Suchergebnisse

THE AUTHOR REEXAMINES THE CONCEPT OF PARTY IDENTIFICATION AND THE SURVEY QUESTIONS USED TO MEASURE IT. HE ARGUES THAT SEVERAL ATTITUDES MUST BE TAKEN INTO ACCOUNT: ATTITUDES TOWARD THE REPUBLICAN PARTY, THE DEMOCRATIC PARTY, POLITICAL INDEPENDENCE, AND POLITICAL PARTIES GENERALLY MAKE UP THE COMPONENTS OF A MULTIDIMENSIONAL CONCEPTION OF PARTY IDENTIFICATION.

SCALING ANALYSIS IS BASED ON A GEOMETRIC METAPHOR. HOW OUR UNDERSTANDING OF THE METAPHOR AFFECTS OUR USE OF SCALING IS EXAMINED. SEVERAL GEOMETRIC MODELS ARE EXPLORED USING REAL & HYPOTHETICAL DATA. INSTANCES WHICH APPEAR TO BE MULTIDIMENSIONAL ARE SHOWN TO BE UNIDIMENSIONAL UNDER OTHER SCALING MODELS. CONVERSELY, SOME APPARENTLY UNIDIMENSIONAL CASES ARE FOUND TO BE BETTER DESCRIBED AS MULTIDIMENSIONAL. PARTICULAR ATTENTION IS GIVEN TO THE DIFFERENCE BETWEEN MULTIDIMENSIONAL SCALING & FACTOR ANALYSIS. FLEXIBILITY IN THE USE OF SCALING MODELS IS INDICATED, AS IS THE KNOWLEDGE THAT VIEWS OF THE REAL WORLD ARE LIMITED BY THE SCIENTIFIC CONCEPTS USED. SCALING SEEKS ONLY PARTIAL IMAGES OF A REAL WORLD THAT MAY BE FUNDAMENTALLY UNKNOWABLE. 4 TABLES, 12 FIGURES. MODIFIED HA.

THE 1980 AND 1982 AMERICAN NATIONAL ELECTION STUDIES INCLUDE A NEW SERIES OF QUESTIONS ABOUT INDIVIDUAL PARTISANSHIP. IT IS POSSIBLE TO CREATE A 5-POINT SCALE OF PARTY SUPPORT/ CLOSENESS FROM THESE QUESTIONS. THE NEW MEASURE PERFORMS REASONABLY AS REGARDS ITS RELATIONSHIP TO OTHER MEASURES OF PARTSANSHIP TO ITS OWN OVER TIME, AND TO DEPENDENT BEHAVIOR. THERE IS ALSO A NEW QUESTION ON INDEPENDENCE. BUT THIS IS BEST TREATED AS A SEPARATE ITEM RATHER THAN BEING INCORPORATED IN THE PARTY SUPPORT/CLOSENESS SCALE. THE NEW MEASURE ALSO PERFORMS WELL IN MEASURING STRENGTH OF PARTISANSHIP.