The gambling landscape is constantly evolving. Digital technologies feature several emerging forms of gambling-like activities (‘gamblification’) including free-to-play social casino games and video game loot boxes, as well as the opportunity to spectate gambling on streaming platforms such as Twitch, Kick and YouTube (Macey & Hamari, 2024). Recent research has begun to characterize links between these activities and real-money gambling. For example, young adults who purchase loot boxes show an increased likelihood of initiating gambling over the next 6 months (Brooks & Clark, 2023). We do not yet understand how early exposure to randomized rewards in these settings draws people to gambling, and the factors associated with risk and resilience to gambling harm. This study will focus on gambling streams, as a popular medium through which young people are exposed to gambling content. Watching a gambling stream does not involve betting or winning money, but the viewer observes authentic gambling, often in a highly stimulating context (e.g. high stakes bets). Streaming content is often broadcast as highlight ‘clips’: short, edited videos that can be viewed on the channel homepage and also shared across social media platforms. For gambling streams, these clips often display the outcomes of single bets showing large 'jackpot' wins, accompanied by the streamer’s intense emotional reaction to winning. It is well recognized that experiencing early big wins in the context of (firsthand) real-money gambling is associated with future gambling problems (Turner et al., 2006), and we propose that the ‘vicarious’ observation of winning may also shift gambling attitudes and intentions. The present study is an online experimental design that will present pre-selected gambling stream clips to participants assigned to two conditions. The experimental group will view a sequence of clips featuring gambling jackpot wins. The control group will view clips with realistic gambling outcomes, i.e. mostly small losses. We will develop these stimuli from a combination of authentic clips and edited Kick footage from longer streams, both as publicly accessible content. The primary dependent variables are (1) attitudes towards gambling, (2) intention to gamble on slot machines, (3) winning probability judgement. We will also measure participants’ affect ratings of the clips as a manipulation check. Trait-level scores on the Gambling-Related Cognitions Scale (GRCS) will be recorded as a possible moderator.