(PS11-24) Decoding Tiktok: The Interplay of Suicide and Self-injury Themes

In the US, about 1 in 5 adolescents and young adults (AYAs) consider suicide, and 1 in 11 attempt it. Many AYAs with suicidal thoughts don't seek help and those who do often turn to informal support such as online social networks. This presents an urgent need to understand how individuals communicate about suicide and self-injury online, specifically via the fastest-growing social media app used predominantly by AYAs, TikTok. The current study aims to investigate how users find and engage with suicide and self-injury-related content on TikTok. First, the top three hashtags associated with suicide and self-injury content on the platform will be identified. This will involve a comprehensive review of existing literature and exploration of "related tags" within the TikTok app. From the compiled list of tags, priority will be given to those with the highest volume of associated videos. Acknowledging the tendency of users to circumvent TikTok's censorship and algorithms, known as "algospeak," this analysis will also consider and incorporate relevant modified tags, such as variations like "su1cide" alongside conventional tags like "suicide."  

Videos using the top three hashtags will be downloaded with the TikTok Research API which provides access to data that is otherwise hidden from users, such as the total number of comments, voice-to-text, and time stamps. Two doctoral students will code the videos and associated comment sections using a comprehensive qualitative coding scheme. Drawing on existing literature, the developed coding scheme includes aspects of both harmful and protective messaging and overall theme of both the video and comment section (i.e., suicidality vs. self-injury). Additionally, information regarding user responses to content will be collected by downloading and recording the number of likes, comments, shares, and saves for each video. Collecting these data points allows for identification of themes that have higher engagement, resonance, and dissemination among users. Inter-rater reliability will be established via joint coding of a random sample of 50 videos and discussion of qualitative codes until consensus has been reached. 

It is hypothesized that suicidality and self-injury will be intertwined themes on TikTok due to users’ interrelated conceptualization of these topics, as compared to the separation imposed by the academic literature. If supported, findings would suggest AYAs may not differentiate between these constructs as researchers do, potentially exposing them to more nuanced and harmful content. Exploratory analyses will investigate which themes are interacted with most frequently on TikTok.  

The current study presents important implications for how AYAs discuss and engage with suicide and self-injury on TikTok and will inform the growing literature surrounding digital communication on these topics.