Abstract: Examines the methodological bases of different findings in the Anglo-American democracies and tries to resolve these aspects of the disputes through: (1) a comparison of the findings obtained in Britain using the British Election Studies versus Gallup polls, and in the US using the American National election Studies, Gallup polls, and the General Social Survey; (2) comparison of the effects of differences in class coding; and (3) various specifications of loglinear models. The analyses span the period from 1936 to 1987 in Britain and 1936 to 1988 in the US. The effects of the sources of difference are interpreted in conjunction with an emphasis on their substantive implications.