Political Behavior Blog

The paper uses an item response model to estimate a latent measure of ideology from a long series of issue questions in the 2000 NES. One of the key findings of the paper, I think, is that this issue-based latent measure differs from respondent’s own ideological self-placement on the traditional liberal-conservative scale – and that the new measure has a significant impact on things like presidential vote choice.

The audience certainly made plenty of suggestions for where to take the project. I guess that's the danger of coming up with a new measure for an important concept - everyone wants to resolve all current debates using it. Suggestions included seeing whether the dimensions of ideology had remained the same over time, whether this measure of ideology had become more correlated over time with party ID or vote choice, and whether mass ideology had changed in response to elite ideology or was in fact distinct from it. People also wanted to know about those respondents who seem cross-pressured between ideological dimensions, or between new issues and theirprevious ideological affiliations.

One comment really struck me, though. Ideology seems to be of increasing interest to pundits and political scientists alike. Yet we seem to have little conceptual guidance on what the term means – to some extent we don’t quite know what we’re looking for even as we develop different measures of it. Respondents to surveys are clearly thinking of something other than just issue preferences when placing themselves on the liberal-conservative scale. But what is that "something else"? What should they be thinking of? And should political scientists be re-evaluating what we mean when we say “ideological�? I’ve already tried to persuade one political theorist to write a dissertation with the working title "The Concept of Ideology", but to no avail. Hopefully this kind of empirical work will persuade someone to fill these gaps in the way we think about what is a substantively important concept.