The Dimensionality of Party Identification: Cross-National Perspectives

Rigorous description of reality is of crucial concern to the social sciences. Frequently the predictions of common sense or the conventional wisdom have proven unfounded when confronted with systematically collected and carefully analyzed data. As important as the establishment of facts is, however, real advances are only made when facts are fitted together in a framework or theory that both renders explicable their relation to one another and leads to the postulation of new hypotheses. The confirmation (or rejection) of these hypotheses then adds to the store of facts and thus to the refinement, modification, or even rejection of the theory leading to their generation. Progress lies in abstraction and generalization from raw facts. One abstraction that tremendously advanced the study of electoral behavior was the concept of identification. Long before the authors of The American Voter operationally defined this concept in the now classic way,1 it had been recognized that voters did not approach elections with completely open minds. Without the benefit of modem social science, Washington, Madison, Hume, and Calhoun all recognized the importance of the spirit of or party attachments in influencing electoral behavior. The advance for which Angus Campbell et al. deserve credit is the casting of this fact in a socialpsychological framework. Once identification was likened to other psychological predispositions, it could be fitted into a more comprehensive theory of electoral choice. Additionally, a variety of hypotheses concerning the acquisition of partisanship were suggested, many of which have been verified by subsequent investigation. Party identification has become a key concept in the study of mass political behavior in democratic countries. For all its operational specificity and theoretical centrality, however, the