In dominant party regimes, voters—particularly those who favor the ruling party—may have little information about the quality of opposition candidates. To correct this informational imbalance, civil society organizations have devised a number of informational interventions aimed at providing candidate information to voters. The authors conducted a large field experiment in 11 constituencies in Uganda during the 2015 ruling party primaries and the 2016 general election for Members of Parliament. In Uganda, the National Resistance Movement (NRM) has held power since 1986 and controls the majority of elected positions across all levels of government. The remaining seats are held either by independent candidates or by members of a handful of opposition parties. Voters are generally poorly informed about candidates’ policy positions, education, occupation, religion, and even ethnicity. Given this lack of knowledge, partisan considerations are highly salient for vote choice. A recent opinion poll found that 75% of respondents would likely vote for the same party as they had in the previous election. Moreover, members of the ruling party are also less informed about opposition candidates relative to ruling party candidates. Informational asymmetries may be heightened by the fact that radio stations, a common source of news for Ugandans, are frequently owned by politicians.
Information about candidates was provided in video recordings in which candidates answered questions about policy preferences, qualifications for office, personal characteristics, and relevant experiences. The videos were edited to give the appearance of a debate in which all candidates answered one question in turn before moving on to the next question. The study was conducted in 11 constituencies, with constituency eligibility being determined by competitiveness, unlikelihood of violence, and other logistical factors affecting the feasibility of film screening. Constituencies from all four regions – central, western, eastern, and northern – were represented.
In a factorial design, each of 480 polling stations was randomly assigned to one of four treatment conditions: whether it was studied in the primary or the general elections, and whether it received a public screening of the debate-style video (treatment) or not (control). Randomization was blocked on constituency.
In a baseline survey conducted shortly before the intervention, the authors collected data on respondent characteristics and priors about candidates in the respondent’s constituency. Data on the primary outcomes of interest, voter-level turnout and vote switching, were collected through a phone survey on the evening of the election and subsequent days for all treatment and control groups. The treatment – the public screening of candidate videos – was conducted in a total of 240 polling station catchment areas. Approximately 50 to 100 people attended each screening, such that in total, an estimated 12,000 to 24,000 people have watched the Meet the Candidates videos.
This study was part of EGAP’s Metaketa Initiative (http://egap.org/metaketa), in which teams of researchers run coordinated evaluations of interventions across contexts that aim to impact the same causal mechanisms. This Metaketa round focused on Information & Accountability (http://egap.org/metaketa/metaketa-information-and-accountability) by implementing a series of experimental projects that assess the role of information in promoting political accountability in developing countries.
Treatment effects on both turnout and vote switching were similar across the two electoral environments. Across both elections, those who learned that their intended vote choice performed worse in the video than their opponents were more likely to stay home on election day. Furthermore, voters in both elections responded to information about candidate quality by switching away from their intended vote choice. In particular, voters in the general election who watched the videos were more likely to switch away from the ruling party candidate and toward opposition candidates. The prevalence of switching away from the ruling party was greatest among ruling-party voters whose preferred candidate had lost the party primary. In the primary election, voters switched to the candidate who was rated by local experts as having performed best in the video. Overall, the intervention led voters to view candidates representing parties they did not initially favor as more likable and reduced uncertainty about opposition candidates.
Overall, the results suggest that provision of information about all candidates may be effective in reducing uncertainty about opposition candidates, thus leveling the informational playing field, even among partisan voters. In a dominant party setting, informational asymmetries can further reduce electoral competitiveness. These findings suggest informational interventions in the form of candidate debates have the potential to reduce those asymmetries, especially in the case of voters who supported non-ruling party candidates who lost in party primaries.