Drug-induced dyskinesias (DID) manifested as hyperkinetic and/or dystonic movements or postures are common problems in Parkinson's disease (PD). Novel therapeutic interventions may offer possibilities to counteract these common adverse effects of an otherwise necessary treatment. To be able to evaluate the effects of such interventions on DID, reliable and relevant clinical assessment tools are needed. We tested the inter- and intrarater reliability of a new clinical dyskinesia rating scale consisting of separate ratings of different body parts, including lateralization and separate ratings of dystonia and hyperkinesias. Interrater reliability was tested both with and without a defined scoring code and clarification of the dystonia section. The nondefined version was also tested for intrarater reliability. Thirteen raters independently reviewed 23 videotape sequences showing PD patients performing standardized motor tests. Inter- and intrarater agreement was significant in all evaluations, and no differences were detected when comparing ratings performed with the defined and nondefined version of the scale. The rationale for, and the role and use of, the present scale are addressed.