Abstract
This article evaluates some of the statistical evidence offered in support of the need for multidimensional conceptualizations of partisan and ideological identification, with primary emphasis on the "partisan continuum." It presents a test of the dimensionality of partisan and ideological affect that confirms that unidimensional conceptualization is associated with higher levels of political sophistication. It further suggests that the lack of strong negative correlations between "feeling thermometer" stimuli analytically conceived as polar opposites is the result of a pervasive positivity bias in political evaluations. Once positivity is taken into account, the notion that the two parties occupy different points on the same continuum appears to be readily grasped by the public in general. By contrast, familiarity with the concept of a liberal-conservative continuum does seem to require some minimal level of sophistication. These findings suggest that for most analytic purposes there is little need to reconceptualize the traditional index of partisanship.