Fault Lines

How to cite

Raffler, Pia; Martin, Lucy. 2015. "Fault Lines." AEA RCT Registry. AEARCTR-0000767. https://doi.org/10.1257/rct.767-2.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Pia Raffler* , Lucy Martin

* Primary investigator

Registry ID
AEARCTR-0000767
DOI
https://doi.org/10.1257/rct.767-2.0
First registered
Last updated
Status
Completed
Trial dates
2014-10-31 – 2015-12-31
Countries
Uganda ()
Registry record
http://www.socialscienceregistry.org/trials/767

Abstract

Models of electoral accountability typically assume a nested principal-agent problem in which voters hold their representatives accountable not only for their own actions but also for the performance of the bureaucracy. In developing countries such as Uganda, the limited capacity of elected officials at the local level means that they cannot always control the behavior of their appointed counterparts; we argue that this may have a serious impact on how well elections serve as an accountability mechanism. This paper outlines the pre-analysis plan for a survey experiment conducted with citizens in Uganda that examine how citizens allocate blame and credit for the provision of public goods at the local level and how this affects electoral accountability. The survey experiment is the third in a series of three, with the first two experiments sampling elite citizens and local government officials in Uganda, respectively. Data from the other two experiments has already been analyzed. The authors do not yet have access to the data from the survey experiment with citizens.

Intervention

The study is a survey experiment. The intervention consists of different vignettes which randomly vary whether (a) a local bureaucrat comes from the same or another district than where she works and (b) whether service provision is of good or low quality and (c) whether the local bureaucrat or the local politician are primarily responsible for the quality of service delivery. The vignettes are described in greater detail in the attached pre-analysis plan.

Experimental design

The study is a survey experiment. The intervention consists of different vignettes which randomly vary whether (a) a local bureaucrat comes from the same or another district than where she works and (b) whether service provision is of good or low quality and (c) whether the local bureaucrat or the local politician are primarily responsible for the quality of service delivery. The vignettes are described in greater detail in the attached pre-analysis plan.

Design details

Randomization method
Randomization done by ODK on a smart phone.
Randomization unit
The unit of randomization is the individual survey respondent.
Clusters
6,000 - 7,000 respondents.
Observations
6,000 - 7,000 respondents.
Arms
Each final cell is expected to have between 750 to 875 respondents.
Primary outcomes
1) Whether voters hold the bureaucrat or the politician responsible for the quality of service delivery prior to having received information. 2) Expected changes in vote shares of the politician 3) Perceived effectiveness of electoral sanctions/rewards 4) Perceived quality of official (index) Secondary: 5) Expected likelihood of informal sanctions/rewards 6) Perceived effectiveness of informal sanctions/rewards

Cited by

No citation data available for this registration yet.

Keywords