Manufacturers of pedometers recommend wearing a pedometer on the midline of the right thigh and this recommendation is used in research. However, there is conflicting evidence regarding the effect of pedometer position on accuracy. The purpose of this study was to systematically evaluate the effect of pedometer position on accuracy for 3 modes of gait. The Yamax SW200 pedometer was evaluated in 20 subjects in 5 different positions simultaneously: left mid-axillary, left mid-thigh, umbilical, right mid-thigh, and right mid-axillary. Each subject was asked to walk on a treadmill (54, 80, and 107 m.min(-1)), to walk overground (slow, preferred, and fast speeds), and to ascend and descend stairs. Pedometer steps were recorded and compared with actual steps observed. Using the same protocol, instantaneous triaxial accelerometry was used to explain position-dependent differences in accuracy. At slow speeds, the left mid-axillary position demonstrated the lowest net mean (+/-SD) error across all modes of gait (7.7% (+/-11.6%) overground;-0.6% (+/-2.2%) stairs). The right mid-thigh position had significantly (p<0.01) higher error for treadmill (18.1% (+/-17.3%)), overground (12.9% (+/-15.2%)), and stairs (2.9% (+/-3.9%)). Pedometer position dependent error was demonstrated, with the left mid-axillary position superior to the recommended position of right mid-thigh. The greater accuracy on the left side was wholly explained by gait asymmetry evident in step-induced accelerations recorded at right and left pedometer positions. A model of absolute error based on the ratio of steps at different gait speeds was generated to demonstrate the importance of this finding in pedometer-based lifestyle intervention studies.