Purpose: This study was designed to compare a low versus high theoretical fidelity pedometer intervention applying social-cognitive theory on step counts and self-efficacy. Method: Fifty-six public university employees participated in a 10-week randomized controlled trial with 2 conditions that varied in theoretical fidelity. Participants in the high theoretical fidelity condition wore a pedometer and participated in a weekly group walk followed by a meeting to discuss cognitive-behavioral strategies targeting self-efficacy. Participants in the low theoretical fidelity condition met for a group walk and also used a pedometer as a motivational tool and to monitor steps. Step counts were assessed throughout the 10-week intervention and after a no-treatment follow-up (20 weeks and 30 weeks). Self-efficacy was measured preintervention and postintervention. Results: Participants in the high theoretical fidelity condition increased daily steps by 2,283 from preintervention to postintervention, whereas participants in the low fidelity condition demonstrated minimal change during the same time period (p = 0.002). Individuals attending at least 80% of the sessions in the high theoretical fidelity condition showed an increase of 3,217 daily steps (d = 1.03), whereas low attenders increased by 925 (d = 0.40). Attendance had minimal impact in the low theoretical fidelity condition. Follow-up data revealed that step counts were at least somewhat maintained. For self-efficacy, participants in the high, compared with those in the low, theoretical fidelity condition showed greater improvements. Conclusion: Findings highlight the importance of basing activity promotion efforts on theory. The high theoretical fidelity intervention that included cognitive-behavioral strategies targeting self-efficacy was more effective than the low theoretical fidelity intervention, especially for those with high attendance.