The shelter director speaks from the perspective of one whose primary concern is that research with battered women do no harm and achieve benefits for the women. This means that researchers should consult with shelter staff and other service providers about research methodology and goals. Methodology should ensure battered women's anonymity, and interview techniques should not increase or aggravate stress for the women interviewed. Contracts should be made with the researchers to ensure that agreed-upon research procedures are followed. All interviewers should be trained in the nature of domestic violence and the needs of its victims. The researcher discusses various issues of safety, confidentiality, and ethics from the perspective of the researcher. The researcher has the difficult task of collecting accurate and valid data while ensuring that collection methods are not harmful to the battered women participating in the research. Researchers must ensure anonymity and use research techniques that are either therapeutic or benign by the standards of treatment specialists. One significant ethical issue discussed by the researcher is the overlap between partner abuse and child abuse in the same family. Often an interview with a battered woman reveals that her children have also been abused. Is the researcher obligated to report this child abuse to authorities or should the information remain confidential between researcher and subject? The researcher discusses the dilemma and how it can be resolved.