We used an RCT design to evaluate an online training program for over 200 educators implementing the Reducing the Risk program. Our primary outcome of interest was implementation fidelity. We used several methods for measuring implementation fidelity: educator implementation logs for each lesson, interviews with educators after implementing specific lessons, in-person observations, audio observations, follow-up interviews, and post surveys. We are analyzing and comparing all data sources to determine how valid the educator self-report logs are, which method(s) appears to provide the most valid data source, how validity appears to vary across the data sources, and what information each data source adds to the measurement of implementation fidelity. The results will provide important information for improving the quality of the various types of measures of implementation fidelity and recommendations for which sources yield the best quality data given various resource constraints. #RandomizedControlTrial #Implementation_Fidelity #HealthEvaluation #2011Conference