Project Abstract Exposure to ultraviolet (UV) light between the ages of 15-20 years is the most important etiological factor in skin cancer, yet adolescents and young adults (AYAs) this age are more likely to engage in health-compromising behaviors like indoor and outdoor tanning without skin protection. As an intervention modality, social media (SM) represents an opportunity to reach AYAs, who are among the most active Facebook, Instagram, and Twitter users. Recent research suggests SM may also be a rich hub for proliferation of skin cancer misinformation; yet the characteristics and diffusion patterns of this misinformation and risk/prevention communication across platforms are unknown. Characterizing the skin cancer communication landscape and creating effective risk/prevention posts, with an “understanding of which messages will resonate with specific groups” (The Surgeon General’s Call to Action to Prevent Skin Cancer), could enable targeted prevention methods for AYAs. Our long-term goal is to reduce health-compromising behaviors (e.g., indoor tanning, sunscreen nonuse) among at-risk AYAs, who are vulnerable to developing skin cancer. This proposal’s main objective is to: 1) characterize the SM landscape regarding skin cancer-related posts; and 2) develop/test messages for skin cancer prevention among AYAs that are clear, specific, consistent, and scientifically up to date. With a robust multidisciplinary team, we will accomplish this objective via three specific aims—AIM 1: Characterize skin cancer-related communication across three popular SM platforms. AIM 2: Build a predictive, explainable health communication model to determine the diffusion potential of skin cancer-related messages. AIM 3: Develop/pilot test sun-protection and indoor tanning-related messages for AYAs for future implementation and evaluation. In AIM 1 we will use content analyses to assess the skin cancer communication/misinformation landscapes on Facebook, Instagram, and Twitter, describing message features, source characteristics, posters/users, and social networks, and identifying network characteristics and diffusion patterns of skin cancer misinformation (exploratory). In AIM 2 we will apply machine learning methods to characterize message features; develop/evaluate a predictive model of a message diffusion potential using large-scale training data; apply the model to predict a set of online diffusion metrics for a given message; and assess its ability to reach skin cancer prevention-relevant populations. In AIM 3 we will engage two stakeholder segments in iterative rounds of message development/testing: 1) cancer organization staff (who post on their SM accounts); and 2) intended recipients of these messages: five sub-groups of AYAs aged 15-20 years—White boys/men, White girls/women, and White, Black, and Hispanic gay and bisexual boys/men. The research is innovative because of its focus on posts across three SM platforms and the Multilevel Model of Meme Diffusio...