Selection bias in prevalence data is difficult to rule out

EDITOR—To sociologists, Boydell et al's findings are counterintuitive.1 One would expect economic deprivation (at neighbourhood level) to be a decisive factor for an increased incidence of mental illness. But it is surprising to learn that the lower the proportion of non-white ethnic minorities in a local area the higher the incidence of schizophrenia in those minorities (controlled for economic deprivation).

As an explanatory hypothesis the authors point to overt discrimination and institutionalised racism as sources of stress, which can be alleviated by people making use of social capital within the ethnic group. This hypothesis surely necessitates further testing and debate. It is a pity that non-white ethnic minority groups had to be considered …