To date, Bachman's (2000) research is the only published attempt to bridge the methodological differences between the NCVS and NVAWS. Her conclusions, however, were based on an erroneous presentation of NVAWS confidence intervals. The current paper builds upon Bachman's research, corrects the identified error, and addresses the following questions: To what extent do estimates of violence against women from the NCVS and NVAWS differ? To the extent that the estimates differ, are the higher estimates more accurate? After reviewing the features of the two surveys, the published estimates of the two surveys are compared for rapes and physical assaults against women, with disparities documented. Methodological differences in the surveys that can account for their different estimates are identified, and the authors advise that some features of each survey tend to make estimates more accurate while other features may introduce error. They conclude that the methodological differences are so numerous that it is impossible to isolate the influence of each factor; however, they do address one methodological difference, i.e., the way the two surveys treat recurring victimizations in the construction of annual incidence estimates. The authors then describe and apply their analytic strategy for making the NCVS and NVAWS annual incidence estimates as comparable as possible and for conducting statistical tests on these estimates to determine whether resulting estimates of violence against women differ statistically. This is done by restricting NCVS estimates to 1995 and including only persons age 18 or older, as well as by applying the NVAWS series victimization counting protocol to NCVS estimates. 6 tables, 1 figure, and 35 references