Expert-based environmental and health risk regulation is widely believed to suffer from a lack of public understanding and legitimacy. On controversial issues such as genetically modified organisms and food-related chemicals, a "lay-expert discrepancy" in the assessment of risks is clearly visible. In this article, we analyze the relationship between scientific experts and ordinary lay citizens in the context of risks from pesticide usage in Denmark. Drawing on concepts from the "sociology of scientific knowledge" (SSK), we contend that differences in risk perception must be understood at the level of social identities. On the basis of qualitative interviews with citizens and experts, respectively, we focus on the multiple ways in which identities come to be employed in actors' risk accounts. Empirically, we identify salient characteristics of "typical" imagined experts and lay-people, while arguing that these conceptions vary identifiably in-between four groups of citizens and experts. On the basis of our findings, some implications for bridging the lay-expert discrepancy on risk issues are sketched out.