This retrospective study was undertaken to determine the test-retest reliability of hand-held dynamometer strength testing for 18 extremity muscle groups. To assess reliability, I compared three dynamometer strength scores recorded for each muscle group on the strength score forms of 30 neurologically involved patients, using the Pearson product-moment correlation and a one-way analysis of variance (ANOVA) for repeated measures. All correlations were significant (p less than .01); the median and modal correlations were all .97 or .98. The ANOVA revealed significant differences in repeated dynamometer test scores only for hip and shoulder abduction. The results suggest that hand-held dynamometry can be a reliable assessment technique when practiced by a single experienced tester.