The increasing availability of georeferenced datasets creates new opportunities to perform spatial analysis of social science and public health survey data, but also raises ethical issues regarding the potential for unintended violation of the confidentiality of respondents. This article examines these ethical challenges by reflecting on the experience of a study mapping the facilities that provide abortion-related services in Cambodia. The technique of masking is examined as a potential method for preventing reidentification of respondents in georeferenced surveys. Broader solutions are offered for ways to balance the potentially conflicting goals of spatial analysis and protection of confidentiality.