The ambiguities of risk which stem from its translation from epidemiological findings into clinical knowledge and practice and thus to lay experiences of health and illness is a clear dilemma. How are risks expressed statistically, or otherwise mathematically, to be interpreted and communicated within the discourse of medico-science, and how within the discourse of an individual's everyday life? An important tool in all risk discourses and in preventive practices such as health information is testing and test results. Test results--presented in mathematical terms as points on a scale, or as a number--are in fact fundamental to preventive practice. But what do we know about how people involved in these tests understand them and how the results are used in the construction of ideas about risk and normalcy? This article attempts to answer part of that question by drawing on an empirical study of the use of numbers as metaphors in talks between a nurse and her potential patients in a directed health survey.