Aim The present study compared the performance of the Jamar and Smedley dynamometers for the measurement of grip strength, and the two dynamometers were used to determine the prevalence rate of weakness in a large, community‐dwelling, older adult population, based on different diagnostic criteria. Methods A total of 467 community‐dwelling older adults aged 69–89 years were recruited into the study. Grip strength was measured with two dynamometers; the Jamar hydraulic hand dynamometer was used as the reference device. To assess agreement between grip strength measurements by the two dynamometers, linear regression and Bland–Altman analyses were carried out. Results In a comparison of the dynamometers, high coefficients of determination (R2) were obtained for grip strength (R2 = 0.80, standard error of the estimate 2.68 kg in male participants, and R2 = 0.75, standard error of the estimate = 1.95 kg in female participants). There was systematic bias with underestimation of grip strength (bias 3.09 kg, 95% confidence interval 2.77 to 3.41 for men; bias 2.60 kg, 95% confidence interval 2.31 to 2.89 for women) by the Smedley dynamometer. In both sexes, no proportional bias was noted between the two dynamometers for measurement of grip strength (r = 0.030, P = 0.611 for male participants; r = −0.033, P = 0.653 for female participants). The prevalence of weakness ranged from 1.8 to 19.9% in male participants, and 0.5 to 27.4% in female participants. Conclusions The present study shows that despite the excellent correlation between the two dynamometers, there is a statistically significant difference in grip strength measurements. The Smedley dynamometer showed a higher prevalence of weakness than did the Jamar dynamometer. Geriatr Gerontol Int 2017; ••: ••–••