A Study of Depression in Women: A Reply to Keith Hope's Critical Note

Abstract

We welcome Dr. Hope's detailed comments on our paper, 'Social Class and psychiatric disturbance among women in an urban population* in Sociology , May 1975, and here take the opportunity of clarifying briefly, and in so far as we can with our present data, the questions he has raised. We shall deal with each of Dr. Hope's numbered points in turn: the first deal with the technical points concerning sampling etc. and the later ones with more fundamental issues. i(i) and (ii). We would not claim that our sampling procedure gives a completely representative sample. The register used as a sampling frame was that drawn up by the local authority for rating purposes. In about 7 per cent of instances the rateable unit was found to contain more than one household (defined by communal eating arrangements). In these cases a household was chosen at random. In 33 (15 per cent) of the 220 selected households there was more than one woman aged between 18 and 65 (the target population). In these cases a woman was selected at random from those eligible. There are two sorts of bias here: (a) women from households at multiple-household addresses (rateable units) would be expected to be underrepresented and those at one-household addresses over-represented, and (b) the procedure would also be expected to result in the under-representation of women in larger-than-average households since the sample was based on a frame unstratified by the number of target subjects in each sampling unit (household size here refers to the number of eligible women in the household). At this stage it is difficult to correct our overall estimates of event and disturbance rates etc. for these biases. These are in any case less critical than the causal model we have outlined and this would suffer only if the sample's shortcomings have resulted in spurious associations or in obscuring real links between the variables examined. We have no reason to believe that this has happened. The sample is too small to allow the sort of detailed examination of household composition Dr. Hope suggests in this regard, but we would point out that in our sample the distribution of the occupations of any male 'head of household' was very close to that provided by the 1961 census for men in Camberwell (see fn. 2, Brown et al., 1975). The defects of the sampling procedure are indeed unfortunate but, in the absence of a complete frame, could be avoided only by the use of more cumbersome methods having their own specific, and equally unwelcome, sources of bias. i (iii). The figure of 18 per cent in owner-occupied houses compares with 19 per cent for the Inner London boroughs, 36 per cent for the Greater London district as a whole, and 50 per cent for England and Wales (G.R.O., 1961). There is no reason to believe that the middleclass group in the old borough of Camberwell, who are concentrated largely around Dulwich, is in any way atypical of middle-class women in South London, although it might well be possible to find districts in London where middle-class women show higher rates of psychiatric disturbance (e.g. areas with heavy concentration of rented accommodation housing separated or divorced women). Their relatively low frequency of disturbance must not be interpreted as representing the frequency of 'psychiatric symptoms' among middle-class women. The 'cases' were, as is pointed out on p. 228 and fn. 3, the more severely disturbed among those having