The reported prevalence of absent ankle jerks in elderly people varies greatly. This variation may be due to differences in the method of testing. Eight physicians examined 12 patients for the presence of ankle jerks using two techniques: plantar strike and tendon strike. Both intra-observer agreement (kappa 0.47 vs 0.20; p = 0.01) and inter-observer agreement (0.57 vs 0.21; p < 0.001) were greater with plantar strike. Reliability of ankle jerk assessment was greater for more experienced examiners. Differences in technique may explain some of the discrepancy between studies examining the prevalence of absent ankle jerks in elderly people.