We designed this study to replicate previous findings which suggest a relationship between the qualities found in family environments and specific eating disorders through the examination of subjects' reports of family environments. We also attempted to refine previous findings of such relationships by adding an additional psychiatric contrast group. Subjects included patients with the following diagnoses: (1) Anorexia nervosa, restrictor type; (2) anorexia nervosa with bulimic features; (3) normal weight bulimia; and, as a psychiatric control population (4) major depression. In contrast to previous findings, subjects in each diagnostic group do not differ statistically either from each other or from a normative population with regard to reported family environments. However, when we grouped subjects by self-reported level of depression, those with a high level of depression described family environments that are significantly different, independent of the severity of illnesses or diagnoses. We discuss the use of self-report instruments in measuring such complex phenomena as family environment.