So far we have dealt with depression in general terms: we now consider our own investigation. Our approach developed from the hypothesis that clinical depression is an understandable response to adversity, and it was this we wished to investigate. We could imagine practically anyone developing depression given a certain set of environmental circumstances. We do not view all psychiatric disorder in these terms and would certainly see it as an inappropriate perspective for schizophrenia. We therefore looked at clinical depression in terms of rates of disorder in a population and sought to explain differences through the everyday lives of particular individuals. It was only by trying to link rates of disorder between social class groups, and similar broad social categories, to detailed knowledge about the lives of particular individuals that we felt at all confident about reaching an understanding of the processes underlying differences in rates. It was in this way that we attempted to bring the survey and clinical traditions closer together. We began by developing a causal model of depression based on the day-to-day experience of each woman. While we did not entirely ignore the woman's past, we concentrated on her recent experience because we considered it had been unwarrantedly neglected elsewhere (perhaps as a result of psychoanalytic influences) and because it was obviously easier to collect accurate information about the present. Only when this model had been established did we attempt to use it to arrive at some understanding of rates of depression in the general population.