THE GOOD CITIZEN

A number of works suggest that at least some partisans update their party identity to reflect their political evaluations some of the time (Allsop & Weisberg, 1988; Brody & Rothenberg, 1988; Carsey & Layman, 2006; Dancey & Goren, 2010; Fiorina, 1981; Franklin, 1984, 1992; Franklin & Jackson, 1983; Highton & Kam, 2011; Jackson, 1975; Lavine, Johnston, & Steenbergen, 2012; MacKuen, Erikson, & Stimson, 1989; Markus & Converse, 1979; Page & Jones, 1979). Even Campbell and colleagues, who stressed the enduring nature of party identification, acknowledged that it is “firm but not immovable” (Campbell et al., 1960, p. 148). Thus, the challenge is to determine the conditions under which partisans are more likely to change their identity to reflect disagreements with their party and those under which they are more likely to rationalize away disagreements to maintain their party allegiance.

For decades, the party identification literature has been preoccupied with the question of whether party identification is predominantly stable or inherently changeable. By developing a dual motivations theory of party identification, this book attempts to push the debate toward the more pertinent underlying question: When is party identification more likely to help and when is it more likely to hurt democracy? The answer lies in examining voters' motivation to hold parties accountable versus their motivation to maintain their team allegiances. Party identification has the potential to help citizens navigate their way through politics, but this requires a willingness to update their party identity. If they do so, their party identity will serve as a running approximation of their evaluations and thus function as an efficient information shortcut.[1] If they fail to update, then their party identity will likely lead them astray.

  • [1] Because citizens possess incomplete information, they will often make errors inchoosing the party with which to align themselves. However, errors attributable solely to incomplete information can be assumed to be random and therefore to cancel out in the aggregate (see, for example, Page & Shapiro, 1992).
 
Source
< Prev   CONTENTS   Source   Next >