Multiplicity, variability and incongruity in the meanings of risk encountered throughout the research process (and beyond) are key foci of inquiry within socio-cultural risk research, which attaches considerable importance to appreciating participants' perspectives, orientations and contextual understandings. These foci are also associated with epistemological and methodological dilemmas since in risk research, as in social science more widely, researchers' theoretical assumptions and problem formulations inevitably serve as key anchor points for the meanings that come to be assigned to phenomena, events and situations under investigation. In this article, we consider the idea of risk framing to explore how issues arising in connection with variability in the meanings of 'risk' are dealt with in three different socio-cultural projects utilising a variety of forms of in-depth qualitative interview and situated, contextual data analysis to investigate risks in intimate relationships, employment/career transitions, and nuclear hazards. The projects share the aim of seeking to understand how risk perceptions play out in people's everyday life situations in order to illuminate how people live with risk. The paper considers the framing of risk at three stages of the research process: constructing the research topic, presenting the research to participants, and interpreting meanings of data. A key argument established through comparative analysis of these projects is that consideration of risk framing is not only theoretically important but a key, reflexive resource for risk researchers.