Significance: With advances in technology manufacturers have introduced voice-announcement (VA, announces the number of steps registered) in various brands of pedometers. These pedometers can prove useful in promoting activity for youth with visual impairments through the ability to notify the wearer of his/her registered steps. However, prior research (e.g., Beets et al., 2005; Crouter et al., 2005) has found pedometer step count accuracy (i.e., actual steps taken versus pedometer recorded steps) varies among pedometer brands. As a result systematic testing is necessary to determine which pedometer recommend for use in research and health-promotion settings. Design: Youth with visual impairments (N = 35, 13.5±2.1yrs, 13 girls) walked three 100-meter distances while wearing two units (right and left placement) of three brands of VA pedometers (Centrios Talking Pedometer, TALKiNG Pedometer, and Sportline Talking Calorie Pedometer 343). Registered pedometer steps for each trial were recorded and compared to actual steps assessed via digital video. Intraclass correlation coefficients (ICC) were calculated to determine inter-unit agreement. Percent error scores (PES) and absolute percent error scores (APE, error score without its sign) were calculated ([registered steps minus actual steps] x 100) for each brand. The PES scores were averaged and plotted for each participant to determine the direction of error. The minimum criterion for inter-unit agreement was set at .80 and the maximum threshold for APE scores set at 3%. Results: Inter-unit agreement ICCs (95CI) were: Centrios .76 (.56-.87); TALKiNG .37 (.05-.63); and Sportline .59 (.32-.77). The APE scores for right and left placements, respectively, were: Centrios 3.3±4.4% and 5.6±8.3%; TALKiNG 2.3±2.2% and 12.2±14.9%; and Sportline 2.1±2.9% and 8.6±11.2%. The PES plots indicated the left placed Centrios (2.0±9.9%) and Sportline (5.0±13.3%) systematically overestimated steps, while the left placed TALKiNG systematically underestimated steps (-11.7±15.4%). Conclusion: No brand reached the minimum criteria for agreement between right and left units, whereas the right placed units for each brand met the criteria for steps registered compared to actual steps taken. A systematic error was observed for all left placed units that cannot be attributable to manufacturer quality (Bassett et al., 1996; Beets et al., 2005). The authors speculate that variables not assed in the current study, such as gait patterns, may be attributable to the systematic error for the left placed units. This suggests right placement of VA pedometers for youth with visual impairments is required for a more accurate count of steps. Keyword(s): adapted physical activity, measurement/evaluation