Social inequalities in mental disorders have been reported. Studies indicate that the lower the social position, the higher the prevalence of mental disorders. However, these inequalities have not always been observed and results may vary according to the examined indicator of social position and mental health outcomes. The objective of this study was to examine the link between social position (educational level and occupation) and mental disorders (psychological distress measured using the GHQ-28 scale and its four sub-dimensions, i.e. depressive, anxiety, and somatic symptoms and social dysfunction, psychotropic drug use, alcohol abuse, and tobacco status) in a random sample of 6,056 male and female employees in the Centre, Pays-de-Loire, and Poitou-Charentes regions of west-central France. A high prevalence of depressive symptoms and psychotropic drug use was found among low-educated employees and blue-collar workers. However, these results did not persist after adjustment for age and classical risk factors for mental disorders (marital status, social support, life events and family history of depression). Strong social gradients were observed for smoking; these social differences persisted after adjustment. Inverse social gradients were observed and persisted after adjustment: more educated employees and professionals/managers were found to be at higher risk of psychological distress, especially anxiety and somatic symptoms and social dysfunction (particularly men). This study suggests that classical risk factors for mental health outcomes may be socially graded and may account at least in part for the social inequalities observed in depressive symptoms and psychotropic drug use.