IT IS ARGUED THAT CATEGORY SCALING - THE MEASUREMENT TECHNIQUE MOST COMMONLY USED FOR DETERMINING DIRECTION AND INTENSITY OF OPINION - IS WEAK, AND MAGNITUDE SCALING OFFERS A MORE VALID TECHNIQUE. THIS IS TESTED IN A FIELD SURVEY EMPLOYING SRC/CPS NATIONAL ELECTION STUDIES DATA. RESULTS ARE PRESENTED AND DISCUSSED.
Virtually all social scientists claim to know about the strength of political beliefs & preferences is based on category scaling. When category & magnitude scales of opinion strength are compared -- either in measuring the perceived strength of physical stimuli or in the evaluation of social-psychological stimuli -- magnitude scaling is almost invariably found to be superior in terms of providing quantitative information about the intensity of people's impressions, preferences, & judgments. An attempt is made to describe magnitude scaling in sufficient detail to provide political scientists with the conceptual basis for appraising the relative costs & benefits of magnitude scaling & to detail the procedural wherewithal so they can employ magnitude scaling in their research. 4 Tables, 4 Figures, 1 Appendix, 49 References. HA.
The measurement technique most commonly used by political scientists for determining the direction and intensity of opinion is category scaling–a procedure with serious weaknesses. Recent developments in psychophysics for the magnitude scaling and validation of sensory eontinua offer a powerful alternative to category scaling. Paralleling explicitly the logic and procedures used to scale psychophysically such variables as the loudness of sound and brightness of light, research methods now make it possible and feasible via a simple paper and pencil technique to obtain accurate, precise, cross-modally valid, magnitude measures of the direction and strength of political opinion from respondents in a survey setting.A field survey, pitting category against magnitude measures for a sampling of the most important items employed in the SRC/CPS national election studies, demonstrates that the category scaling of political variables results in (1) the loss of significant portions of information and on occasion (2) the misclassification of respondents. The results of this scale-confrontation study demonstrate the superior utility of magnitude over category scaling for the description and quantitative analysis of political judgments and preferences.
