Political Attitudes II
Democracy substitutes selection by the incompetent many for appointment by the corrupt few.—George Bernard Shaw.
Our government is a government of political parties under the guiding influence of public opinion. There does not seem to be any other method by which a representative government canfunction.—Calvin Coolidge.
I always voted at my party's call, and I never thought of thinking for myself at all. —W. S. Gilbert.
In this chapter we discuss two aspects of voting behavior: individual voting decisions and aggregate voting patterns. Because many of the demographic factors that influence voting behavior as well as political attitudes per se were examined in the preceding chapter, those factors are mentioned only briefly here. Instead, we focus on several other major determinants of individual voting decisions that have been discovered in a series of major election studies.
Historically, the first such factor to be analyzed came from the earliest large-scale scientific election study, conducted by researchers at Columbia University (Lazarsfeld et al., 1948). This research on the 1940 election, reported in a landmark volume called The People's Choice, was a panel study in which 600 residents of Erie County, Ohio, were each interviewed seven times between May and November to investigate factors involved in changing voting preferences. (Interestingly, and unexpectedly, nearly 70% of the respondents showed no changes in voting intentions from start to finish of the study; and similar results have been found for elections in the 1990s—Erikson & Wlezien, 1999. )
Because Lazarsfeld was a sociologist, he adopted a sociological approach in this research, focusing primarily on demographic groups. Predictions of the respondents' voting patterns were made from a score called the Index of Political Predisposition (IPP), based on a combination of three demographic variables—religion, social class, and urban or rural residence. At the various levels of this index, the proportion of respondents voting Democratic ranged from 26% at one extreme to 83% at the other, indicating a strong relationship of the index to voting decisions. (Although most demographic variables no longer relate as closely to voting behavior as they did in the 1930s and 1940s, it is still possible to make reasonably accurate forecasts of election outcomes based solely on demographic 318