A theory of multiattribute search and choice

Area of Science:

• Cognitive Science
• Decision Science
• Computational Neuroscience

Background:

• Decision-making often involves numerous options with multiple attributes, requiring efficient information processing.
• Individuals must limit information search to manage cognitive load during complex choices.

Purpose of the Study:

• To propose and test a novel hierarchical Bayesian cognitive model of information search in multiattribute, multialternative decisions.
• To explain how attention is allocated in a goal-directed, efficient manner during decision-making.

Main Methods:

• Development of a hierarchical Bayesian cognitive model for information search.
• Conducting a preregistered eye-tracking experiment minimizing bottom-up attention influences.
• Testing the model's ability to account for empirical findings and predict choice dynamics.

Main Results:

• The proposed model successfully explains established findings like the correlation between gaze time and choice probability.
• The model accounts for attention-choice interactions in multiattribute decisions, outperforming alternative theories.
• The model predicts decision speed dependencies on option quantity and value.

Conclusions:

• The hierarchical Bayesian model offers a robust framework for understanding attention and search in complex decisions.
• The findings highlight the goal-directed nature of information search, influenced by evidence, importance, and uncertainty.
• This research advances our understanding of the interplay between search, attention, and valuation mechanisms.