Voting Behaviour

You are here

Voting Behaviour

Studies of voting behaviour are based upon finding the important variable that cause people to vote one way or another.

There is a belief that political alignments reflect social cleavages. So research tries to identify these significant social cleavages.

There have been discernible changes in voting patterns. So, in order to identify change, a point of reference is required. This is the so called sociological, or class based model.

An essential element of the British political party model of government is that parties aim to win office by competing for the vote of the electorate on the basis of a programme of policies. It is was then argued that individual voters are rational and informed, they have an appreciation of their own best interests and so vote accordingly.

Ballot box

The electorate has not matched this model of an active and informed public involvement in politics. Studies revealed just how limited the information that voters had concerning the policy position of the parties was, and that they often supported a party in spite of its policies, instead of because of them (Dearlove and Saunders, 1984).

In a landmark text - Political Change in Britain, Butler & Stokes (1963), challenged the idea of an informed electorate. They argued that the behaviour of the electorate is shaped less by particular issues than it was by generalised attitudes and beliefs about party image.

It is this idea of a party image that was taken by most commentators of the time as evidence that Britain, at least until the late 1970s, was the purist example of class voting in any Western industrial democracy.

Pulzer (1967), claimed:

'Class is the basis of British party politics; all else is embellishment and detail.'

Traditionally then, political sociologists have identified social class as the most important factor associated with voting behaviour. This is hardly surprising given that one of the major political parties - Labour - was founded upon a commitment to a class, and whose origins lie in the organised trade union movement.

The sociological model of voting behaviour, associated with Butler and Stokes, was based on the consistent finding that social class was the most accurate indicator of likely voting intention. In general, working class voters did vote for the Labour Party, and middle class voters voted Conservative. Usually, about two-thirds of the working class voted Labour, while four-fifths of the middle class voted Conservative.

In the period 1946-70, it was assumed that stable two party voting was the norm. During this period, Marsh, Developments in Sociology, argued that the following situation prevailed:

  • Two party dominance. The overwhelming majority of the electorate voted Conservative or Labour.
  • Partisan alignment. Most people identify with the party for which they vote. Party membership high compared with other countries. The majority supported the policies of party for which they vote.
  • Electoral volatility was low, few people changed their vote between elections.

Thus the sociological model of voting behaviour, put forward by Butler and Stokes (1963), started with what was seen as the major social cleavage in British society - social class:

  1. Social cleavage (especially social class) indicated.
  2. The type of political socialisation in a family, which lead to.
  3. Party identification, leading to.
  4. Political attitudes and voting intentions.

Most analysts, while still supported to some extent, by sociologists who argue for the continuing importance of social class as the main indicator of likely voting intention, have largely abandoned this model. Put simply, the reason that the model was abandoned was that it was losing its predictive value.

The main criticisms of the sociological model are:

  1. Its inability to explain partisan de-alignment.
  2. The over emphasis on the utility of social class as an indicator of voting behaviour.
  3. A failure to appreciate the significance of individual policy preferences and assessment of government performance on voting.

There have been no major upheavals, or extensive social mobility to explain these changes, but there has been slow but steady change:

The manual working class has never voted solidly for Labour, and the proportion voting Labour has been falling. Working class Labour vote in 1959 was 62%, in 1983 it was 38%. And this is at a time when working class jobs have been in decline.


The traditional support base for Labour has been getting smaller. Technological change, and the decline of the British economy has reduced manual working class occupations. It is no longer the case that solid working class support would win power for Labour. This was first recognised in the 1950s, when the Conservatives won three successive general elections.

While the Labour vote remains mostly working class, the majority of the working class are no longer Labour voters. In 1983, Labour received 38% of the manual working class vote. By 1983, the Labour Party was - argues Crewe (1983) - more accurately viewed as the party of the traditional working class in Scotland and the north, the public sector and council estates.

Dunleavy (1983) identifies two key elements in new patterns of voting behaviour:

Class De-alignment

- A major weakening of the association between people's occupational class positions and their political alignments.

Up to 1974, class de-alignment mainly involved a weakening of the Conservative leanings of the middle class. Labour voters from the middle class grew from one in six in the mid 1960s to over one in four in the mid 1970s. A second phase of class de-alignment occurred in the late 1970s when there was a large growth in Conservative support among skilled working class people.

