The epidemiological data examined in Chapter 2 revealed a gendered landscape of identified mental disorder with the gender differences varying according to diagnosis, age, social class, marital status and ethnicity as well as historical period. However, in the light of the examination of constructs of mental disorder in the last chapter, we need to consider precisely what weight can be attached to the data. Certainly they show some marked associations, such as those between gender and diagnosis. But do they reveal any more than this? Can we make any inferences about gender differences in the distribution of medically defined mental disorder from them? Can we say that overall more women than men are disordered in mind as the patient statistics might seem to imply. How does the fact that psychiatric categories vary across time and place affect the conclusions we can draw from epidemiological studies? In this chapter I argue that the social and cultural specificity of medical categories of mental disorder, which are embedded in the epidemiological data, severely restricts the inferences we can make. Even associations supported by a range of empirical data are construct specific, and claims about the social distribution of disorder, even when they recognise that specificity, have to be treated with great caution because of the limitations of the data on which they are based.