Measures of intervention fidelity can be used to identify specific intervention components promoting desired change - the active ingredients - yet such applications are rare. In the context of a social-emotional learning intervention, we illustrate how fidelity measures can be used to identify program active ingredients. We applied one customary and two novel approaches to creating indices of fidelity. In the customary approach, we averaged fidelity ratings across all core components. In the novel approaches, we computed fidelity indices for specific components by (a) averaging responses from like-items and (b) deriving factor scores from a multitrait, multimethod factor analysis. We then tested indices in relation to achievement gains (N = 1442). Indices derived using novel approaches explained more outcome variance than indices from the customary approach. Further, novel approaches revealed one component as a potential active ingredient. Discussion highlights strengths and limitations of the indices and implications for identifying program active ingredients.