(PS8-25) Bystander Intervention to Life-threatening Posts on a Social Media Platform: Exploring Bystander Threat Assessment and Reported Willingness to Intervene

Abstract Text: Over the past decade, the United States has witnessed a surge in social media use. While designed to connect users through a virtual platform, these platforms have become a common place for individuals to share intimate personal details, which includes posting life-threatening thoughts and urges (Chan et al., 2017; Peterson et al., 2023). In social media, these posts are shared to an audience of followers who subsequently become bystanders to an individual threatening to lethally injure oneself or others. As a bystander, one may have to decide whether to intervene (e.g., reporting the post, direct messaging the author, leaving a supportive comment) with the author of the post. Implications of not intervening with a direct threat towards oneself or others may include the death of the poster, or other innocent individuals. This is supported by previous research indicating expressions of distress or help-seeking behavior posted on social media are positively related with suicide-related behavior (Chan et al., 2017; Sueki, 2015). Additionally, mass threats towards others (e.g., mass shootings, school shootings) are often leaked on social media prior to the perpetration of violence (Muschert & Sumiala, 2012; Peterson et al., 2023). To date, there is little known literature exploring such bystander intervention. As such, the current study sought to evaluate bystanders’ reported willingness to intervene with life-threatening posts on a social media platform, and factors which may implicate this willingness. The participants of the study (N = 722) were from an undergraduate pool of students at a university in the United States, and mostly identified as female (83.4%) and White (50.6%). Participants were shown a series of 72 vignettes which mocked individual social media posts found on X/Twitter, followed by a series of questions to assess perception of threat and willingness to intervene with the post. Posts were categorized into three categories of risk: imminent (within next 24 hours), near-future (within next 1-30 days), and general future (does not specify timeframe). Half of the posts were life-threatening toward oneself, and half were life-threatening toward others. Additionally, half of the posts explicitly stated one’s wish to die or inflict death to others (e.g., “die,” “murder,” “kill,” “suicide,” “dead,” “shoot(er)”, “massacre”). Linguistic characteristics for each post were developed using research from previous literature. Data analysis plans to use ANOVAs to assess for significant differences between mean threat ratings of post by manipulation category. Additionally, ANOVAs will be utilized to assess for significant differences between mean ratings of reported willingness to intervene across manipulation categories. Findings from this study hope to provide preliminary insight into social media post factors that may influence a bystander’s threat assessment of life-threatening posts and willingness to intervene with such a post. If we can better understand factors affecting bystander intervention in life-threatening cyber situations, there may be opportunity to address the state of current intervention and prevention efforts to ultimately reduce life-threatening behaviors that follow certain social media posts.