Violence and its consequences on health is increasingly part of the daily activity of health professionals. According to the literature, a large proportion of victims of violence do use emergency services. Those services are often the first contact in the health care system for victims of violence. This potential opportunity to follow the victims demands improvements in the quality of screening, because persons affected by violence will not easily reveal a past or current history of violence. This delicate problem often leads to underdiagnosis. This paper presents a research protocol developed to assess the prevalence of violence in an emergency unit in a university hospital. We obtained a high participation rate of 77% and a high proportion of violence (11.4%), confirming the public health importance of the issue.