With recent efforts to link environmental factors to human behavior, specifically people's physical activity space (PAS), knowledge in special statistical techniques are needed to handle this type of data. Although commonly used univariate and multivariate statistics can still be used, more specialized spatial statistical techniques have been developed for analyzing activity space. These techniques are based on theories with an emphasis on creating new geographical and personal variables. While routinely used in geography, they are relatively new to physical activity research. The purpose of this presentation is to introduce and review spatial statistical techniques that can be used to evaluate geographic areas and people's time-space interactions. The focus will be on several geographical statistical variables, including confidence ellipses, Kernel densities and minimum spanning trees. Confidence ellipses are travel probability fields for individuals calculated by confidence intervals from longitude and latitude directional data obtained through a global positioning system. Kernel densities combine location choice and frequency of visits to a geographical area in order to represent the routinely used activity space. Total area covered and volume of activity can be calculated for the individual and the area, which provides valuable information about spatial behaviors of a community. With confidence ellipses, kernel densities and detailed information about available resources through GIS, minimum spanning tree, a network of travel paths, can be created to represent the person's perception of available physical activity space. The person's space-time interaction can also be examined with the creation of time-space prisms. This technique allows for the examination of available space for an individual, considering time constraints between activities. This technique evaluates the available space an individual has at times when physical activities can be performed between habitual activities. These prisms can be evaluated for size, quantity and quality. Because of the hierarchical nature of activity space, it is likely that new and advanced multilevel statistical methods, such as hierarchical linear and growth-curve models will also be useful statistical tools for analyzing spatial data. These geographical statistical variables will help identify differences in various groups' activity space and the role different environmental factors play with location and activity choices. Patterns and space, based on season, type of day and weather can also be examined.Keyword(s): community-based programs, measurement/evaluation, physical activity