Exploring general outcome measures for decision making using a wearable eye tracker

Abstract

Studying gaze behavior using eye tracking can give insights in decision making processes regarding food choices. However, commonly used analyses that involve area of interests (AOI’s) are difficult to implement in measurements made in complex environments like a supermarket. In these cases it could be beneficial to use general outcome measures that do not use AOI’s. In this study, we investigated with a wearable eye tracker if general outcome measures can account for differences in gaze behavior for customers that already decided what to buy (resembling a search task) or did not yet decided what to buy (resembling a decision task). Additionally, we took into account their familiarity with the supermarket (resembling learning effects). This study substantiates the Natural Decision Segmentation Model by providing evidence for differences in gaze behavior during the ‘orientation’ compared to the ‘evaluation and verification’ phase using general outcome measures. More specifically, undecided unfamiliar customers had higher standard deviations of saccadic amplitude and velocity, and longer mean saccadic durations in the orientation phase compared to other customers. Although the results should be considered with nuance, they seem to support the feasibility of wearable eye tracking research in complex and applied settings by means of general outcome measures analyses.