Eye tracking is a tool that captures where an individual is looking on a screen, which can tell us about various types of perceptual, social, cognitive, and emotional processing. While eye tracking systems are typically designed for adults, they are increasingly used with human and primate infants, whose data may be noisier and less precise. Yet there is a lack of consensus on the most effective methodology for improving data quality in screen-based infant eye tracking studies, particularly those with different species and age groups. Here we examined how some key methodological choices, including enlarging and prolonging the areas of interest (areas on the screen in which the eye tracker detects looks), influence data quality using a Tobii TX300 eye tracker. We tested humans (N=119) at 2, 4, 6, 8, and 14 months of age and rhesus macaques (N=21) at 2 weeks, 3 weeks, and 6 months of age. We found that larger and longer areas of interest improved the proportion of looks detected-increasing valid looks relative to errors-suggesting improvements in capturing eye gaze. These benefits varied by age and species, suggesting that the ideal methods to use may vary depending on the populations studied. To maximize data quality and minimize data loss, we recommend adjusting data collection and extraction approaches for the age groups and species studied, which may better standardize and improve the replicability of eye tracking research.
