Abstract
In Reply. —The "Data and Methods" section of our article noted five criteria for inclusion of data in our review, including (of importance here) relevance to key policy choices, consistency with other survey findings, and avoidance of biased or confusing questions. We specifically noted the problems with the wording of Roe v Wade questions because, while we thought any discussion of abortion and its legality must include available data about the public's views of Roe v Wade, such a discussion should also acknowledge a potential problem in relying solely on these questions to understand the public's views of the legality of abortion. In keeping with our other criteria, we looked at other surveys for confirmation of what we saw as a general trend in favor of legality that persisted despite the public's apparent misgivings about the circumstances under which legal abortion services are utilized. Indeed, as we presented, several general