However, while class membership became less useful as a predictor of voting intention, certain class corollaries still seemed useful guides. Three examples are worth noting:

  • Members of Trade Unions are much more likely to vote Labour than non-members.
  • Housing tenure continued to be closely associated with voting, homeowners were twice as likely to vote Conservative than were council tenants.
  • Voters in a constituency are influenced towards the characteristic alignment of the majority occupational class in the area.

Partisan De-alignment

- A significant reduction in the fit between people's choice of party and their views on political issues.

The political science orthodoxy of the early 1960s, suggested that most voters come to identify with political parties, mainly via family socialisation in early life and imitating the views of others with whom they come into contact in their daily lives, at work or in local neighbourhoods. Voters' loyalties were, on the whole, affective or emotional in character, rather than based on any closely rationalised philosophy or calculation.

Voting was seen as meaningful, but not as rational. Once formed, these party loyalties tended to endure over successive elections, producing stable blocks of voters. Voters tended to adjust their views on individual issues to fit in with their party loyalties, rather than vice versa.

Three developments have been identified that have weakened the earlier association between party identification and attitudes on specific policies.

Third party voting/Two party decline

The most obvious indicator is the falling share of the poll going to the two major parties, falling from 97% of the vote in 1955 to 77% in the mid 1970s. In addition, the proportion of people with strong attachments halved over the 1970s to just one voter in every five. It seems that people's party loyalties are becoming more conditional.

Issue changes

Three changes are involved. First, some issues have declined in importance. Crewe et al (1977) suggest that some issue decay particularly affected Labour Party support on questions of basic principle for the party such as, further nationalisation, the extension of the welfare state and its close links with the TU movement. Support for these policies decayed significantly between the mid 1960s and 70s. It is further argued that Labour failed to find new policies with major appeal.

A number of new issues grew in importance from the mid 1970s, and they were associated with the Conservatives. These are taxes and law 'n' order.

Voting Behaviour

Electoral volatility

With third party voting, there is obviously an increase in electoral volatility. For example, The British Elections Study, which interviewed a panel of electors after each election between 1974 and 1979, found that over the four elections of the 1970s, half the electors entitled to vote on all four occasions changed vote, or abstained, at least once, and often more.

There are two new explanations:

  1. Social base explanations-based on class dealignment.
  2. Issue changes-based on partisan dealignment.

These concentrate on changes in the character of British society, in particular, which new lines of social cleavage have opened up, which crosscut occupational class divisions (Dunleavy, 1979, 1980). This new cleavage is a sectoral one, with public/private sector conflicts at its core. In employment terms, which sector you work in has a major impact upon whether or not you are likely to join a trade union, attitudes to incomes policy and towards the extension or rolling back of state intervention. These divisions also exist in consumption processes - how people gain access to goods and services - such as housing transport, health care and education.

Labour has become the public sector and trade union party, rather than a party orientated towards the working class as a whole.

Thus working class people who work in the private sector, own their own homes and have a car, may find themselves on the Conservative side of the divide. Conversely, middle class people who work in the public sector, rely on public transport and are union members, may support Labour.

The emphasis is on an issue approach to voting. Voters are seen as moving towards a more critical stance, deciding issues on their merits and hence acting more rationally than formally. This is basically a consumer model, people actively search for the party which seems most likely to implement policies which they favour and whose style of government they can respect (Himmelweit, 1985).

Another supporter of this approach is Whiteley, with his policy preference model. Essentially, Whiteley argues that attitudes play a crucial role in determining behaviour. Voters are assumed to have preferred policy positions and know the policy positions of parties. They then vote for the party nearest to them in terms of policy preference.

In Harrap's, Voting and the electorate (1988), the two models above are listed, but also includes the Party Identification model. This is a return of the idea that social class is a strong indicator of voting behaviour. Marshall et al (1988) Social Class in Modern Britain, argue the continuing utility of a class model of voting behaviour, albeit with a more sophisticated model of social class (The Goldthorpe model).

Marshall et al argue on the basis of their 1984 data, that the working class has divided politically - half of the skilled manual workers they interviewed did not support Labour, and that factors like home ownership do seem to be associated with political alignments.

However, they do not accept that this indicates increased working class antipathy to socialism. They point out that the core working class is still quite strongly Labour. Furthermore, Labour voting is strong among those who identify themselves as working class. They conclude from this that class is still the key to political sympathies and values in Britain, and that the core working class has been deserting Labour, not because it no longer feels it is working class, but because Labour has been expressing poorly its class values and interests.

It is not, therefore, the working class which has changed, but the Labour Party.