Get access

Convergent validity of MCMI-III clinical syndrome scales


Correspondence should be addressed to Morten Hesse, Artillerivej 90, 2nd, 2300 Copenhagen S, Copenhagen, Denmark (e-mail:


Objectives. This study tested the convergent validity of the Millon Clinical Multiaxial Inventory-III (MCMI-III) clinical syndrome scales.

Design. Cross-sectional survey.

Methods. Using a sample of 186 substance abusers from one single town referred for assessment, convergent and discriminant validity of the MCMI-III and Mini International Neuropsychiatric Interview (MINI) diagnoses was conducted. Additional measures included the Montgomery-Åsberg Depression Rating Scale and the Beck Anxiety Inventory.

Results. A single Axis I factor based on the raw scores correlated adequately with the factor based on the other scales (r= .85), whereas the correlation between the factor based on the MCMI-III baserate scores was somewhat lower (r= .74), but still indicated substantial convergent validity. For individual disorders, area under the curve (AUC) analyses suggested that the convergent validity of the MCMI-III and the MINI was adequate. The raw score scales were superior to the baserate adjusted scores in all but one case. Discriminant validity was good for alcohol and drug dependence, moderate for major depression and delusion, and poor for thought disorder and anxiety.

Conclusions. The MCMI-III clinical syndrome scales generally measure the constructs they were intended for. The data did not support that the adjustments used in calculating the baserate scores improved validity.