In this paper we exploit detailed data on over one million individuals in England to more precisely identify and separate the associations between several measures of physical inactivity, different aspects of socioeconomic position and a wide range of local geographical factors. (University of Bristol)