The schizophrenia-crime relationship was studied in 151 research participants meeting DSM-IV criteria for schizophrenia or schizoaffective disorder and with histories positive or negative for criminal charges, convictions and offences involving violence. These crime-related variables were regressed on a block of nine predictors reflecting non-specific illness context (e.g. demographic, social) and a block of 14 predictors reflecting specific illness content (e.g. symptoms). Context variables predicted charges, with unique contributions from employment status, education and substance use. Further significant validity was provided by content-related predictors including symptoms (paranoia, depression, low energy), but not cognitive performance (verbal and non-verbal ability, working memory, processing speed, verbal memory, word fluency, inhibition, practical cognition). In contrast, neither convictions nor violence were predicted by illness context or content variables. These results suggest that specific contextual and intrinsic aspects of schizophrenic illness make interaction with law enforcement and therefore criminal charges more likely.