According to the nationally based 1992 British Crime Survey (BCS), one in ten women who have lived with a partner have at some time experienced physical violence. This article examines some of the problems involved in estimating the extent of domestic violence through the use of surveys and outlines the factors to be taken into account when comparisons are made between the BCS and other survey estimates. These factors include the differing definitions of domestic violence; whether single or repeated victimization is counted; the use of different survey methods; the geographical areas covered by the survey; and the nature of the samples examined. The BCS estimates of the extent of domestic violence in 1991 are summarized briefly. Recommendations for future research include measuring domestic violence outside of the context of a criminal survey with a specially designed questionnaire with carefully selected interviewers. Tables, references