Electoral Fraud, Voter Turnout, and Perceived Legitimacy

How to cite

Arbuzova, Anastasiia. 2024. "Electoral Fraud, Voter Turnout, and Perceived Legitimacy." AEA RCT Registry. AEARCTR-0013916. https://doi.org/10.1257/rct.13916-1.0

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

Anastasiia Arbuzova*

* Primary investigator

Registry ID
AEARCTR-0013916
DOI
https://doi.org/10.1257/rct.13916-1.0
First registered
Last updated
Status
Completed
Trial dates
2024-06-30 – 2024-09-30
Countries
Russian Federation ()
JEL codes
C93, D72, P00
Registry record
http://www.socialscienceregistry.org/trials/13916

Abstract

The use of electoral fraud as a tool for securing victory in elections is well-documented. However, its use for purposes that go beyond winning an election remains relatively understudied. Particularly puzzling is its prevalence in the context of elections in non-democracies, where electoral results are often manipulated despite the incumbent's guaranteed success. Why do autocratic leaders engage in fraud when victory is assured? In this project, I study improving perceived legitimacy as one of the reasons for engaging in fraud. I examine the following mechanism: if turnout influences the population's perception of the autocrat's legitimacy, then he might want to employ fraud to achieve a desired turnout value. I will call this the legitimacy hypothesis. First, by leveraging data from the 2016-2021 Russian parliamentary elections, I confirm the presence of electoral fraud. I focus on manipulations of turnout values and perform several statistical tests to detect those, including last digit distribution and the excess integer values test. After gathering suggestive evidence of the use of fraud to influence turnout, I proceed with testing whether knowledge of turnout actually affects perceived legitimacy. To answer this question, I propose undertaking a survey with a representative sample of the Russian voter population. The key component of the survey will involve a randomized information intervention where treated respondents receive information about outcomes of a hypothetical election. Subsequently, their evaluation of legitimacy will be measured.

Intervention

Each respondent will be randomly assigned to one of five treatment arms: Control (C), Turnout Low (TL), Turnout High (TH), Turnout Low + Result (RL), and Turnout High + Result (RH). Respondents in the Control group (C) will not be provided with any information about elections. Respondents in the Turnout groups (TL, TH) will be provided with information about the turnout, either Low or High. Respondents in the Turnout + Result groups (RL, RH) will be provided with information about the turnout, either Low or High, and leading party vote share.

Experimental design

1,600 participants will be enrolled to ensure a representative sample of Russian voters. Participants must satisfy two inclusion criteria: (1) adults (over 18 years old); (2) Russian citizens with voting rights. In addition, quotas on federal district, city size, gender, and age cohort, will be implemented to insure that the sample is nationally representative.
The survey questionnaire consisted of three sections and should take approximately 10 minutes to complete. The first section will address socio-demographic characteristics. The second section will include questions on various forms of political participation, such as signing petitions and voting. At the end of this section, respondents will be asked about their participation in the 2021 Federal Duma elections and whether they recalled the results at both the national (proportional system) and constituency (majoritarian system) levels. To elicit actual political preferences, respondents will be presented with a scenario in which elections were scheduled to occur the following day and asked about their willingness to participate and, if yes, which party would they vote for. Subsequently, respondents will exposed to a hypothetical outcome of these election results and their perceptions of government elected with these results will be measured.

Design details

Randomization method
Randomization in software
Randomization unit
Individual
Clusters
-
Observations
1600 individuals
Arms
320 Control (C), 320 Turnout Low (TL), 320 Turnout High (TH), 320 Turnout Low + Result (RL), 320 Turnout High + Result (RL)
Primary outcomes
Trust in government, approval and willingness to abide by hypothetical laws respondents do not personally approve of.

Cited by

No citation data available for this registration yet.

Keywords