According to surveys, the proportion of people believing that AIDS could be caught by donating blood has increased dramatically, from 28.9% in October 1988 to 43.5% in November 1989. One possible explanation for this change was that an inadvertent change in the order in which the question about donating blood was asked had contributed to the increase in the number of those responding that it could lead to infection with the AIDS virus. A December 1989 experiment revealed that a substantial part of the total increase in misinformation about AIDS transmission appears to be due to a form effect, a “confusion” factor contributed by the context in which the question was asked. The results of the experiment are of interest for methodological as well as substantive reasons. Methodologically, they illustrate how question order in a list of items may help clarify the meaning of an otherwise ambiguous item by means of a contrast effect. Substantively, clarifying and counteracting respondents' misperceptions about the risks of donating blood are necessary to maintain the flow of such donations.