The Millon Clinical Multiaxial Inventory (MCMI) has become an important and commonly used instrument to assess personality functioning. Several studies report significant changes on MCMI personality disorder scales after psychological treatment. The aim of the study was to investigate whether pre-post-treatment changes in 39-session psychodynamic group psychotherapy as measured with the MCMI reflect real personality change or primarily reflect symptomatic state changes. Pre-post-treatment design included 236 psychotherapy outpatients. Personality changes were measured on the MCMI-II and symptomatic state changes on the Symptom Check List 90-R (SCL-90-R). The MCMI Schizoid, Avoidant, Self-defeating, and severe personality disorder scales revealed substantial changes, which could be predicted from changes on SCL-90-R global symptomatology (GSI) and on the SCL-90-R Depression scale. The MCMI Dependent personality score was the only MCMI personality scale showing significant change when the SCL-90-R Depression change score was included as a covariate. Splitting patients into those with and without personality disorders did not change the results. Observed changes on MCMI-II personality disorder scales in short-term psychotherapy reflect change in symptomatic state. The MCMI-II Base Rate cut-off points probably include too many patients, justifying the introduction of new scoring procedures in the MCMI-III.