This study used multiple regression analyses to identify the socioeconomic, demographic and attitudinal correlates of neighborhood differences in the rate of child abuse and neglect. Comparison of actual vs. predicted rates defined high- and low-risk neighborhoods. These "risk scores" then become variables for inclusion in additional neighborhood-level analyses, as well as in analyzing family-level interview data concerning stresses and support systems. A pair of neighborhoods matched for socioeconomic level was selected, one high-risk, the other low-risk. Interviews with expert informants ranging from elementary school principals to mailmen were used to develop neighborhood profiles. Samples of families were drawn from each neighborhood, and interviews conducted to identify stresses and supports, with special emphasis on sources of help, social networks, evaluation of the neighborhood and use of formal family support systems. The results lend support to the neighborhood risk scores. Families in the high-risk neighborhood, though socioeconomically similar to families in the low-risk neighborhood, report less positive evaluation of the neighborhood as a context for child and family development. Furthermore, they reveal a general pattern of "social impoverishment" that parallels findings at the neighborhood level. (Author)