The current work assesses the relationship between respondent perceptions of collective efficacy and neighborhood violence. Data used in the analysis combined a community survey from Mesa, Arizona, with census data. Factor analysis provided mixed evidence regarding the factor structure of collective efficacy; therefore, separate regression models were used to test the influence of collective efficacy, social cohesion, and willingness to intervene on levels of neighborhood violence. Analyses found that community structural characteristics including concentrated disadvantage and residential instability significantly predicted perceptions of collective efficacy, social cohesion, and willingness to intervene. In turn each of these variables was related to violent crime after controlling for levels of concentrated disadvantage, residential instability, and individual demographic characteristics. When social cohesion and willingness to intervene were included in a single regression model, only social cohesion was predictive of neighborhood violence. Social cohesion and violent crime had reciprocal effects that were both negative and statistically significant.