Two raters performed hand-held dynamometer testing of six muscle groups of 30 patients to determine the interrater reliability of the procedure. The six muscle groups tested were the shoulder external rotators, elbow flexors, wrist extensors, hip flexors, knee extensors, and ankle dorsiflexors. The patients were of eight different diagnostic groups, with most having hemiparesis secondary to cerebrovascular accidents. We used Pearson product-moment correlations and t tests to compare the two raters' score. The correlation coefficients ranged from .84 to .94 (p less than .001). Only two muscle groups (ie, shoulder external rotators and wrist extensors) had mean scores that differed (p less than .05) between raters. Although the interrater reliability of the procedure was found to be good to high in the six muscle groups tested, the results of the t tests indicate that hand-held dynamometry should undergo further evaluation.