Recent reviews of the relevant literature have examined studies conducted from 1995 to 2003 and confirmed past meta-analyses. Children who witness IPV are at risk for a range of behavioral, emotional, and developmental problems (Edleson, Mbilinyi, Beeman, and Hagemeister, 2003; Kitzmann, Gaylord, Holt, and Kenny, 2003; Sox, 2004). This paper reviews research findings on physical and biological functioning of children exposed to IPV; their behavioral, emotional, and cognitive functioning; and social adjustment. A review of research on factors that mediate the effects of violence focuses on children's age and gender. Research on the differential effects of witnessing IPV compared with being a direct victim of child abuse is briefly reviewed. A review of research methods addresses retrospective reporting, hypothetical situations, and parent and child assessments. A discussion of future research directions notes that all methods used in past research are limited in their external validity. Potential difficulties in the area of response bias are significant obstacles to progress in this research area. Moreover, studies that use only one source to establish a child's witness status risk both false positives and false negatives in the assignment of witness and victim status. A lack of appropriate operational definitions of violence adds to the already problematic research design. Central to these design problems is the use of samples taken from domestic-violence shelters, since only a small percentage of women victimized by domestic violence access shelter services. More clearly operationalized definitions and more strictly controlled experimental designs are needed if valid and clear conclusions about the effects of children witnessing IPV are to be drawn. 2 tables and 102 references