Patterns of Partisanship in a Nonpartisan Representational Setting: The Illinois Constitutional Convention

Abstract

Roll call voting the convention was highly structured. Using MacRae's clustering and scaling techniques, 78 per cent of the roll calls with disagreement fitted into 13 substantive dimensions. Although members were elected on a nonpartisan ballot, and did not organize the convention a partisan fashion, partisan identification of the members was the best of several political, social, career, and constituency variables accounting for roll call voting variance on most dimensions. Multiple regression analysis reveals that the other independent variables account for little roll call voting variance on dimensions not explained by partisanship. Recent literature on roll call voting has indicated that the key explanatory variable is ordinarily the party identification of the legislators. Where partisans are clearly identified, this variable seems to outweigh the effects of others both structuring and explaining the roll call voting behavior. In essentially one party legislatures, researchers have found relatively little structure to voting.' In his study of the Oklahoma House of Representatives, Patterson found that there were few discernible dimensions of voting, and that those which could be identified related only weakly to one another. What dimensions there were, were modestly related to a variety of independent variables and the explanations offered for the various dimensions were unique. The inference was that, in the absence of party as a reference group, the legislator is likely, consciously or unconsciously, to respond to