Community problem solving is believed to help coalitions achieve community changes and subsequent population-level reductions in targeted community health problems. This study empirically examined a community problem solving model used by CADCA, a national coalition training organization, to determine if the model explains how coalitions become effective community change agents and the extent to which it fits for different types of coalitions. The sample consisted of 551 substance abuse coalitions, and SEM was used to test the research questions. Results suggest the overall conceptual model fits the data and provides an adequate representation of the community problem solving process for coalitions. Additionally multi-group modeling indicates the model fits for a variety of coalitions. CADCA's model appears to have wide utility and applicability and appears to have honed in on critical elements of community problem solving that may increase the likelihood of coalition success at reducing substance abuse within their community. (Contains 4 figures and 3 tables.)