The Instagram/Facebook ban on graphic self-harm imagery: A sentiment analysis and topic modeling approach
Abstract
Research indicates that adolescents who are vulnerable to self-harm are likely to access social media specifically related to self-harm, which often contains health myths and fails to provide linkages to appropriate resources. In this study, we examined the emotional discourse before and following a ban of graphic images of self-harm on Facebook and Instagram.
This observational case series study assesses a corpus of 8,013 tweets that specifically mention Facebook and/or Instagram between June 1, 2018, and May 31, 2019 (the ban occurred on February 7, 2019). Topical content was elucidated via statistical modeling to extract abstract topics in discourse.
The graphic self-harm ban produced emotional shifts in users of social media. A time series plot produced peaks of tweets on the banning of graphic images of self-harm including political events, news coverage, and media exemplars of suicide (both celebrity and noncelebrity).
We observed raw proportional differences with elevated emotions before the ban (i.e., disgust, joy, surprise, and trust) followed by elevated emotions occurring after the ban (i.e., anger, anticipation, and sadness). This study provides insight into how the policy-change relating to self-harm was received by those with a vested interest.
These insights can be valuable to policymakers and advocates for those affected by self-harm.