Principal Investigators: Marcus Holmlund, Malte Lierl
Country: Burkina Faso

Registration: 20151120AA

Dates of Intervention: April – May 2016

Background: Voters and candidates in Burkina Faso have been faced with a highly unusual situation after the coup in 2014. Incumbent local governments were removed from office and new municipal elections were scheduled to be held in 2016. While the previous municipal elections were single-party dominated, previous national ruling party has now become one political force among many in a much more fragmented political landscape. As a consequence, local-level politicians are forced to (re-)position themselves in a changed environment. This raises a number of important questions: Can local politicians who were affiliated with the former party in power gain from switching parties? What candidate attributes can countervail the effects of being associated with the previous regime? Are there dimensions of candidate differentiation that are such strong determinants of voters’ preferences that any other attributes are irrelevant?

Research Design: The researchers carry out a survey experiment with voting-age citizens in 118 rural municipalities. Respondents are asked about their preference between two imaginary mayoral candidates, who are characterized by eight randomly varied attributes, along which the two candidates may or may not differ. Furthermore, respondents are asked about the extent to which their preference might be altered by additional information about the candidates, and about the types of information that would have the greatest influence over their preference. The survey experiment was designed to address two distinct sets of research questions. The first set of questions concerns the relationship between candidate differentiation and voters' receptiveness to information: (1) Are there dimensions of candidate differentiation which will cause voters to ignore further information? (2) What type of information will sway indifferent voters? (3) What is the (observational) relationship between voter indifference and intent to vote? The second set of questions focuses on the mechanisms by which local political support is sustained, specifically in the Burkinabe context. In order to explain the local-level impact of the national-level political transition in Burkina Faso, it is crucial to understand whether voting in local-level elections is partisan or personalistic. We will therefore investigate the following additional questions: (4) In the municipal elections, will voters turn away from local-level supporters of former president Blaise Compaore, even though many established local politicians were and continue to be affiliated with the CDP? (5) Can local-level supporters of former president Blaise Compaore gain from switching parties? (6) What dimensions of candidate differentiation can countervail an association with the previous regime?

Hypotheses:
  • The more enthusiastic a voter is about her most-preferred candidate/party, the greater is her motivation to vote, conditional on the voter’s indifference between candidates/parties.
  • The more indifferent a voter is between her most-preferred candidate/party and other candidates/parties, the lower is her motivation to vote, conditional on the voter’s enthusiasm for her most-preferred candidate/party.
  • Access to performance information increases pro-incumbent voting, if the incumbent’s performance is better than the voter had expected.
  • Access to performance information decreases pro-incumbent voting, if the incumbent’s performance is worse than the voter had expected.
  • If voters were previously uncertain about the incumbent’s performance, access to performance information about the incumbent causes voters to place greater weight on expected performance in forming their candidate preferences.
  • Personal invitations to municipal council/delegation speciale meetings reinforce the effect of performance information on vote choice.
  • Personal invitations to municipal council meetings cause voters to place greater weight on expected performance in forming their candidate preferences.
  • Personal invitations to municipal council meetings increase citizens’ interest in/attentiveness to performance information.
  • Personal invitations to municipal council meetings increase citizens’ motivation to vote.
  • Actual attendance of a municipal council meeting increases citizens’ motivation to vote.