The St. Hans Rating Scale (SHRS) is a multidimensional rating scale for the evaluation of neuroleptic-induced hyperkinesia, parkinsonism, akathisia and dystonia. This scale and the Abnormal Involuntary Movement Scale (AIMS) were tested by 7 raters (2 experienced, 2 less experienced and 3 totally inexperienced) in 30 psychiatric patients with tardive dyskinesia (TD). The test was performed 3 times in the same patients: 1) live evaluation during a video recording, 2) evaluation 2 weeks later from the videotape, and 3) evaluation after another 2 weeks from the same videotape. The intrarater reliability was high in the experienced group (0.91-0.96 for SHRS hyperkinesia scale, 0.80-0.84 for AIMS, and 0.82-0.97 for SHRS total parkinsonism). No significant changes occurred from live to video evaluation. The interrater reliability coefficient for the experienced group was also high: 0.89-0.95 for the SHRS hyperkinesia scale, 0.76-0.85 for the AIMS scale and 0.95-0.98 for the SHRS parkinsonism scale. The less experienced and the inexperienced raters had coefficients for intra- and interrater reliability that were 0.10 and 0.20 lower, respectively. The SHRS parkinsonism scale had a high construct validity, as determined by the homogeneity coefficients of Cronbach (0.82) and Loevinger (0.43). The corresponding coefficients for the hyperkinesia scales were low, in agreement with the individual distribution of TD (only about 50% present extremity dyskinesia and less than 25% facial, head and trunk dyskinesia, independent of the severity of the syndrome). Finally, convergent validity was found between the SHRS hyperkinesia scale and AIMS and divergent validity between all of the other scales.(ABSTRACT TRUNCATED AT 250 WORDS)