Social deficit is the core symptom of pervasive developmental disorder. In other child psychiatric disorders, social problems are also described but mainly as a result of the disease symptomatology. However, some recent studies suspect that in several disorders such as attention deficit hyperactive disorder, patients have an endogenous social disturbance. The aim of our research was to study abnormal child social behaviour in several disorders, using a dimensional approach. It is a preliminary validation study of the French version of the Children's Social Behaviour Questionnaire, a dimensional instrument constructed by Luteijn, Minderaa et al. Five clinical groups, according to the DSM IV criteria, formed a population of 103 children aged 6 to 16 years old: autistic disorder, attention deficit hyperactive disorder (ADHD), emotional disorder (anxious, depressed), mental retardation and normal children. Parents completed the Child Behaviour Checklist (CBCL) and the Children's Social Behaviour Questionnaire (CSBQ). The research worker and the child's physician completed a data form. The data form included information about medical history, development and socio-demographic criteria. The CBCL explored children's behaviours and general psychopathology, and included social dimensions (withdrawn, social problems, aggressive/delinquent behaviours, thought problems). The CSBQ, a dimensional questionnaire, explored children's social behaviours and included five dimensions: <<acting-out>>, <<social contact>>, <<social insight>>, <<social anxiety>>, <<social stereotypes>>. The English version of the CSBQ, validated with in the Netherlands Dutch population was translated into French and the translation was validated (double back translation). As the CBCL and CSBQ questionnaires are both dimensional instruments, dimensions have been compared. All instrument results were analysed separately; correlations and comparisons were made between groups. Correlations between CSBQ and CBCL dimensions are consistent. Positive correlations exist for: <<acting-out>> dimension with <<external behaviours>>, <<aggressive behaviour>> and <<delinquent behaviour>>; <<social contact>> with <<internal behaviours>> and <<withdrawn>>; <<social Insight>> with <<social problems>> and <<attention problems>>; <<social anxiety>> with <<anxious/depressed>>, <<thought problems>> and <<internal behaviours>>; social stereotypes>> with <<thought problems>>. Mean CSBQ results are as follows: 1. autistic group has the highest score for the <<social contact>> dimension, ADHD group has the highest score for the <<acting-out>> dimension, mental retardation group has the highest score for the <<social insight>> dimension. 2. comparisons between groups shows: significant difference between the autistic and ADHD groups for <<social contact>> and <<social anxiety>> but not for <<social insight>> and <<acting-out>>; between the autistic and mental retardation groups, there is a significant difference for <<social contact>> but not for the other dimensions; between the ADHD and mental retardation groups, there is a significant difference only for <<acting-out>>; there is no significant difference between the ADHD and emotional groups; control group has very low scores. CBCL results are: abnormal scores in all groups except normal control group, for <<social problems>> and <<attention problems>>; abnormal scores in the autistic and emotional groups for <<anxious/depressed>>, <<withdrawn>> and <<internal behaviours>>; abnormal scores in the ADHD group for <<aggressive behaviour>>, <<delinquent behaviour>> and <<external behaviours>>; the <<internal behaviours>> score is borderline. Social behaviour profiles are different and characteristic for each disorder. However, social symptoms are not specific for one disorder and common social signs do exist between different disorders. Our results are concordant with the Luteijn study and literature data. The results support the hypothesis of a dimensional pathogenesis in social behaviour disturbance. We discuss the benefit of a dimensional approach to complete the categorical one. The Children's Social Behaviour Questionnaire seems to be an interesting instrument to explore social behaviour disturbances in several child disorders.