Eye movements are intricately coupled with decision-making, serving as a window into a decision-maker's attentional processes. In social settings, we intuitively use this principle to infer other persons' preferences while they are deciding. Showing the gaze pattern of one participant to another in a coordination game leads them to better understand each other's preferred choices. In our study, we are using this ability to test whether a seller can reach better agreements with a buyer if they can see their eye movements in a cooperative bargaining setting. In the task, the seller tries to sell products with multiple attributes to a buyer. The seller’s goal is to figure out the buyer's preference to be able to offer (and ultimately sell) appropriate products. For making such inferences, the seller has access to choices, response times, and eye movements. We hypothesize that sellers can use buyers' eye movements to infer their preferences, thereby improving the efficiency and outcomes of bargaining processes.

To explore how sellers can learn from buyers' decisions, we developed a model for the buyer that assumes a probabilistic cutoff value at which the buyer accepts an offer. We also designed a learning model for sellers according to which the seller infers attribute preferences from the buyer’s decisions. Simulation results suggest that sellers relying on feedback from buyers' choices perform better. The study is currently in the piloting stage, and we expect to be able to present the data from the main experiment by the start of the conference.