Abstract Purpose The present article investigated individual and aggregated effects of cannabis-related perceptions and other cannabis-related indicators on 12-month cannabis use prevalence and frequency among 15–16 year olds using multilevel analysis across 32 European countries. Methods Data on cannabis use, perceptions of availability, risks and friends’ use as well as socio-demographic characteristics were taken from the 2007 European School Survey Project on Alcohol and Other Drugs. At the country level, aggregated measures of the perceptions were used. Data on cannabis price and 12-month cannabis use prevalence in the total population were taken from the World Drug Report. The analytical sample comprised 86,107 students (82.5% of the overall 2007 international database). Results Strong and persistent individual-level effects were identified for perceived availability, perceived harm, and the number of cannabis using friends. The effects on cannabis use prevalence and frequency were more pronounced than country-level effects. At the country level, aggregated perceived peer consumption and population prevalence were significant predictors, whereas price was not found to be related to both outcome variables. The association between perceived friends’ use and cannabis use was moderated by aggregated perceived availability. Conclusions Proximal influences related to the immediate social situation seem to be more strongly associated with cannabis use than do distal influences related to social contexts, emphasizing the importance of personal attitudes and perceptions in substance use behavior. Prevention programs may focus on informing adolescents about the potential risks of cannabis and on correcting misperceptions of social norms. Policy measures may target on reducing visibility of drug use.