This study investigates the communicative practices of youth volunteers involved in Surabaya’s Kelurahan Bersinar anti-drug initiative, highlighting the role of grassroots messaging in shaping public health discourse. Grounded in framing theory, narrative theory, and social capital theory, the research explores how volunteers construct persuasive anti-drug messages through culturally rooted storytelling, emotional framing, and symbolic communication. Using a qualitative case study approach, data were gathered through interviews, focus groups, observations, and campaign materials with 18 youth volunteers. The findings reveal that volunteers effectively used personal experiences and local narratives to foster trust and collective identity, transforming individual challenges into shared stories of hope and resistance. Emotional strategies, such as gain-framed and fear-based appeals, were deliberately crafted to trigger empathy and promote behavioral change. Volunteers also capitalized on their deep integration within community networks to enhance message credibility and reach, cultivating both bonding and bridging social capital. The study challenges top-down approaches in health campaigns, illustrating the power of peer-led, culturally resonant communication. It contributes a layered communication model for grassroots health advocacy and underscores the transformative impact of narrative-driven volunteer engagement. The study recommends integrating cultural framing in training programs and exploring digital channels to scale grassroots messaging in future research.

This study discusses the form of code mixing content creators on the TikTok application. Data in the form of words, phrases, and clauses, the data source of this study is code mixing content creators in the video account @furkysyahroni found in the TikTok application. The data collection method uses the reading method. Basic tapping techniques and advanced reading techniques free from conversational involvement . The results of the study found code mixing forms in the form of words, phrases, and clauses. The conclusion of the author's research found 21 data from the video of 1 content creator uploaded by @furkysyahroni on the TikTok application . Of these, 13 data are included in the category of code mixing forms of word insertion, 7 data are code mixing forms of phrase insertion, and 1 data is code mixing forms of clause insertion.