Compared to the general U.S. population, individuals incarcerated in the United States are at greater risk of self-directed violence (SDV) and depression. SDV risk factors lie within multiple domains, and critical risk factors include depression and the absence of social connection. Research has yet to comprehensively examine variation in SDV risk factors based on individual characteristics (i.e., age) or investigate the influence of modifiable risk factors that are multilevel in nature (e.g., social connectedness) on depressive symptoms and SDV in prison settings. Therefore, this dissertation research has two aims: 1) to examine demographic-based variation in SDV risk factors, including depressive symptoms and social connection, and 2) to 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 aims. Framed within the context of the SEM, study one is a systematic review that synthesizes findings from fifteen studies investigating the association between social connection with depression and SDV in U.S. prisons. In study two, data from the development of the Self Injury Risk Assessment Protocol for Corrections (SIRAP-C) is used to explore variations in SDV-related risk factors based on demographic characteristics such as age, race, and sex. Study three uses SIRAP-C data and two hierarchical regression models to investigate the relationships between multilevel social connection with depressive symptoms and SDV events for a sample of adults incarcerated in a U.S. state prison system. Altogether, the findings speak to the applicability of the SIRAP-C for incarcerated individuals of differing demographic backgrounds, whether there is significant variation in risk factors based on demographic characteristics, and provide additional insight regarding the role of social connection and how it can be used in the future address depressive symptoms and SDV and depression for this population.