A dynamic computational model of gaze and choice in multi-attribute decisions

Area of Science:

• Cognitive Psychology
• Decision Science
• Neuroscience

Background:

• Attention significantly influences decision-making, with a tendency to favor options viewed longer.
• Existing models like the attentional drift-diffusion model (aDDM) capture option-wise attention but not attribute-wise attention.
• The interaction between option-level and attribute-level attention in choice remains poorly understood.

Purpose of the Study:

• To propose and test a novel multi-attribute attentional drift-diffusion model (maaDDM).
• To investigate how attention to both options and their attributes influences preferential choice.
• To examine the interplay between gaze allocation, attribute weighting, and decision outcomes.

Main Methods:

• Development of the multi-attribute attentional drift-diffusion model (maaDDM).
• Analysis of five eye-tracking datasets from two-alternative, two-attribute preferential choice tasks.
• Empirical testing of the maaDDM across diverse choice domains.

Main Results:

• Consistent option-level and attribute-level attentional discount factors were observed across datasets.
• Nonfixated options were discounted more than nonfixated attributes.
• A multiplicative discounting of the nonfixated attribute of the nonfixated option was found, indicating it was most discounted.
• Gaze allocation positively correlated with attribute weights, with higher-weighted attributes receiving more attention.

Conclusions:

• The maaDDM successfully accounts for attentional effects at both option and attribute levels.
• Attention plays a crucial role in modulating attribute importance during decision-making.
• Gaze behavior is a reliable indicator of underlying attribute valuations and influences preferential choice outcomes.