Measuring the Median: The Risks of Inferring Beliefs from Votes

Abstract

AbstractA large number of studies of ideological congruence, and of the effect of public opinion on policy outcomes more generally, have relied on the Kim-Fording (KF) measure of median voter opinion. This measure has the great virtue of being readily calculable – no direct measurement of voter opinion is required – but it rests on assumptions concerning party locations and voter behaviour that are unquestionably incorrect, at least some of the time. This article explores the sensitivity of the KF measure to violations of its core assumptions through simulation experiments. It then uses public opinion data to assess the degree to which consequential levels of violation occur in actual democratic systems. The article concludes with a discussion of what the KF median really measures and where it can – and cannot – be safely used.