Voices from the margins: computational analysis of prison-related posts in social media

Abstract Utilizing data from over 11,000 posts and 127,000 responses on Reddit groups over a 10-year period, we employed AI-powered content analysis to examine patterns in the discussions of Reddit users who post about prison and who the model classifies as before/after incarceration across various topics. Our findings revealed significant shifts regarding personal issues. These included issues relative to users’ mental health, substance use, spirituality, family relationships, and guilt and blame expression. The results offered insights into the human challenges faced by individuals navigating the carceral system as expressed through their virtual identities. This study showcases how computational analysis of social media can reveal nuanced, longitudinal patterns in perceived carceral experiences while also highlighting methodological trade-offs between data richness and anonymity.