This paper demonstrates that ethical problems may occur in qualitative research even if it draws on seemingly safe, "exempt" data. Drawing on J.L. Jacobs' concept of "double vision," I describe a number of ethical issues that arose in a project using archived oral Holocaust testimonies; these included concerns about using pseudonyms and identification numbers instead of subjects' real names and problems with coding the testimonies and writing up the results. Despite these concerns, I conclude that double vision need not always be problematic, and can actually help identify guidelines for producing work that is both scientifically and ethically sound.