Activity spaces are used to capture patterns in urban mobility and to portray the spatial distribution of day-to-day activities. The literature exploring variation in individual activity spaces identifies strong associations between several activity space characteristics and the built environment of the residential location. This cross-sectional study adds to this evidence by examining whether these associations persist after adjusting for residential self- selection. Adults’ everyday mobility was studied using public participation GIS, a participatory mapping method allowing the large-scale collection of laymen-produced spatial data. Activity spaces were defined with a customized minimum convex polygon modelled on the respondents’ frequently visited locations. We used linear regression and multinomial logistic regression analyses to study the associations between residential preferences, residential location, and activity space size and centricity. According to our results, residential location signif- icantly influences activity space size and polycentricity in models adjusted for stated residential preferences and individual-level covariates.