The structure of social attitudes in three countries: Tests of a criterial referent theory(1)

Abstract

A criterial referents theory of attitudes was tested cross‐culturally by administering a social attitude referents (single words and short phrases) summated‐rating scale, suitably “transformed”, to 500 graduate students of education in the United States, 470 mature students of psychology in Spain, and 270 students of the social sciences and a stratified random sample of 685 persons in the Netherlands. The theory claims a dualistic (two second‐order factors) structure of social attitudes, thus challenging the common assumption of attitude bipolarity, which arises from the notion that differing attitude orientations oppose each other, e. g., conservatism versus liberalism. First‐order factor analysis of the correlations among the items and second‐order factor analysis of the correlations among the first‐order obliquely rotated factors supported the theoretical predictions with certain exceptions. In general, the first‐order factors of the referents scales contained only conservatism or only liberalism items; negative loading were mostly low in magnitude. Four similar conservative factors and four similar liberalism factors appeared in the three countries, with one, two, or three unique factors in each sample. The second‐order factor analyses yielded two main orthogonal factors, one identifiable as conservatism, the other as liberalism; a third factor was complementary to one of the two main factors. Bipolarity was a relatively minor phenomenon. Thus the results indicated the empirical cross‐cultural validity of the theory, but also the need for its further specification.