A hierarchical Bayesian modeling approach to searching and stopping in multi-attribute judgment

Area of Science:

• Cognitive Psychology
• Decision Science
• Behavioral Economics

Background:

• Information overload is common in decision-making.
• Determining information search strategies and stopping points is complex.
• Existing models like take-the-best heuristics offer limited explanations.

Purpose of the Study:

• To investigate how individuals search for and select information during decision-making.
• To compare the explanatory power of a novel Bayesian inference model against conventional heuristics.
• To analyze decision strategies in a modified German cities task.

Main Methods:

• Administered a version of the German cities task in two experiments.
• Participants decided which of two cities had a larger population using available cues.
• Varied participant freedom in determining the number of cues examined.

Main Results:

• A novel hierarchical latent mixtures and Bayesian inference model provided a superior fit to the experimental data.
• The proposed model explained decision-making processes more comprehensively than conventional strategies.
• Findings highlight the effectiveness of Bayesian inference in modeling information search.

Conclusions:

• The novel Bayesian inference model offers a more complete account of information search and decision-making.
• Understanding these cognitive processes is crucial for improving decision support systems.
• Future research should explore the generalizability of this model across different decision domains.