Diffusion can contribute to the spread of preventive intervention effects from participants to non-participants, but best practices for randomized trials prevent contamination of conditions. These practices conflict with cultural values of community benefit, which are salient among American Indians. This study embedded social network measures within a randomized trial of the Bii-Zin-Da-De-Dah (BZDDD) family-focused prevention program to characterize youth's social networks, describe the nature and content of sharing, and test for diffusion effects on cultural engagement (ethnic identification, cultural socialization, cultural practices) and substance use. Participants were 256 American Indian youths enrolled in the trial who provided self-reports of their social networks and indicated whether specific program content was shared with or received from others, while completing cultural engagement and substance use questionnaires across three waves. Results indicated that social networks were comprised mainly of peers and same-age family members (e.g., cousins). Program sharing was not uncommon. For example, 51% of responding intervention youth reported talking with non-participants about BZDDD at wave 2, typically (53%) with similar-age friends and family who were, most often (71%), out of the home. Evidence for diffusion effects was limited, but did indicate that control youth who had some exposure to BZDDD had a significantly higher average cultural/ethnic identity scale score at wave 2 and were more likely to ask an elder for advice than control youth who had no BZDDD exposure in adjusted analyses. Findings illustrate the value of measuring and testing for potential effects of diffusion in prevention trials with American Indians.