Traditionally, anthropologists have sympathized with marginalized objects of study – people who do not have the power to speak for themselves. Acting as advocates of subordinate groups has thus seemed much more attractive to anthropologists than studying institutions and powerful informants. Though many recent studies of Western institutions and elite groups have modified this viewpoint, a marked tendency still exists to criticise these informants for being unambiguously hegemonic, whereas minorities are almost automatically rendered as complex. The article considers the methodological implications of letting the status of the informants influence which ethical codex is to be followed – a sympathetic approach to marginalized people, or a cr...