We addressed the following three questions: (i) are there sex differences in outcome in the functional psychoses?, (ii) what is their effect size, and which variables mediate the effect of sex on outcome?, (iii) is the effect of sex diagnosis-specific? In a prospective study of 166 patients with recent onset psychosis, we established that 4-year outcome was more favourable for women. Female patients more often had a remitting illness course (OR = 3.0; 95% CI: 1.5-5.9), were living independently 14% (4-24%) more of the time, had less evidence of negative symptoms over the follow-up period (OR = 0.3; 0.2-0.7) and were more likely to be employed at follow-up (3.6; 1.8-7.6). The findings did not appear diagnosis-specific, although the sample size was small to test for interaction with diagnostic category. Baseline occupational and social adjustment, clinical expression of illness and age and type of onset explained up to 60% of the sex effect. The processes underlying these factors mediate the effect of sex on outcome.