Background: Pedometers have been identified as a potential motivational aid for increasing physical activity, but their efficacy has not been demonstrated in a community-based, nonclinical sample. Design: A randomized controlled trial was conducted from August to December 2005. Analysis was completed in June 2006. Setting/Participants: Inactive adults aged 30-65 years (n=369) recruited from the community. Intervention: Comparison of a theoretically based self-help walking program (WP) and weekly diaries (sent by mail); the same walking program with a pedometer (WPP) (also by mail); and a no-treatment control group (C). Measures: Change in self-reported leisure time in any sports/recreation in the last 3 months, and all-purpose walking (APW) for exercise, recreation, and travel, and other moderate, vigorous physical activity in the last week. Proportions meeting physical activity recommendations (equal to or greater than 150 minutes and equal to or greater than five sessions/week-1) were determined. Results: A 3-month follow-up interview was conducted with 314 (85%) participants. Intention-to-treat analyses indicated significance within-group increases of APW and leisure-time walking (LTW), but mean and median sessions and minutes changes were greatest in the WPP group. There were no significant between-group differences in regular LTW (walked equal to or greater than 5 sessions/week-1 for at least 30 minutes/session), but the WPP group increased significantly participation in other sports/recreations and was more likely than the control group to meet physical activity recommendations by all leisure-time physical activity (adjusted odds ratio=2.40, 95% CI=1.17-4.93), by APW (adjusted odds ratio=1.75 95% CI=0.92-3.34) and all physical activity (adjusted odds ratio=1.59 95% CI=0.92-2.79) in the last week. Conclusions: Pedometers enhanced the effects of the self-help walking program. This low-cost intervention should be tested for sustainability.