Some Problems of Design and Inference in Studies of Community Tenure

Three important problems for research on community tenure among former mental patients are identified: 1) the use of a “significant other” as informant, 2) the use of “success-failure” as a major dependent variable, and 3) the use of measures of psychiatric symptoms to the exclusion of social dynamics surrounding hospitalization. These are discussed as exemplified by the methodology and findings of three major empirical works of this genre. The author concludes that there is no single community study sampling both sexes in a wide range of living settings from which valid inferences can be drawn and urges the recognition of the consequences of drawing policy inferences from research conducted within the design limitations discussed.