The assumed relationship between role performance and expectations is tested for discharged female psychiatric patients. A positive relationship is found between both sets of expectations and all areas of posthospital performance. Even though the relationship between level of expectations and performance level is affected by psychiatric and sociological variables, the significant relationship between expectations and performance still remains when the intervening variables are controlled. A central issue of interpretation is raised in the form of a “corrosion” hypothesis: the role theory assumption that performance is a function of expectations may be only partially explanatory; the opposite possibility that expectations are adjusted to suit the performance level of the patient must also be taken into account.