Most prisoners come from economically and socially disadvantaged circumstances in which violence, substance abuse, family disruption, and other traumatic experiences are common. In addition, prisoners exhibit high rates of psychological disorders. Thus, incarceration is but one potentially damaging hardship in many inmates’ troubled lives. The current study examined variation in symptoms indicative of distress among a sample of 208 men who were residents of work release facilities in a Midwestern State. It also examined the direct and indirect effects of support from free-world friends and family, past traumatic experiences, past exposure to violence, self-control, age, race, and time served. Data for the study were collected between September 10 and December 4, 2001. Three hypotheses were tested: 1) that prison victimization is a significant predictor of depressive and post-traumatic stress (PTS) controlling for other variables; 2) previous trauma significantly affects distress measures; and 3) prison victimization links pre-existing characteristics to post-prison distress. The two dependent variables for the study were measures of depressive symptoms and post-traumatic stress symptoms. The data were examined using path analysis. The findings support the first hypothesis and reveal that prison victimization is a significant predictor of PTS symptoms; race, previous trauma, and self-control also had direct effects on PTS symptoms. The direct relationships between PTS symptoms and social support, age, exposure to violence, and time served were insignificant controlling for effects of other variables. The results also found that previous trauma, race, and self-control all had direct effects on depressive symptoms just as they did on PTS symptoms, while previous trauma, exposure to violence, and race had significant indirect effects. This finding supports the second hypothesis. The analysis also revealed that while prison victimization adds to the pains of pre-existing events, it does not mediate the effects of previous events on outcomes, thus disconfirming the third hypothesis. Study limitations are discussed. 3 tables, 2 figures, 6 notes, and 76 references