Search results

Previous research has found that some individuals give different answers when asked about their national, state and/or local party identifications. These response inconsistencies suggest that there is a level of government dimension to party identification. It is possible, however, that the consistency of individuals' responses to questions about their different party identifications is instead the product of measurement error. The validity of measures of mixed party identifications is investigated in this study. The results of the analysis support the conclusion that there is a level of government dimension to party identification.

Data from the 1988 Canadian Election Study (N = 3,610 interview respondents [Rs]) are drawn on to explore the role of party identification in the 1988 election campaign. Analysis reveals party commitments remarkably like those found in studies of US Rs, which subsequently shaped Rs' reaction to the events of the campaign: issue positions lined up with party identification, as did perceptions of the leader rehabilitated by the campaign. But the campaign also cut through party groups. It is suggested that such a mixture of reinforcement & displacement might well typify campaigns in Canada & elsewhere; implications for academic observers of US campaigns are discussed. 3 Tables, 7 Figures. Adapted from the source document.

Party identification has long been thought to be the unmoved mover of political evaluations. The research presented here shows that party identification is changeable. In fact, over a 2-year period (1990-1992), we can see that there is a substantively important amount of movement. MacKuen, Erikson, and Stimson argue that expectations, not retrospections, lie at the core of political economy. This work takes that logic and shows that the individual-level changes in party identification are explained by both prospective and retrospective economic evaluations. A novel finding in this article is that egocentric economic evaluations are more important than sociotropic evaluations. People are looking at what the government has done to them rather than at what the government has done to the national economy when making political decisions.

National Election Studies data 1952-1988 are used to examine party identification of 3 US nonblack voter generations: those who first voted before, during, & after the New Deal era. It is found that political turmoil from the early 1960s to the mid-1980s shaped basic party identifications in the generation who first voted after 1968, but did not affect dispositions of older generations. This finding confirms the concept of enduring predisposition to a political party, largely unaffected by changing election-specific circumstances. Also discussed are the realignment of party identification in the South 1960-1988, & the circumstances of the national realignment 1980-1988. 6 Tables. Adapted from the source document.

Studies of voting behaviour have been dominated by the Anglo-American experience. With more complex institutional contexts there is reason to expect more complex voter behaviour. The Irish STV system allows voters to choose more than one candidate from more than one party. Examines what Michigan-style party identification should look like under such an electoral system. (SJK)

THE PRESENT ANALYSIS USES DATA FROM 1974 AND 1981 U. S. CROSS SECTIONS, WHICH INCORPORATE A PANEL, TO COMPARE THE STANDARD NES MEASURE OF PARTY IDENTIFICATION (ID) WITH A MEASURE OF PARTISANSHIP DERIVED FROM A PARTY CLOSENESS QUESTION WIDELY EMPLOYED IN CROSS-NATIONAL RESEARCH. IMPORTANT FEATURES OF THE TWO SCALES ARE EXAMINED BY TRANSFORMING THE CLOSENESS MEASURE INTO A SCALE OF VERY CLOSE, FAIRLY CLOSE, NOT VERY CLOSE, AND NO PREFERENCE CORRESPONDING TO THE SEVEN-POINT ID SCALE. THE SCALES ARE HIGHLY CORRELATED AND ARE SIMILAR IN THEIR RELIABILITY. MORE THAN 75% OF THE "INDEPENDENTS" IN THE ID SCALE CHOOSE A PARTY IN THE CLOSENESS VERSION, AND OVER HALF OF THESE SELECT THE "FAIRLY CLOSE" CATEGORY. RESPONDENTS DO NOT VOLUNTEER THAT THEY ARE INDEPENDENTS WHEN THAT ALTERNATIVE IS NOT STATED IN THE QUESTION.