Adverse Mental Health Events as a Trigger for Online Harassment: Evidence from Online Conversations of Youth Experiencing Psychiatric Hospitalizations

Abstract

Online harassment is a pervasive issue that disproportionately affects individuals with mental health challenges. Yet, modeling under what circumstances an individual might become susceptible to harassment remains under-explored. This study investigates how psychiatric hospitalization may causally impact the likelihood of experiencing online harassment, using over 360,000 Instagram direct messages shared by individuals with and without mental health diagnoses. Our analysis reveals distinct linguistic patterns in harassment directed at mental health patients, with a significant increase in harassment following hospitalization. A Difference-in-Differences analysis highlights the critical two-week period post-hospitalization as a time of heightened risk to harassment. Additionally, we explore the protective role of social support and resilience through Vector Autoregressive modeling, showing that emotional and informational support significantly reduce the likelihood of harassment. These findings emphasize the need for context-aware interventions and trauma-informed platform designs to protect vulnerable individuals online. By understanding the interplay between mental health events, social support, and online harassment, this research offers new directions for developing strategies to foster safer digital environments. Content Warning: This paper includes mentions and descriptions of online harassment and mental health challenges.