Consensual Attitudes and Attitude Structure

In: Public opinion quarterly: journal of the American Association for Public Opinion Research, Band 40, Heft 3, S. 349-359

Abstract

The analysis of attitude structure has frequently sought proof of structure by searching for high nonrandom statistical association between questionnaire items. This strategy will frequently mislead because: (1) highly consensual attitudes are excluded from analysis as "uninteresting," & (2) if such items are employed, most measures of statistical association give misleading indications of item consistency. The 1st argument is considered in terms of what constitutes "interesting" political analysis. The 2nd argument is demonstrated with both hypothetical data & 1968 survey data from the U of Michigan's Survey Research Center on race related questions among blacks & whites. Low correlations frequently mask extensive structure. Overall, current methods for determining attitude structure are unlikely to show this structure when it does exist. 5 Tables, 2 Figures. AA.