The sociological and psychological influences on public support for the European Union in Britain, 1983–2005

Abstract

Previous research has suggested that in Britain older, less educated and working class voters tend to be less supportive of Europe than the young, educated and middle class. This research examines the sociological and psychological influences on opinions towards the European Union (EU) over time in Britain. It builds on existing insights from both single-country and cross-national studies of the sources of support for the EU and examines the main theoretical approaches based on interests, identities and domestic politics. It uses six BES cross-section surveys, binomial logistic regression and a multi-stage causal model for the period 1983–2005 to assess the apparent impact of a range of variables on public support for Europe. This enables estimation of the total and indirect effect for each explanatory variable and provides a more detailed and comprehensive review of the causes of opinion on the European issue. As the analysis covers a period of over two decades, it is also able to gauge whether the factors underpinning support or opposition for the EU have changed over time.

Appendix: Coding of explanatory variables

Sociological variables

• Age (cohort): older cohorts are less supportive of Europe as a result of their wartime experience and innate conservatism. Age is scored to represent membership of a series of cohorts (18–24, 25–34, 35–44, 45–54, 55–64 and 65 plus).

• Age (lifecycle): older voters are less supportive of the EU because they become more conservative as they grow older.

• Gender: the theoretical expectations here are not strong. Women are either more (or less) supportive of the EU as a result of their internationalism (or responsibility for the family budget). This is scored as a dichotomous variable (1 if male, 0 otherwise).

• Education: more highly educated individuals are more supportive of the EU, because they tend to understand the potential benefits or because they have highly paid jobs, higher levels of ‘human capital’, and can directly benefit from the opportunities created by the Single Market. This is measured by the age that the respondent left full-time education (terminal educational age or ‘TEA’)

• Religion: as earlier research has shown, Catholics are expected to be consistently more supportive of the EU, compared to adherents of other religious denominations and those of no religion. Religion is measured as a series of dummy variables for Church of England/Anglican, Catholics, other Christian and non-Christian (where available) The base category consists of those respondents with no religion.

• Trade union membership: in the 1970s and early 1980s, trade unions members were less likely to support the EU because it was seen as a ‘capitalist club’. In the late 1980s and 1990s, they are more supportive, following the cues of the union leaderships. Trade union membership is scored as a dichotomous variable (1 if a member, 0 otherwise).

• Social class: those in the higher social grades (such as the salariat) are more likely to be supportive of the EU. Those in lower social grades (working class) are more likely to be fearful of the changes brought about by the Single Market and, therefore, less supportive. Social class is measured by a series of dummy variables (for salariat, routine non-manual (RNM), petty bourgeoisie, manual foremen and supervisors, and working class), using the Goldthorpe-Heath five-category schema.

• Controls: this study also assesses the influence of those variables that have in previous studies been found to be associated with opinions on the EU. Including these socio-demographic background characteristics can also help us in trying to avoid the possibility of drawing spurious inferences. These include race (scored as a dichotomous variable, 1 if not white and 0 otherwise), housing tenure (dummy variables for homeowner, council tenant or private rental), public or private sector employment (1 if in public sector, 0 otherwise), region (dummy variables for Scotland, Wales, Northern England, the Midlands and Southern England), and religiosity (scored on a 0–1 scale, with 1 being attend services once a week or more, and 0 for never attend services, or attend less than once a year).

Psychological variables

• Partisanship: partisan voters often tend to follow the cues provided by the parties. This research uses answers to the traditional BES question (‘Generally speaking, do you think of yourself as Conservative, Labour, Liberal Democrat or what?’). This is scored as a series of dummy variables (Labour, Conservative, Liberal Democrats (or Liberals), other identifiers, and non-identifiers).

• Ideological positions: measures of an individual's position on both the left-right scale and liberal-authoritarian scale are available – two important dimensions of enduring ideological conflict within British society. Each scale is generally constructed from a battery of five or six separate agree-disagree statements. For three of the six cross-sections, 1983, 1992 and 1997, a variable measuring individuals’ materialist or post-materialist value priorities can be constructed (coded 1 if a materialist value priority is chosen, 0 if a post-materialist value priority is selected).

• Newspaper readership: here, a series of dummy variables are used to control for the national daily newspaper that a respondent reads.

• Interest: where available, the level of interest an individual has towards politics can be measured. Interest is measured on a scale (scored from 0 to 1) where higher scores indicate greater levels of interest in politics.

• Knowledge: the indicator of political knowledge used (where available) is one that measures chronic knowledge – an index based on a series of true or false statements about British politics and government – rather than any measure of domain-specific knowledge. Here, a scale is constructed (scored from 0 to 1) to measure levels of political knowledge, where higher scores represent higher levels of knowledge. It is worth stressing here that political knowledge has been measured irregularly and inadequately in previous BES studies.

• Economic evaluations: in line with existing research, measures are included for subjective economic evaluations, both prospective and retrospective (though these are not measured consistently in the BES surveys used here). These are generally treated as relatively short-term appraisals, grounded in citizen's daily lives, which can impact on other opinions. These are scored on a scale from 0 to 1, so that those with positive evaluations are assigned higher scores. Previous research has tended to find that positive appraisals of economic circumstances are associated with higher levels of support for the EU. Unfortunately, a full battery of economic assessments – prospective and retrospective, sociotropic and egocentric – are only available in the 1997, 2001 and 2005 BES data sets.

• Other factors: also included are variables measuring whether a respondent cared who won the general election (1 if yes, 0 otherwise), tapping into people's engagement with their political environment, and their preferences for either single-party or coalition government (1 for coalition, 0 otherwise), an important institutional difference between Britain's political system and many continental polities. Those respondents who express a preference for coalition government can be expected to show more support for the EU, as this mode of administration may be perceived to be a ‘desirable’ feature of other political systems within the EU from which Britain can learn.