Voting is the main form of political participation in liberal democratic societies and the study of voting behaviour is a highly specialized sub-field within political science. The analysis of voting patterns invariably focuses on the determinants of why people vote as they do and how they arrive at the decisions they make. Sociologists tend to look to the socio-economic determinants of support for political parties, observing the correlations between class, occupation, ethnicity, sex, age and vote; political scientists have concentrated on the influence of political factors such as issues, political programmes, electoral campaigns, and the popularity of party leaders on voting behaviour. However, both disciplines share much the same terrain, and increasingly have tended to overlap in their analytical approaches (see M. Harrop and W. L. Miller, Elections and Voters: A Comparative Perspective, 1987).
A number of different (but not mutually exclusive) approaches to the explanation of voting behaviour can be distinguished in the literature. Structural (or sociological) approaches concentrate on the relationship between individual and social structure, place the vote in a social context, and examine the effects on voting of such variables as social class, language, nationalism, religion, and rural-urban contrasts. Ecological (or aggregate statistical) approaches relate voting patterns to the characteristic features of a geographical area (ward, constituency, state, or whatever). Social psychological approaches relate voting decisions to the voter's psychological predispositions or attitudes, for example his or her party identification, attitudes to candidates, and such like. Finally, rational-choice approaches attempt to explain voting behaviour as the outcome of a series of instrumental cost-benefit calculations by the individual, assessing the relative desirability of specific electoral outcomes in terms of the issues addressed and policies espoused by the different parties or candidates. Each of these broad approaches tends to be associated with different research techniques and each makes different assumptions about what motivates political behaviour.
In Britain there has been a long-running debate about whether the influence of social class on voting behaviour has declined (the so-called ‘class dealignment thesis’), and about the extent to which this process is associated with the dilution of loyalty to the two major parties (the Conservative Party and the Labour Party) which have dominated the political system since the Second World War (the ‘partisan dealignment thesis’). Proponents of these arguments (see, for example, B. Sarlvik and I. Crewe, Decade of Dealignment, 1983) argue that both absolute class voting (the overall proportion of the electorate who vote for their ‘natural’ class party) and relative class voting (the relative strength of the parties in different classes) have declined continuously since the late 1960s and that this is connected to the decline in the share of the Conservative and Labour Party votes. They attribute this dealignment to a number of underlying social changes: changes in the occupational structure, the decline in the size of the manual working class, social mobility, and growth of cross-class families—all of which are said to undermine the socio-economic cohesiveness of class. As a result of class fragmentation, issues have become a more important influence on how electors vote, and voters evaluate the political parties as self-interested individuals rather than on a collective or class basis.