An Analysis of American Electoral Data

An Analysis of American Electoral Data The displacement of old beliefs by knowledge rarely comes easily, as Kuhn has shown, nor do ideas dispel older ones simply by reconsidering the same relevant facts. Rather, when views clash with ancient truths, the ways specify, at least in part, what now shall be considered clear-eyed observation as opposed to earlier illusion, and sometimes specify even the measuring instruments that will allow facts to be seen.1 A minor revolution of this Kuhnian order, suggestive of paradigmatic change, has occurred in the field of American electoral history. Few recent scholars working in this field would deny how different current discourse about the history of American elections is from what it was just thirty-or even twentyyears ago. Back then, elections were described largely in terms of the winning party, its platform, and candidates, with hardly a look at election statistics beyond the simplest vote totals and electoral count. Today, a statistical analysis of just those voting returns, at several levels of disaggregation, is part of the normal course of research-with a concomitant shift to computer technology and technical language.2 This minor revolution was well under way by 1968 when Bogue named it the new political history, a term which hinged initially on Benson's innovative case-study of voting patterns in

Political science · Electoral Systems and Political Participation