Most research on representation in the US Congress examines the relationship between the voting behaviour of legislators and citizens' preferences. Using a roll call-based measure (DW-NOMINATE scores) of US Senators' ideologies and citizen assessments of Senators' ideologies (from the American National Election Study: Pooled Senate Election Study 1988, 1990, 1992), we demonstrate that judgements about Senators' responsiveness (the extent to which more liberal/conservative Senators represent more liberal/conservative states) is sensitive to measurement issues. In addition, using citizen assessments, we move beyond responsiveness to examine ‘representativeness’ (the extent to which Senators' ideologies match the average ideologies of their constituents) as a dimension of representation. We conclude that citizen assessments should not be overlooked since representation may ultimately be a phenomenon that rests in the eye of the beholder and since governmental legitimacy itself rests on citizens' attitudes an...