Community-based action research (CBAR) encourages students to view internships and other civic-engagement opportunities as the first steps in collaborative projects that build long-term relationships with community partners and model how democracy works. We marshal longitudinal data, qualitative and quantitative, on both attitudes and behaviors to demonstrate meaningful impacts of community-based work by university students. Our approach to this difficult question involves the use of a treatment group contrasted with two distinct control groups: one situated somewhat closely to the experiences of the treatment group and one more distantly. Despite the difficulty of developing metrics of attitude and behavior change over relatively short periods of time, we achieve leverage on the question of how project pedagogies and experiential learning shape civic attitude formation and associated changed behaviors.