Electoral volatility, political sophistication, trust and efficacy: A study on changes in voter preferences during the Belgian regional elections of 2009

Abstract

In this article, we investigate voter volatility and analyse the causes and motives of switching vote intentions. We test two main sets of variables linked to volatility in literature; political sophistication and political disaffection. Results show that voters with low levels of political efficacy tend to switch more often, both within a campaign and between elections. In the analysis, we differentiate between campaign volatility and inter-election volatility, and by doing so show that the dynamics of a campaign have a profound impact on volatility. The campaign period is when the lowly sophisticated switch their vote intention. Those with higher levels of interest in politics have switched their intention before the campaign has started. The data for this analysis are from the three wave PartiRep Belgian Election Study (2009).

Appendices

Appendix A

Table A1

Table A1 Inter-election volatility (recall 2007 – wave 3 2009)

Appendix B

Language: Language of the survey, either Dutch (0) or French (1).

Gender: Male (0) or female (1).

Age: In years, calculated by distracting the reported year of birth from 2009 (the year the survey was held in).

Education: No degree or elementary school degree (1), unfinished high school degree (2), finished high school degree (3) and higher education or university degree (4).

Religious practice: Self-reported scale of religious practice, with values from 1 (never) to 7 (at least once a week).

Social class: Four categories, being non-active (1=retired, unemployed, student or housewife/-men), labour (2), middle class (3=farmer, entrepreneur with less than six employees and civil servants), higher class (4=entrepreneur with more than six employees, profession, board of directors and staff member).

Ideological profile (left–right): Self-reported position on scale from 0 (left) to 10 (right).

Ideological extremeness: Distance from the ideological centre (5) on the left–right scale. Ranging from 0 (when 5 was reported) to 5 (when 0 or 10 was reported).

Political knowledge: Score on five knowledge questions in the survey, thus ranging from 0 (no knowledge) to 5 (high knowledge). The five questions in the Partirep survey are sufficient for a solid measurement and analysis of political knowledge (Hooghe and Walgrave, 2010).

Interest in politics: Self-reported score for interest in politics, ranging from 0 (no interest at all) to 10 (very much interested).

Political participation: Variable composed by means of the scores on 10 participation items (excluding membership in political parties), with scores ranging from 1 (never), 2 (seldom) and 3 (sometimes) to 4 (often). The items questioned were the following: How much have you done the following things during the past 12 months? ‘boycotting certain products’, ‘participating in manifestations’, ‘partaking in illegal protest actions’, ‘sending a letter to a politician’, ‘sending an e-mail to a politician’, ‘signing a petition’, ‘being in the media’, ‘being active in a volunteer organisation’, ‘debating on political issues on internet forums or online discussion groups’, ‘supporting a charity organisation’. As the distribution on these participation items was very unequal, the item scores were converted into dummies. The sum of those dummies then resulted in our participation variable, ranging from 0 to 10. (Cronbach's α 0.62; Eigenvalue 2.45; 24.5 per cent explained variance.)

Exposure to the media: Scores of self-reported exposure to three different media channels (newspaper, news bulletin and radio) during the past 2 weeks. Questions were part of the first wave of the survey. Scores ranging from 1 (never) to 6 (systematically, each week day).

Talking about politics: Sum scale of self-reported scores for talking about politics during the previous month to friends, colleagues and family members. Scores ranging from 1 (never) to 4 (often). Converted into a 0–10 scale. (Cronbach's α 0.76; Eigenvalue 2.27; 67.55 per cent explained variance.)

Internal political efficacy: Composed out of the scores on four items in the survey, ‘I consider myself capable of participating in politics’, ‘I think I would do as good a job as most politicians we elect’, ‘I think I’m better informed about politics and government than most people’, ‘I think I have a rather good understanding of important problems society is dealing with’. Respondents gave a score from 1 (completely disagree) to 5 (fully agree) on each question, the sum of the scores on each question was computed and put on a 0 to 10 scale. (Cronbach's α 0.69; Eigenvalue 2.07; 51.72 per cent explained variance.)

Political trust: Average score of self-reported trust in six political institutions (political parties, the regional government, the regional parliament, the federal government, the federal parliament and politicians). The scores were ranging from 0 (no trust at all) to 10 (complete trust). (Cronbach's α 0.81; Eigenvalue 3.11; 54.76 per cent explained variance.)

External political efficacy: Composed by means of the scores on three questions in the survey, ‘An average citizen does have an impact on politics and what the government is doing’, ‘Voting makes no sense, parties do whatever they want anyway’ and ‘In election times, one party promises more than the other, but eventually, nothing happens anyway’. The scores were ranging from 1 (completely disagree) to 5 (completely agree) and were converted so that all high scores meant high efficacy and low scores meant low efficacy. The sum score was put on a 0 to 10 scale. (Cronbach's α 0.59; Eigenvalue 1.65; 54.93 per cent explained variance.)

Appendix C

Table C1

Table C1 Crosstab inter-election and campaign volatility (N=1300)

Appendix D

Table D1

Table D1 Descriptives independent variables