Abstract
Originally developed for adult humans, modern corneal-reflection eye-tracking has become applicable to diverse populations. This has driven its popularity in developmental studies with preverbal infants and comparative research with other animals, including nonhuman primates. However, these novel applications introduce unique methodological challenges, raising critical questions about data quality and the validity of cross-group comparisons. Here, we systematically investigated eye-tracking data quality (accuracy, precision, robustness) in 4-, 6-, 9-, and 18-month-old infants, adults, and zoo-housed chimpanzees. In Experiment 1, most strikingly, chimpanzees showed markedly lower accuracy compared to all human groups, including even 4-month-olds, who showed poorer accuracy than older infants and adults. Furthermore, some data quality metrics co-varied with key eye-tracking measures (e.g., number and duration of fixations, relative looking times), underscoring how data quality shapes inferences about psychological processes. Experiments 2 and 3 explored whether calibration adjustments can optimize data quality. In Experiment 2, modifying the conventional calibration procedure for chimpanzees yielded significant improvements. Compared to using an individual’s own 2-point calibration, another conspecific’s 5-point or a human 9-point calibration substantially enhanced accuracy. Conversely, in Experiment 3, a conceptually similar approach for human infants (using a peer’s or adult’s 9-point calibration) did not improve data quality compared to infants’ standard baseline (own 5-point calibration). Together, these findings offer profound implications and practical recommendations for navigating data quality challenges in comparative and developmental eye-tracking. Moving forward, we emphasize the necessity of species-specific method development and call for closer interdisciplinary collaboration to establish shared standards across application fields.