Individuals incarcerated in the United States face significantly higher rates of self-directed violence (SDV) and depression compared to the general U.S. population. Despite this, there is a lack of comprehensive research on variation in SDV risk factors based on individual characteristics (i.e., age) and regarding how modifiable and multilevel risk factors, such as social connectedness, influence depressive symptoms and SDV in prison settings. This dissertation research aims to examine demographic-based differences in SDV risk factors and investigate the relationship between multilevel social connectedness with depressive symptoms and SDV within the context of the social-ecological model (SEM). Three studies were conducted to meet these objectives. Study one is a systematic review reporting findings from fifteen studies that investigated the association between social connection with depression and SDV in U.S. prisons within the context of the SEM. Study two uses data from the development of the Self Injury Risk Assessment Protocol for Corrections (SIRAP-C) to explore variations in SDV-related risk factors based on demographic characteristics like age, race, and sex. Study three uses SIRAP-C data and hierarchical modeling to explore the relationships between social connection across the SEM with depressive symptoms and SDV for a sample of adults incarcerated in a state prison system. Altogether, the findings speak to the applicability of the SIRAP-C across incarcerated individuals from different demographic backgrounds, whether there are significant differences in risk factors based on demographic characteristics, and the importance of social connection at various levels for this population. The findings also demonstrate the potential of socially-based interventions in the promotion of social connection and subsequent mitigation of depressive symptoms and SDV-related behaviors.