This study aims to identify the locations where unstructured socializing is related to adolescent offending. 'Locations' refer to private, semi-public and public spaces, which are further categorized into public entertainment, public transportation, other semi-public settings, streets, shopping centres and open spaces. Detailed longitudinal data, derived from space-time budget interviews among 615 respondents in the age range 11-20 years, about hourly activities and the whereabouts of adolescents are analysed. A random intercept panel model is used to control for selection effects that occur when crime-prone individuals prefer crime-conducive locations to other locations. Findings indicate that unstructured socializing is positively related to offending and that this relationship strongly depends on the location in which it occurs.