Revisiting Ideology Measures in Political Surveys

The empirical evidence in this article demonstrates that responses to the regularly included questions in the American National Election Study asking respondents to place themselves on an Extremely Liberal to Extremely Conservative scale (LibCon) increasingly contain valid and useful information for a wide range of political analyses. Responses to this question in recent decades provide reliable summaries of respondents’ positions on a range of specific policy issues. These responses are now also strongly related to important political behaviors, such as turnout and vote. The evidence comes from a novel, more comprehensive and nuanced model of the full set of responses to the LibCon question. The results do not contradict earlier analyses of the LibCon responses that argued against using data on ideology but instead show that something has changed: A far broader set of the population can and does understand and use the questions’ ideological labels.