EDITOR—Shaw et al add welcome breadth to the literature concerning psychiatric disorders in African-Caribbeans.1 However, there are methodological problems which make interpretation of the results difficult.

Contrary to BMJ guidelines, there is no rationale for the choice of ethnic variables used.2 Hypotheses for why depression in the inner city should be expected to be related to Office for National Statistics' definitions of ethnicity may have led to the measurement of possible explanations or confounders such as financial worry, religion, housing, racism, …