The informational effect of candidate’s traits on voter behavior. A survey experiment for the municipal elections of Rome (Italy).

How to cite

Pinto, Gabriele; Garzarelli, Giampaolo; Galli, Emma. 2021. "The informational effect of candidate’s traits on voter behavior. A survey experiment for the municipal elections of Rome (Italy).." AEA RCT Registry. AEARCTR-0007763. https://doi.org/10.1257/rct.7763-2.2

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

Gabriele Pinto* , Giampaolo Garzarelli , Emma Galli

* Primary investigator

Registry ID
AEARCTR-0007763
DOI
https://doi.org/10.1257/rct.7763-2.2
First registered
Last updated
Status
Completed
Trial dates
2021-09-29 – 2021-10-16
Countries
Italy (Lazio (Rome))
JEL codes
C93, D72, O10, O50
Registry record
http://www.socialscienceregistry.org/trials/7763

Abstract

We evaluate the effect of providing voters with a set of information about candidates' curriculum. The information set is comprised of basic information derived from candidates’ curriculum (education, work, and political experience). The set mimics transparency requirements introduced by the recently introduced Italian law against corruption (“Legge spazzacorrotti”).
To evaluate the informational effect on voting behavior we organized a dedicated campaign on transparency of the candidates. The evaluation is made trough a panel survey experiment where we will expose treatment group individuals to the campaign while covering the campaign to control group individuals. The informational treatment is an invitation to visit a website with information about candidates.

Intervention

To evaluate the impact of political transparency on voting behavior we organized a dedicated campaign on transparency of the candidates. The campaign reconstructs information from all curricula vitae presented by all the candidates for transparency requirements. Since the information comes in the form of long curriculum vitae, we extrapolate the more salient characteristics (education, job experience, political experience) and structure this information in a tabular format. This information is then converted into a simple and light website where the visitor can see information about the candidates (HIDDEN). To evaluate the effectiveness of the campaign, we organized a panel survey experiment where we will expose treatment group individuals to the campaign while covering the campaign to control group individuals.
The sample of our panel survey experiment is comprised of N=1000 eligible voters in the *HIDDEN* area.
We will ask respondents to fill two questionnaires before and after the election day. In the first questionnaire (3-4 days before the election day) we will ask the demographic and attitude characteristics of the respondents. In the second questionnaire (a week after the election day) we will ask all respondents to report their vote choice. The core of the experiment consists of the submission of an informational treatment to a sub-sample of treated units T=500 that we will randomly select from the main sample.
The informational treatment is an invitation to visit a website with information about candidates through a link. This treatment is submitted just after the last question of the first survey.

Experimental design

A panel survey with two rounds before and after the elections.
Target:
a representative sample of 1000 voters in the Municipality of *HIDDEN* over the age of 18 with gender quotas. Interviews carried in CAWI.
Date:
The first phase of the municipal elections of *HIDDEN*. 3-4 October 2021
First Round:
To be held within 4 days before the election date.
The interviewees will be subjected to a questionnaire of about 18 questions, lasting about 6 minutes.
Treatment:
At the end of the questionnaire, 50% of the interviewees, chosen randomly, will receive via text message or e-mail a link to a website containing basic information about candidates, such as education, previous political experience, age, and criminal record.
Second Round
To be carried out within ten days following the date of the elections.
The interviewees will be subjected to a questionnaire of about 5 questions, lasting about 2 minutes. The questionnaire will also ask for the choice of the vote for the candidate for mayor and candidates for the city council

Design details

Randomization method
randomization done by survey software
Randomization unit
individual
Clusters
1
Observations
1000 +200 additional control interviewed only for the second questionnaire.
Arms
500 treated individuals, 500 control individuals
Primary outcomes
-Turnout -Expression of a preference for a candidate mayor/councilor -Characteristic of the candidate mayor/councilor voted -Consistency between characteristics of the ideal candidate and characteristics of the candidate actually voted (mayor and councilor)

Cited by

No citation data available for this registration yet.

Keywords