Social capital is one of the most important concepts developed in social science in recent years. Coleman (1988, 1990) has defined social capital as resources existing in social structure and relationships that facilitate social action. Sampson and Laub's work was the first to apply the concept of social capital in criminology. In their theory on turning points in an offender's life course, Sampson and Laub explained how social capital plays an important role. Their theory proposes that a supportive marriage and employment reduce the risk of further offending. The data for the current research were obtained from a survey of inmates conducted by faculty members at the Center of Criminology at China's Tianjin Academy of Social Sciences. The survey was conducted in the fall of 1992. Based on the available resources, a 25-percent sample was randomly drawn from the roster of all inmates admitted into Tianjin prisons in 1991. This procedure yielded a total of 1,063 inmates. The analysis for this study was limited to the 269 recidivists who had experienced at least one rearrest. The procedure analyzed how social-capital indicators predicted the risk of reoffending since last release. The analysis extends the concept of social capital by proposing a new concept of "negative social capital," which exists in unconventional relationships such as gangs. The unconventional relations seem especially salient within Chinese gangs because of their special cultural roots. The findings show that, after controlling for the usual covariates of reoffending-risk used in Western literature, social-capital and negative-social-capital variables show significant effects. Being single entails a significantly higher risk of reoffending than being married. Job arrangement after release significantly reduces reoffending risk. Participation in legal education reduces the risk of reoffending. Being a Chinese gang member significantly increases the risk of recidivism. 1 table and 66 references