Recent research on attempts to motivate large-scale safety belt use has documented a number of shortcoming, including limited long-term evaluation data, excessive costs, short-lived intervention effects, and program delivery by outside agencies rather than indigenous personnel. The present study attempted to overcome these disadvantages. Specifically, the "Buckle-up for Bucks" safety belt promotion campaign conducted at a community hospital incorporated the following: a) indigenous hospital staff as program sponsors, delivery agents, and co-coordinators; b) a year-long program evaluation; and c) a combination incentive and commitment-based intervention program. Directed and coordinated through the Office of Community Relations, the hospital-based intervention included awareness sessions, randomly determined five-dollar a week cash incentives, and a commitment-based pledge card strategy. To be eligible to win the incentives, the staff members met the following contingencies: a) wore a safety belt; b) signed a pledge card; c) displayed the signed pledge card on their dashboard; and d) pledged for a duration that ensured eligibility. The evaluation data were collected for four phases: baseline, intervention, withdrawal, and a long-term, follow-up. For the overall sample, usage increased from a baseline mean of 15.6% to 34.7% during the intervention, decreased to 25.6 at withdrawal, and increased to a long-term follow-up mean of 28.6%. For the Pledge card signers and the Non-singers, usage increased from baseline means of 29.4% and 11.8% to intervention usage rates of 75.1% and 17.7%, respectively, demonstrating that the intervention had a differential effect on the signers and non-signers. Withdrawal and Follow-up usage rates were 56.0% and 44.9% for the Pledge group, and 17.2% and 22.1% for the Non-pledge group. A chi-square test for white noise indicated the data were autocorrelated. A time-series analysis was conducted to remove the serial dependency. Statistical significance of the intervention was examined from the time-series perspective and traditional analysis of variance procedures. Differences between approaches are addressed and theoretical explanations for the intervention effects are considered. Finally, suggestions for future research are offered. / M.S.