The sample consisted of 854 questionnaires completed by readers of an article in Women's Day magazine, plus 146 once-battered wives who were the basis for this article. Women in this sample were drawn from all major groups in American society. Some descriptors were: 93 percent white; 74 percent high school graduates; 11 percent college graduates; and a mean number of 1.8 children with the batterer. Almost 43 percent of the women and 38 percent of the batterers held at least one job during their years together that was classified as professional, technical, or managerial. The level of violence suffered by these women was quite high. Of the 772 women who reported having children with the batterer, 70 percent were aware that he also had assaulted the children. Of these battered wives, 76 percent sought shelter with a relative, a friend, or a formal shelter. One reason why formal shelters were used less frequently than other resources was that few women had a shelter nearby when they needed it. Using a broad definition of shelter services, the survey found that 26 percent of the sample received some kind of service from a shelter. Of this gorup, 44 percent rated the shelter as very effective in helping them to decrease or end the violence, 12 percent as somewhat effective, 16 percent as slightly effective, and 22 percent as not effective. The battered women were approximately twice as likely to turn to the police and social service or counseling agencies for help as they were to visit a battered women's shelter, another indication of the low availability of shelters. Shelters, however, were more likely to be rated very effective than any other formal sources of help. Policy implications of the survey's findings are discussed. The article includes 29 footnotes.