The purposes of this research were to identify risk factors for reported child abuse or neglect and to examine the roles of stress and social support in the etiology of child maltreatment. Mothers of newborn infants with biomedical and sociodemographic risk factors were recruited from community and regional hospitals and local health departments in 42 counties of North and South Carolina selected for geographic distribution and for large numbers of such newborns. For every four such mothers, the next mother to deliver an otherwise normal newborn was sought. Mothers were interviewed shortly after giving birth, and state Central Registries of Child Abuse and Neglect were reviewed when each infant was 1 year of age. Eight hundred forty-two of 1,111 recruited mothers were successfully interviewed in their homes between March 1986 and June 1987. Seven hundred forty-nine North Carolina births who resided in the state more than 6 months were eligible for inclusion in the analysis. Logistic regression with backward elimination procedures was used in the analysis. Maternal education (p < .01), number of other dependent children in the home (p < .01), receipt of Medicaid (p < .01), maternal depression (p < .05), and whether the maternal subject lived with her own mother at age 14 years (p < .05) were the best predictors of a maltreatment report. Further examination revealed an interaction effect between stressful life events, as measured by life event scores, and social well-being (p < .01). For children born at risk for social and/or medical problems, extreme low income (participation in public income support programs), low maternal education, maternal depression, the presence of any other young children in the home, and a mother's separation at age 14 years from her own mother significantly predict child maltreatment reports in the first year of life. In addition, stressful life events, even if perceived positively, may increase or decrease the risk of maltreatment reports, depending upon the presence of social support.