Operating characteristics describe the validity of tests that attempt
to dichotomously predict a diagnosis. These statistics are not fully p
ublished in the Millon Clinical Multiaxial Inventory-III Manual (MCMI-
III manual; Millon, 1994). When calculated from available statistics,
the positive predictive powers of the MCMI-III scales are poor both in
absolute terms and relative to the MCMI-II (Millon, 1987). There were
a number of problems, however, with the initial MCMI-III validity stu
dy both inherently and in execution. Although it is doubtful that the
MCMI-III is weaker than the MCMI-II, a new validity study is needed.