Purpose: The purpose of this study was twofold: (1) to determine if the Digi-Walker SW-200 (SW-200), New Lifestyles NL-2000 (NL-2000), and Omron HJ-303 (HJ-303) yield similar daily step counts compared to the StepWatch-3; and (2) to determine if pedometer error is influenced by adiposity and/or stepping rate in African American women. Methods: 60 participants (28.0 ± 9.8 y) wore the devices for three weekdays. ANOVAs were performed to determine if body mass index (BMI) and device were related to steps per day, and to determine if BMI and device were related to error. Stepwise linear regressions were performed to determine which variables contributed to pedometer error. Results: StepWatch-3 counted significantly more steps than all other devices within each BMI category (p < .01). The NL-2000 had significantly less error in the normal (−13.4%) and overweight (−14.9%) groups compared to the SW-200 (−26.2% and −33.3%) and HJ-303 (−32.5% ad −31.5%) (p < .05). The SW-200 had significantly more error in the obese group (−50.7%) compared to the NL-2000 (−17.1%) and HJ-303 (−26.0%) (p < .05). NL-2000 error was not related to any variables while the SW-200 error was related to waist circumference (WC) and the HJ-303 error was related to percentage of slow steps. Conclusion: In African American women adiposity is more strongly related to more pedometer error in a device using a spring-levered mechanism (SW-200). Accumulating steps at a slow rate is related to more pedometer error when using a device with a step filter (HJ-303).