close× Call Us
close×
Title Presidential Candidate Debates in Malawi
Post date 04/26/2019
C1 Background and Explanation of Rationale Candidate debates are often important, and highly publicized, campaign events. Increasingly, governmental and non-governmental organizations are organizing them in new democracies around the world, with the goals of increasing voter knowledge and access to information about policy platforms and promises, promoting policy-based campaigning and voting (as opposed to clientelistic or identity based campaigning and voting), and enhancing political tolerance. Do candidate debates achieve these goals? To address this question, this project will conduct an experiment around presidential candidate debates organized in advance of Malawi's May 2019 elections.
C2 What are the hypotheses to be tested?

Family A: Hypotheses on vote choice and candidate evaluations

HA1: On average, voters exposed to debates will become more supportive of (or more likely to vote for) the three candidates that participated in the debates.

HA2: On average, voters exposed to debates will become less supportive of (or less likely to vote for) the incumbent president, who chose not to participate in the debates.

-- The effects predicted above will be less pronounced among strong supporters of the president's party (DPP): DPP supporters will punish the president less for not appearing in the debates and will not reward the challengers as much for participating

-- Will also test HA1 and HA2 separately for swing versus partisan voters

HA3: Voting across party lines: Debates make partisan voters more supportive of (or more likely to vote for) candidates from different political parties.

HA4: Voting across ethno-regional lines: Debates make voters more supportive of (or more likely to vote for) candidates that are associated with different ethno-regional groups.

HA5: Debate Performance: Participants will become more supportive of (or likely to vote for) candidates that perform well in the debate.

-- Will test whether effects above are different for partisans versus swing/independent voters.
-- If there is variation in performance ratings across the three different policy segment conditions, will also use this random source of variation in performance to test this hypothesis.

HA6: Debate Performance + Priors:
-- When candidates perform better than the voter expected them to before the debate, the debates increase support for that candidate.
-- When candidates perform worse than the voter expected them to before the debate, the debates decrease support for that candidate.
-- Will test whether effects above are different for partisans versus swing/independent voters.
-- If there is variation in performance ratings across the three different policy segment conditions, will also use this random source of variation in performance to test this hypothesis.

HA7: Policy importance: When participants are randomly assigned to view a policy segment in the policy area most important to them, the impact of debates (as predicted above) will be larger than when the participant is randomly assigned to view a policy segment that is less important to them.
-- Will test whether effects above are different for partisans versus swing/independent voters.

Family B: Hypotheses related to causal mechanisms

HB1: The impact of debates will be greater for participants who report learning a lot about candidates' policy positions

HB2: Debates impact perceptions about the strength of candidates' policy positions.

HB3: Debates make voters more confident that they understand candidate policy positions.

HB4: Debates improve voter knowledge of candidate policy positions.

HB5: Debates increase the chances that voters vote for candidates that share their policy preferences.

HB6: The impact of debates will be greater for participants who report learning a lot about candidates' quality (qualifications and trustworthiness)

HB7: Debates impact perceptions about candidate quality (qualifications and trustworthiness).

Family C: Hypotheses on additional outcomes

HC1: Voter turnout: Voters exposed to the debates will be more likely to turnout to vote.

HC2: Engagement: Voters exposed to the debates will be more engaged in the campaign and more likely to discuss politics with those in their family or social group.

HC3: Efficacy: Voters exposed to the debates will have a greater sense that they are informed and understand politics, and that their vote is meaningful.

HC4: Policy-based voting versus Identity or Clientelism:
-- Voters exposed to the debates will be more likely to report that candidate/party policies or platforms were important in determining their vote.
-- Voters exposed to the debates will be less likely to report that identity or handouts (clientelism) were important in determining their vote

HC5: Tolerance: Debates enhance political tolerance and reduce fear of out-group candidates

HC6: Legitimacy of the electoral process: Voters exposed to the debates will have more positive perceptions of the legitimacy and quality of the electoral and democratic process in Malawi.

Family D: Community-Level Hypotheses and Persistence of Effects

HD1: The effects of debates will persist more strongly when voters live in electorally competitive districts

HD2: The effects of debates will persist more strongly when voters have partisan diverse social networks.

HD3: The effects of debates will persist more strongly when voters live in ethnically diverse enumeration areas.

HD4: Effect of debates will dissipate most when voters live in a census enumeration area where the majority of the population is a coethnic of one of the main presidential candidates.

C3 How will these hypotheses be tested? * I am conducting a panel survey with 1,500 Malawians sampled from 12 districts across the country. Some of these participants are randomly assigned to view components of the presidential debates. Participants are surveyed before debate exposure, immediately after debate exposure, and then 6-8 weeks later in the weeks following the May 21, 2019 elections.
C4 Country Malawi
C5 Scale (# of Units) 1,500
C6 Was a power analysis conducted prior to data collection? Yes
C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval? Yes
C8 IRB Number George Washington University IRB #: 190971; Malawi National Committee on Research in the Social Science and Humanities (NCRSH) Protocol #: P.03/19/357
C9 Date of IRB Approval George Washington University: February 28, 2019; Malawian NCRSH: April 7, 2019
C10 Will the intervention be implemented by the researcher or a third party? The debates were organized by a taskforce of 12 Malawian media and civil society partners, supported by international NGOs and donors. The researchers are involved in the dissemination of opportunities to watch the debates to study participants.
C11 Did any of the research team receive remuneration from the implementing agency for taking part in this research? No
C12 If relevant, is there an advance agreement with the implementation group that all results can be published? Yes
C13 JEL Classification(s) not provided by authors