Information Search within a Webpage: Boundedly Rational Models of Eye Movements and Clicks

Abstract

We build a model of information search within a web-page using eye-tracking data collected during two incentive-compatible, online shopping experiments. We assume that shoppers make eye fixation decisions based on the attractiveness of the visual elements on the page and the effort of moving their eyes towards those elements. These elements contain product information and clickable links to new pages. After each fixation, shoppers choose either to continue search or to end search by clicking on a link within a fixated element when its attractiveness crosses a decision threshold. Our model contains dynamics that allow for Bayesian updating of product information, decision threshold trends, and one-stage ahead decision inputs. Using posterior predictive checks, we demonstrate that our model makes accurate in-sample predictions regarding the sequence of fixations and clicks. In Experiment 1, we use counterfactual simulations to predict the effects of hypothetical product layouts. In Experiment 2, we empirically test these counterfactual predictions. We also conduct sensitivity analyses to assess how near to optimally shoppers searched. These results provide guidance for the design of online product displays.