The methodologically oriented reader who takes the time to read the initial report by Dr Dohrenwend and his colleagues1 and our response2 will note that the preceding letter does not even seek to refute our central point. Instead, Dohrenwend et al focused on the validity of the computerized program for psychiatric diagnosis (DIAGNO), which does not even use the same set of summary scales that they criticized.
Our central point was that high reliability, although always desirable, is not an absolute requisite for case finding. That is because the task of identifying deviant "cases" is quite different from the task of distinguishing degrees of psychopathology among noncases.
We do not dispute the proposition that at the current time there are other procedures that are more efficient for finding cases or making diagnoses in the general population. What we objected to (without refutation) is the standards they applied, which