Deficits in social attention constitute one of the core symptoms of autism in toddlers. While a number of hypotheses have been advanced, the mechanisms underlying poor attention to people in naturalistic settings are poorly understood. In recent years, this area of research has been actively investigated using eye-tracking technology. Studies presented in this panel examine factors responsible for toddlers’ atypical attention to people (Talks#1,2,&4), investigate individual variability in social attention and their links to phenotypic features (Talk#3&4), and illustrate the utility of fine-grained and less assumption-laden analytical methods to further our understanding of the factors driving visual behavior of toddlers with ASD (Talks#2&4) based on a series of conceptually- linked tasks. This panel will address several methodological considerations inherent in eye- tracking methodology: importance of accounting for both top-down and bottom-up influences on visual attention, approaches to parsing variability within ASD samples, as well as methodological aspects of analyzing eye-tracking on dynamic scenes. Taken together, this panel will 1) present a comprehensive account on abnormalities of visual scanning in response to complex social scenes, 2) address the issue of heterogeneity of attentional responses; and 3) offer insights into fine-grained analytic approaches to analysis of dynamic eye-tracking data.