Background: Psychological models of conversion disorder (CD) traditionally assume that psychosocial stressors are identifiable around symptom onset. In the face of limited supportive evidence such models are being challenged. Method: Forty-three motor CD patients, 28 depression patients and 28 healthy controls were assessed using the Life Events and Difficulties Schedule in the year before symptom onset. A novel ‘escape’ rating for events was developed to test the Freudian theory that physical symptoms of CD could provide escape from stressors, a form of ‘secondary gain’. Results: CD patients had significantly more severe life events and ‘escape’ events than controls. In the month before symptom onset at least one severe event was identified in 56% of CD patients – significantly more than 21% of depression patients [odds ratio (OR) 4.63, 95% confidence interval (CI) 1.56–13.70] and healthy controls (OR 5.81, 95% CI 1.86–18.2). In the same time period 53% of CD patients had at least one ‘high escape’ event – again significantly higher than 14% in depression patients (OR 6.90, 95% CI 2.05–23.6) and 0% in healthy controls. Previous sexual abuse was more commonly reported in CD than controls, and in one third of female patients was contextually relevant to life events at symptom onset. The majority (88%) of life events of potential aetiological relevance were not identified by routine clinical assessments. Nine per cent of CD patients had no identifiable severe life events. Conclusions: Evidence was found supporting the psychological model of CD, the Freudian notion of escape and the potential aetiological relevance of childhood traumas in some patients. Uncovering stressors of potential aetiological relevance requires thorough psychosocial evaluation.