Improving democratic accountability in Sierra Leone


Published 26.06.17

A study evaluates the influence watching electoral debates has on voter behaviour and finds that it contributes to political accountability

Eighty-four million people watched the first Presidential debate between Hillary Clinton and Donald Trump during the 2016 US presidential election. Whether it, or similar debates, changed how people voted is hotly disputed. Many political scientists believe that in an environment where people are constantly buffeted by new headline stories, the effect of debates on voting intentions fades quickly. But what about countries where there is very little information about candidates? Can debates help improve electoral accountability there? Or do debates focus attention on the wrong attributes, like physical attractiveness or who the voter would like to have a beer with?

These questions are important as we struggle to make democracy work for people. According to the Polity IV project, between 1970 and 2015 the number of democracies jumped from 35 to 96. Yet our faith in democracy has faltered. The ability to vote non-performing politicians out of power has not prevented corruption or poor public services.

Why isn’t democracy delivering better government?

A common concern is that voting is determined more by party, ethnic, or regional loyalty than by politician performance. This is not just a concern in poor countries: the US Democratic incumbent is almost certain to win in Boston and the US Republican incumbent almost certain to win in west Texas.

An alternative explanation is that voters can’t hold politicians to account because they have very little information about the candidates. Maybe some voters are voting for the party of their ethnicity/region/caste because they don’t have any other information to go on. In Sierra Leone, for example, voters are much more willing to vote across ethnic lines in elections for local councillors (where they know the candidates) than they are in elections for Members of Parliament (MPs).

Impartial information can change how people vote

Information from audits of local government finance in Brazil, and report cards on MPs in India (which showed how MPs spent their time and money) changed how people voted. But in many poor countries there is no objective information from independent sources that can be released. In these situations, candidates serve as the sources of information. They are asked about their priorities and experience: they debate.

The study: Sierra Leone and democratic accountability

Sierra Leone is one of the poorest countries in the world, public service provision is weak, and ethnicity is highly predictive of how people vote. The All People’s Congress wins seats in the North where Temnes predominate, while the Sierra Leone People’s Party wins seats in the South, the traditional homeland of Mendes.

Information is also scarce. In the rural areas where the study was conducted, 70% of the population had never been to school, 30% had no radio (TV and newspapers are extremely rare), just 3% knew how much money MPs are given to spend in their constituencies, and less than half could tell who was the incumbent. What’s more, there is no requirement for MPs to disclose how they spend the money they are given for their constituency so there is no information to fill out an MP report card.

Bringing electoral debates to rural voters

During the 2012 election, Search for Common Ground—an NGO with a local reputation for independent news coverage—organised debates between candidates for MPs in 14 constituencies. Candidates were asked about their priority for government spending, how they would spend their constituency funds, as well as their positions on topical policy issues. These debates were videotaped in the common national language, Krio, and a mobile cinema took the videos on a tour of local communities, often screening them on the external wall of a local school at night. Hundreds of people would crowd around and watch the debate screening, which was translated into local languages. In total, roughly 19,000 people watched the debates in the weeks before the election.

To evaluate the impact of the videotaped debates on voting, the mobile cinema visited 112 villages, randomly picked from a list of 224. People in all 224 villages were then surveyed on or just after election day.

The findings

1) Voter knowledge

Substantial impacts were found on voters’ knowledge about MP candidates and their policy positions thanks to the debates. People in the villages where debates had been screened were more likely to be able to name the candidates, know which one was more educated, who was the incumbent, and how much money MPs were given to spend. Knowledge of policy positions taken by candidates improved dramatically too: for example, 29% could name the first priority for spending by the SLPP candidate in debate areas, compared to 14% in the control villages.

2) Voter decision making

Voters were more likely to vote for someone who had the same policy priorities as they did. In addition, the candidate who was seen to have performed best in the debate got five percentage points more of the vote in debate villages than nondebate villages. Who performed better was determined by asking the audience at debate screenings as well as an independent panel of experts. There was striking agreement between the two groups about who was the best performer.

3) Voting patterns: Charisma or information

To understand what voters are responding to, another experiment was performed where some individuals were given only part of the information coming from the debates. In 40 large villages, individuals were randomly picked to watch the full debate on tablets. Others saw only the “getting to know you” introductory questions. This meant they could see which candidate had charisma but they learned nothing about their policies. A final group heard a “just the facts” summary of the debate. They did not hear the candidates themselves but learned about their policy positions and experience. In a survey at election time, all three groups knew more about the candidates’ qualities, such as their level of education, than the control group. Those who watched the debate on tablet or heard just the facts summary also knew more about the candidates’ policy positions. Only those who watched the debate, however, changed how they voted. Therefore, getting to know a candidate on its own and facts on their own are not enough, only the combination of the two changed voting patterns.

4) Informed voting and electoral accountability

Our results suggest that debates change the behaviour of politicians in office. Because the debates were held in 14 constituencies, randomly chosen from a list of 28, we could tell how participating in a debate changes performance. A survey team asked MPs how they spent the money given to them for their constituency, and then went to the constituency to find evidence of that spending. If the MP said they had paid school fees, the survey team asked the school principal how much of the fees had been paid. If the MP said they paid for road repairs, the community was questioned about how much was spent on the repairs and receipts requested. In treatment constituencies, evidence was found of two and a half times more money being spent than in control constituencies.

Don’t give up on democracy, improve it

When voters know little about the responsibilities and actions of elected officials it is not surprising that politicians under-deliver. But we should not give up on democracy. Instead we should find ways for voters to learn about their representatives. Debates appear to be an effective way for voters who may be illiterate and live in societies without freedom of information laws, to learn about politicians. Once voters are better informed, politicians may behave better.

Editor’s Note: This article is based on an IGC project

Photo credit: Marco Longari/Getty.