Violence against women as a social or psychological issue did not receive much attention in the professional literature until the mid-1970s. Previous research regarding battered women has been limited to the use of subjects drawn from shelters/safe houses specifically designed for this population. In view of this potential limitation, this study was designed to expand the generalization of the research findings to include a sample of battered women that was drawn from the non-sheltered, general population. Thus, three samples of subjects were tested: a sample of abused women (N=24) drawn from a sheltered population, a sample of battered college women (N=44) drawn from a non-sheltered population, and a group of non-battered women (N=113). Subjects were drawn from introductory psychology classes and from shelters for battered women. Through the evaluation of level of self-esteem of each sample this study ascertained similarities and/or differences between the two samples of battered women, as well as providing a replication of the previously reported self-esteem differential that exists between battered and non-battered women. Results indicated the non-battered women had significantly higher self-esteem than did the sheltered women or the battered women not from a shelter. Further research is needed to explain the increased verbal abuse in the non-shelter battered women versus the shelter battered women. (ABL)