Discriminant validity of the Inventory to Diagnose Depression

Measures of depression and anxiety correlate highly with one another. It has been hypothesized that this shared variance partly reflects poor discriminant validity, which could be improved by linking item content more closely to diagnostic criteria. We tested this speculation by comparing the Inventory to Diagnose Depression (Zimmerman, Coryell, Corenthal, & Wilson, 1986), developed to correspond with diagnostic criteria for major depression, with the Beck Depression Inventory (Beck, Rush, Shaw, & Emery, 1979) in terms of discriminant validity. These measures correlated more strongly with each other than with anxiety but did not differ in their relations with anxiety.