Understanding the context of children's schoolyard activity can inform approaches by which to increase physical activity participation. Two major approaches exist for assessing this context- based on location or activity type. Location-based methods (e.g., Global Positioning Systems) may miss variations in activity type if children participate in multiple types of activities in each location (e.g., children engage in something other than playing with sand in a sandbox). Activity-type based methods for contextualizing activity (e.g., Observational System for Recording Physical Activity in Children, OSRAC) capture activity type, but not location, and children might participate in the same type(s) of activities in multiple schoolyard locations. The purpose of this study was to compare location- vs activity type-based approaches to characterizing children's schoolyard behavior. Preschoolers' (N = 50) schoolyard location and activity type (based on OSRAC) were video-coded and matched with physical activity monitor data (counts/s) for 1 outdoor period (~43 min). The results showed that multiple activities occurred within each location (median = 5), but the predominant activity typically matched the expected activity type (e.g., 66.9% fixed equipment play in the fixed equipment location). Children spent the majority of their time in open space (type or location; 37.6%–48.9%), but there were differences in the setting that promoted the most physical activity (teacher arranged type, but fixed equipment location). Activity types elicited varying physical activity levels depending on location (e.g., open space activities in open spaces vs on paths: 40.9 vs 60.7 counts/s). Activity type- and location-based approaches provide some similar, but potentially additive information about schoolyard behavior, which is an important methodological consideration for future studies. (author abstract)