Explaining the Vote: Micro-Level Analyses and the Problem of Endogeneity
Much of what we know about voting behavior derives from micro-level analyses of survey data. Political scientists regularly focus on measures of performance evaluation and ideological placement as predictors of the vote, but studies have shown that the cross-sectional variation in these survey evaluations is largely random noise and does not reflect actual preferences. This paper extends this analysis to the formation of the predictors. Essentially, we examine whether performance evaluations, ideological formation, and the vote are all capturing the same phenomenon. We look for culprits in this crime of endogeneity, focusing on rationalization in survey responses, the development of heuristics or cognitive shortcuts, and the ideologicalization of party politics. This research has methodological as well as substantive implications; we question the value of micro as opposed to macro analyses as well as much of the conventional wisdom about why people vote the way they do. These methodological concerns have implications for individual analyses in all of the social sciences.
Keywords: Voting Behavior, Methodology, Endogeneity
Professor, Political Studies, University of Illinois at Springfield
National School for Political Science and Administration in Bucharest