The study aimed to determine whether degree of participation of children with cerebral palsy (CP) is influenced by where they live, as predicted by the social model of disability. Ninety-two per cent children with CP resident in Northern England and born 1991–1996 were entered into the study. Participation was measured by the Lifestyle Assessment Score and its six component domain scores. Regression analysis was used to investigate variations in participation. There were 443 children (265 male, 178 female; mean age 4 years 8 months [SD 1 year 1 month] at time of assessment) in the study. In the regression analysis the following factors remained significant with regard to level of participation: type of CP (167 with hemiplegia, and of those remaining 240 with bilateral spasticity); intellectual impairment (105 with IQ<50, 113 with IQ 50 to 70, and 225 with IQ>70); presence of seizures (115 with active epilepsy); walking disability (114 unable to walk, 81 restricted and needing aids, 186 restricted but unaided, 62 unrestricted); communication problems (61 no formal communication, 51 use alternative formal methods, 126 some delay or difficulty, 205 no communication problems). After adjustment for these factors, there were significant variations with regard to level of participation in the Lifestyle Assessment Score by district of residence. The magnitude of these variations in Lifestyle Assessment Score between districts is similar to that accounted for by severe intellectual impairment. Similar models were obtained for four of the six domain scores. For one of these four, restriction of social interaction, the significant variation between districts was minimally influenced by the underlying type of CP, walking ability, or presence of seizures. Higher levels of participation among children with CP are associated with residence in certain districts. This is not attributable to variations in case-mix or functional capacity of the children. Participation of children with disability is partly a product of their environment.