This exploratory study examined the relationship between patient and staff attributions for psychiatric rehospitalization. The sample consisted of 120 recently rehospitalized schizophrenics and 168 psychiatric staff who in structured interviews were asked to state their reasons for why rehospitalization occurs. As a possible secondary source of data, subjects completed instruments adapted from Russell's (1982) Causal Dimension Scale. Demographic information was collected in order to study the relationship between patient and staff characteristics and the attributional dimensions.

Several approaches were used in the analysis of data. In the principle method comparing patient and staff responses coded in a standard approach (Elig and Frieze, 1975), significant differences between patient and staff attributions were found on the stability and controllability dimensions. Patient and staff internal attributions did not differ significantly. The slightly stable attributions of patients were significantly different (p = .03) from the staff's slightly unstable attributions. Patients' uncontrollable attributions significantly differed (p = .00) from the controllable staff reasons. Data analysis using the Patient and Helper Attribution Scales indicated that patient and staff responses were significantly different on the locus of causality dimension, although confidence in these results was weakened by the poor reliability of these instruments with study subjects. Comparisons of patient and staff responses coded into content categories also yielded differences between the two groups. While the majority of subjects cited factors within the person as the chief cause for readmission, staff primarily attributed rehospitalization to non-compliance, while patients named a variety of factors but most commonly identified chronic psychological factors. Finally, multiple regression analysis indicated that for patients and staff, a few variables predicted attributional dimensions but differences between group mean scores were not significant using Scheffe tests. Staff and patient variables made only a small contribution to the variance in attribution dimension scores.

Although no direct clinical applications can be drawn from the study, the results support findings in other studies of differences between helpers and helpee. The possible affective and behavioral consequences of patient and staff rehospitalization attributions, their relevance for nursing and implications for further research are discussed.