https://orcid.org/0000-0002-6039-801X
![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
ABSTRACT
South Africa's recent democratic elections have witnessed a decline in political participation. Yet few studies examine the influence of political efficacy on the decision to participate. Using 2014 and 2019 post-election public opinion survey data captured after these respective national and provincial elections, the paper addresses a research gap by demonstrating a relationship between a respondent's level of internal and external political efficacy and their likelihood of voting across two consecutive elections. It finds that external political efficacy, and to a lesser extent, internal efficacy shapes the decision to vote. The responsiveness of the political system is a key determinant of voting even after other key predictors of participation are accounted for. It concludes that years of state-based failures have eroded citizen perceptions of political system responsiveness which has, in turn, undermined how citizens perceive the utility of their vote.
Acknowledgements
The financial assistance of the National Research Foundation (NRF) towards this research is hereby acknowledged. Opinions expressed and conclusions arrived at, are those of the author and are not necessarily to be attributed to the NRF.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 A more detailed contextual analysis of both elections can be found in Schulz-Herzenberg, C. and R. Southall (eds) Election 2019: Change and Stability in South Africa's Democracy. Johannesburg: Jacana Media, and Schulz-Herzenberg, C. and R. Southall (eds) Election 2014: The Campaigns, Results and Future Prospects. Johannesburg: Jacana Media.
2 Institute for Democracy and Electoral Assistance, Voter Turnout Database, https://www.idea.int/data-tools/data/voter-turnout.
3 CNEP is a multi-national project, run from the Mershon Centre at The Ohio State University, in which teams of scholars study voting behaviour using compatible research designs and a common core of survey questions. CNEP enables both detailed country studies built around factors that are important for those sites, and the conceptual precision and measurement parallelism applied to quite different and often complex types of political experience that supports cross-national comparison. CNEP is the oldest and the third largest cross-national survey project in the world (behind the World Values Study and the Comparative Study of Electoral Systems) and is now one of the largest and most ambitious cross-national elections projects in the world. Data has been fielded in 30 countries with over 59 surveys and has completed three waves (CNEP I, II, III) since 1990 and it is currently administrating the fourth wave (2009 to present) with the addition of non-democratic, illiberal-democratic countries or weak democratic countries. The South African CNEP series has been held shortly after the national and provincial elections in 2004, 2009, 2014 and 2019. Data collection for the 2014 post-election survey took place in early 2015 between February and March while fieldwork for the 2019 post-election survey took place between 10 August and 6 September 2019. The surveys are nationally representative to the entire adult population. The South African surveys were fielded in person, with face-to-face interviews that were conducted in the respondent's home. In order to reach a nationally representative sample, the CNEP used a random, stratified and area probability cluster samples of 1300 and 1625 adults aged 18 and older, for the 2014 and 2019 post-election surveys, respectively. This means that the primary sampling units were selected randomly with probability proportional to the population size. The primary sampling units were stratified by (1) Province, (2) Urban or rural area and (3) race (Citizen Surveys 2019). The number of respondents are slightly higher in 2019 because budget allowed for a larger sample. This is beneficial in survey studies for investigations that wish to gain greater accuracy for sub-group analyses. In the descriptive analysis of the variables, the figures are based on the weighted data required by the CNEP sample design. The development and calculation of the sample weights were assigned to extrapolate the obtained sample to national population statistics using a scientific method called a multi-stage stratified sample design. The weights were adjusted to avoid the under-representation of certain sections of the population. Datasets and further technical information on weights can be viewed at: https://u.osu.edu/cnep/.
4 Respondents who report ‘Don't know’ on the three core efficacy items for both the 2014 and 2019 surveys range between 3 and 5 percent across all the items and are excluded from the analysis.
5 Respondents who report ‘Don't know’ if they voted are included into the ‘Did not vote’ category. This is a standard practice. If respondents cannot recall if they voted, it is most likely they did not vote but prefer not to be honest about their behaviour because of social desirability bias (Dahlgaard et al. Citation2019). Respondents who ‘prefer not to say’ are removed from the analysis.
6 In our study, the dichotomous turnout variable is cross tabulated with an ordinal variables (political efficacy) When one variable is dichotomous the other variable should determine the choice of the coefficient (De Vaus Citation2013, 262). Therefore, we treat both as ordinal measures which means that Gamma is the most appropriate summary statistic.
7 To check for the robustness of this finding we excluded all key predictors from the 2019 model and found that government performance remained statistically significant in the opposite direction.