The aim of this study was twofold: 1) to use estimates of random and systematic error to ascertain the test-retest reliability of grip strength measurements obtained with the Jamar hand dynamometer in healthy and disabled women, and 2) to determine the size of the change required to detect a genuine change in grip strength for accurate and meaningful clinical interpretation. Previous research has shown grip strength measurements obtained with a Jamar hand dynamometer from healthy and disabled subjects on different occasions to be reliable. However, the test-retest reliability has been based on correlation coefficients rather than on the actual size of the test-retest differences required to detect a genuine change in grip. The test-retest reliability of maximum grip strength measurements in 32 healthy women and painfree grip in 10 disabled women with nonspecific regional pain (NSRP) was determined. Reliability, based on estimates of systematic and random error, was high in both subject groups. There was no statistically significant systematic error between tests. Test-retest measurement error was +/-5.7 kg (12.5 lb) and +/-5.9 kg (13.0 lb) in healthy and disabled subjects, respectively, 95% of the time. In this population of healthy women and women with NSRP, any change in grip of less than 6 kg (13.2 lb) could have occurred by chance. The results of our study suggest that a change of more than 6 kg (13.2 lb) is necessary to detect a genuine change in grip strength 95% of the time.