a1 From the Department of Sociology, Bedford College, University of London
a2 Institute of Psychiatry, University of London
The paper focuses on recent criticisms of the study of the role of life-events in the onset of psychiatric conditions and suggests that measurement error and bias can be reasonably well controlled by various methodological procedures. Failure to comply with these may be expected, however, to increase rather than decrease the chances of establishing a ‘positive’ resuit. Three further factors to do with the design of studies and the analysis of data are discussed which are likely to mask real differences between patient and comparison group, and which therefore might explain the ‘negative’ results reported in the literature. They concern: (1) the choice of an appropriate comparison group; (2) specification of the length of the period between event and onset; and (3) specification of the event in terms of some measure of severity. Results from two London studies of schizophrenic and depressive patients are presented to illustrate the argument. The studies suggest that life-events do play an important causal role in bringing about both disorders.
1 The work on schizophrenia and depression has been supported by the Foundation Fund for Research in Psychiatry, the Medical Research Council, and the Social Science Research Council.