A major focus of studies of child abuse has been the identification of children who are at high risk for abuse. Despite this emphasis, little has been written about the research methodology in such studies. This paper reviews the three major research strategies to investigate risk factors: (1) randomized controlled trials, (2) prospective or observational cohort studies, and (3) retrospective or case-control studies. In addition, eight methodological standards are presented that should help to minimize bias in studies of risk factors for child abuse. These standards are: (1) clear description of abuse, (2) choice of a specific control group, (3) equal demographic and clinical susceptibility, (4) clear definition of the risk factor or protective factor, (5) unbiased ascertainment of the risk factor, (6) clear temporal sequence between risk factor and abuser, (7) equal detection of child abuse, and (8) unequal review of abused and nonabused subjects. Most studies of risk factors for child abuse have used either a case-control or, less frequently, an observational cohort design, both of which are nonexperimental research strategies. In such studies, the use of comparable control groups (standards 2 and 3) and equal detection of abuse in exposed and nonexposed children (standard 7) are of major importance in minimizing bias.