Back to History Current Version

Voting or abstaining in "managed" elections? An RCT in Bangladesh

Last registered on December 20, 2018

Pre-Trial

Trial Information

General Information

Title
Voting or abstaining in "managed" elections? An RCT in Bangladesh
RCT ID
AEARCTR-0003509
Initial registration date
December 14, 2018

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
December 20, 2018, 9:31 PM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Region

Primary Investigator

Affiliation
Monash University

Other Primary Investigator(s)

PI Affiliation
University of St.Gallen

Additional Trial Information

Status
In development
Start date
2018-12-03
End date
2019-10-30
Secondary IDs
Abstract
In many recent elections across the globe, the incumbent government has taken actions prior to or on election day to make its defeat very unlikely. We call such elections "managed" (or authoritarian) elections. We aim to understand the determinants of whether and how people vote in managed elections. In such elections, citizens who oppose the incumbent government have mainly two options: They can vote for an opposition candidate to signal their policy preferences to the regime; or they can abstain from voting to reduce the regime's post-election legitimacy. Our research questions are how the salience of these two options affect whether and how people vote in managed elections; and how this effect differs between areas in which the party of the current incumbent prime minister enjoyed strong support in the past and areas in which it did not. We study these questions in the general elections in Bangladesh that are scheduled for December 30, 2018. Given the current political climate, Bangladesh offers an ideal setting for studying voting behavior in managed elections.
External Link(s)

Registration Citation

Citation
Hodler, Roland and Asad Islam. 2018. "Voting or abstaining in "managed" elections? An RCT in Bangladesh." AEA RCT Registry. December 20. https://doi.org/10.1257/rct.3509-1.1
Former Citation
Hodler, Roland and Asad Islam. 2018. "Voting or abstaining in "managed" elections? An RCT in Bangladesh." AEA RCT Registry. December 20. https://www.socialscienceregistry.org/trials/3509/history/208212
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
There are two types of treatments. The first is the message that voting outcomes may affect policy-making (T-policy). The second is the message that high turnout may increase the political legitimacy of the government (T-legit).
Intervention (Hidden)
We deliver the two different messages, T-policy and T-legit, around one week prior to the general elections. We plan to deliver each message by the same combination of three channels: First, door-to-door visits to talk directly with randomly selected individuals. Second, stickers with these messages given to these individuals. Third, articles with these messages in a well-known (non-partisan) local newspaper that we distribute to these individuals.
Intervention Start Date
2018-12-23
Intervention End Date
2019-04-30

Primary Outcomes

Primary Outcomes (end points)
Individual voting participation coded based on the observation of an ink mark on the finger.
Official polling station-level data on turnout and the incumbent vote share.
Survey information about the voting participation of other voters within the respondents' family.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Changes between pre-treatment and post-vote surveys in the respondents' view on the value of democracy and their opinion about the role of members of parliament.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
We choose 150 villages that supported the party of the current incumbent government in past elections and 150 villages that supported the main opposition party. From within these two blocks of villages, we randomly choose 50 villages that receive treatment T-policy, 50 villages that receive treatment T-legit, and 50 control villages. We then randomly select 35-40 individuals per village, with roughly equal gender balance.
Experimental Design Details
The study area includes 4--5 constituencies containing around 800 polling stations (PS). Many of these PS consist of more than one village. Our sampling and randomization strategy contains the following stages:
1. We collect information on the vote share of the party of the current incumbent government (AL) and the main opposition party (BNP) at the PS-level in the general elections in 1996, 2001 and 2008. For PS consisting of multiple villages, we complement this information with local knowledge about the support for AL and BNP across villages within PS. We then choose 300 villages from the same number of PS such that we have 150 villages that supported AL in past elections and 150 villages that supported BNP.
2. We randomly divide both groups of 150 villages into 50 villages that receive treatment T-policy, 50 villages that receive treatment T-legit, and 50 control villages.
3. We randomly select 35-40 individuals per village, with roughly equal gender balance. These are the individuals we survey twice prior to the general elections and thereafter. In addition, we deliver the respective treatment messages to these 35-40 individuals in the treatment villages.
Randomization Method
Randomization done in office by a computer.
Randomization Unit
Village/polling station
Was the treatment clustered?
Yes

Experiment Characteristics

Sample size: planned number of clusters
300 villages
Sample size: planned number of observations
around 11,000 households/individuals
Sample size (or number of clusters) by treatment arms
150 villages that supported AL in past elections and 150 villages that supported BNP. We randomly divide both groups of 150 villages into 50 villages that receive treatment T-policy, 50 villages that receive treatment T-legit, and 50 control villages.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
Monash University Human Research Ethics Committee
IRB Approval Date
2018-10-29
IRB Approval Number
17640
Analysis Plan

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
Yes
Intervention Completion Date
December 29, 2018, 12:00 +00:00
Data Collection Complete
Yes
Data Collection Completion Date
January 07, 2019, 12:00 +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
302 villages, of which 154 are government villages and 148 opposition villages
Was attrition correlated with treatment status?
No
Final Sample Size: Total Number of Observations
11,843 individuals
Final Sample Size (or Number of Clusters) by Treatment Arms
52 government control villages, 51 government villages with policy treatment, and 51 government villages with legitimacy treatment. 48 opposition control villages, 50 opposition villages with policy treatment, and 50 opposition villages with legitimacy treatment.
Data Publication

Data Publication

Is public data available?
Yes

Program Files

Program Files
Yes
Reports, Papers & Other Materials

Relevant Paper(s)

Abstract
To study the effects of non-partisan information and get-out-the-vote (GOTV) campaigns on the partisan composition of the voting population in competitive authoritarian elections, we conducted a large-scale field experiment prior to the 2018 Bangladeshi general election. Our two treatments highlight that high turnout increases the winning party’s legitimacy and that election outcomes matter for policy outcomes. Both treatments increase turnout (measured by ink marks) in government strongholds but decrease turnout in opposition strongholds. We explain the withdrawal of treated opposition supporters and conclude that non-partisan information and GOTV campaigns can further tilt the uneven playing field in competitive authoritarian elections
Citation
Ahmed F., Hodler R., Islam A., 2024. Partisan effects of information campaigns in competitive authoritarian elections: Evidence from Bangladesh. Economic Journal, forthcoming

Reports & Other Materials