Attitudinal referents, statement items and response set: The effect of using differential item-formats on the structure of an ideological domain

• Published in: Quality & quantity 1981-06, Vol.15 (3), p.249-277
• Main Authors: Middendorp, C. P., de Vries, G. C.
• Format: Article
• Language: English
• Subjects: Attitude/Attitudes/Attitudinal; Dutch; Political/Politically/ Politicalization (see also Politics)
• ISSN: 0033-5177; 1573-7845
• DOI: 10.1007/BF00164640

The effects of different item formats on responses to surveys dealing with various issues have been inconclusive, leading many researchers to ignore the problem. However, studies by F. N. Kerlinger ("The Structure and Content of Social Attitude Referents," Educational & Psychological Measurement, 1972, 32, 613-630) & G. D. Wilson et al (in Wilson, G. D. [Ed], The Psychology of Conservatism, New York: Academic Press, 1973) on political attitudes appear to reflect problems in this area. A model of attitudes is constructed according to which conservatives seek freedom in economic matters & equality in noneconomic matters, while progressives seek the reverse. A set of 80 referents was included in a survey of 815 Dutch Rs; for 51 of these referents, statement items were found in another survey of 1,927 Dutch Rs, which could be linked to a referent in 1 of 11 ways. Exploratory factor analyses were conducted on both surveys. For the first survey, 4 factors were identified: a populist or nationalist/socialist factor, a factor reflecting libertarian noneconomic values, a factor reflecting progressive economic values, & a factor including various conservative referents. A somewhat different factor structure is obtained for responses to the second survey. The first factor on the first survey appears to reflect a high valuation of consensus. The emergence of this factor is the outstanding result of using referents rather than statements as stimuli; it appears possible for all political parties to appeal to it, increasing the political confusion beyond what is already inherent in a multidimensional political spectrum. 5 Tables, Appendix. W. H. Stoddard.