
• To investigate how memory influences choice preferences and response times.

Main Methods:

• Development and comparison of single-process and dual-process computational models.
• Empirical testing using a memory-based decision-making task.
• Analysis of behavioral data, including choices and response times.
• Application of hierarchical Bayesian parameter estimation for model comparison.

Main Results:

• Both single-process and dual-process models predicted the memory bias effect when analyzing choices alone.
• Qualitative and quantitative differences in response times distinguished the models.
• The single-process account provided a superior explanation for the observed behavioral data.

Conclusions:

• A single-process model, involving biased evidence accumulation, better explains the memory bias in decision-making.
• Response times are crucial for differentiating between single- and dual-process models of cognitive bias.
• This study advances the understanding of memory's role in decision-making and offers a methodological framework for comparing cognitive models.