Nonattitudes and American Public Opinion: The Examination of a Thesis

Abstract

This paper utilizes the 1956–58–60 SRC panel study to examine the degree to which Americans hold attitudes on issues of public policy. The conclusions reject the thesis that only 20 to 30 per cent of the American public have true attitudes and that the remainder either refuse to take a position or respond randomly. The nonattitude thesis is rejected on the basis of: (1) a conceptualization of attitudes which allows for variation in responses through time without necessarily indicating the absence of attitudes or their random fluctuation; (2) an evaluation of the major statement of the nonattitude thesis; (3) a probability model for measuring attitudes in a panel study based on the assumption of twin samples, i.e., a sample of the population at one point in time, and a sample of the individual's attitude through time; and (4) the application of the probability model, leading to the conclusion that the number of individuals with attitudes has been severely underestimated. The implications of that finding are drawn for the relation of responses to attitudes and for democratic elitism.