The Quebec Child Mental Health Survey (QCMHS) was conducted in 1992 on a representative sample of 2400 children and adolescents aged 6 to 14 years from throughout Quebec. Prevalences of nine Axis-I DSM-III-R (American Psychiatric Association, 1987) mental health disorders were calculated based on each informant (for 6–11-year-olds: child, parent, and teacher; for 12–14-year-olds: child and parent). Informant parallelism allows the classification of results of the demographic variables associated with disorders in the logistic regression models. This strategy applies to group variables (correlates of disorders) whereas informant agreement applies to individual diagnoses. Informant parallelism implies that results for two informants or more are in the same direction and significant. In the QCMHS, informant parallelism exists for disruptive disorders, i.e. in two ADHD regression models (child and parent) higher rates among boys and young children, and in three oppositional/conduct disorders regression models (child, parent, and teacher) higher rates among boys. No informant parallelism is observed in the logistic regression models for internalizing disorders, i.e. the patterns of association of demographic variables with anxiety and depressive disorders vary across informants. Urban-rural residence does not emerge as a significant variable in any of the logistic regression models. The overall 6-month prevalences reach 19.9% according to the parent and 15.8% according to the child. The implications of the results for policy makers and clinicians are discussed.