Reducing the Gender GAP towards a More Representative Democracy

Last registered on September 26, 2024

View Trial History

Pre-Trial

Trial Information

General Information

Title

Reducing the Gender GAP towards a More Representative Democracy

RCT ID

AEARCTR-0014405

Initial registration date

September 20, 2024

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

September 26, 2024, 12:27 PM EDT

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

Locations

Country

Brazil

Region

Primary Investigator

Name

Frederico Finan

Affiliation

UC-Berkeley

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Amanda de Albuquerque

PI Affiliation

a ponte

Contact Investigator

PI Name

Anubhav Jha

PI Affiliation

Princeton University

Contact Investigator

PI Name

Laura Karpuska

PI Affiliation

Insper

Contact Investigator

PI Name

Francesco Trebbi

PI Affiliation

UC-Berkeley

Contact Investigator

Additional Trial Information

Status

In development

Start date

2024-09-28

End date

2024-10-30

Keywords

Electoral, Gender

Additional Keywords

JEL code(s)

P00, J16, D72

Secondary IDs

Prior work

This trial does not extend or rely on any prior RCTs.

Abstract

The underrepresentation of women in politics is a serious concern throughout the world. Official data from the Inter-Parliamentary Union (an international organization of national parliaments of 179 countries) show that the global participation rate of women in national-level parliaments is 26.4%, and only 5 countries have 50% or more of seats in the lower house occupied by women.

Brazil, one of the largest democracies in the world and the site of this project, is no exception. Currently, women occupy only 14.8% of the seats in the lower house of the Brazilian Congress. Moreover, 955 out of 5,565 municipalities did not elect a single woman to their local chamber in the last election, thus leaving over 10.2 million Brazilian women without political representation at the local level.

The goal of this project is to study the demand-side factors that affect voting for female candidates in Brazil. Most standard voting models assume that voters evaluate candidates along several dimensions when deciding for whom to vote. These dimensions often include the candidate’s Gender, their perceived Ability, and their Policy positions, (GAP). Based on these three dimensions, we have developed a statistical model to decouple their effects on voting decisions. Importantly, our framework can separately identify the roles of preferences versus beliefs of individual voters in this multidimensional decision. A key empirical contribution of this work will be its ability to disentangle which dimensions matter the most for voters when choosing a candidate and how gender (both of the candidate and of the voter) plays a role in this choice. Campaigns to promote gender representation require a clear understanding of the roots of underrepresentation to be effective.

To estimate our model, we have designed a randomized control trial (RCT) that exploits Instagram’s capabilities to micro-target individual voters with messages in the form of ads. The use of the Instagram platform is an exciting feature of the project, as it has become commonly used by political parties for their own political campaigning. As we describe in more detail below, our experiment will consist of several gender-specific treatments that will vary in their level of informativeness. This distinction in informativeness across messages is key for not only isolating changes in salience and beliefs about a particular candidate’s dimension (e.g., gender, ability, and policy), but it will also allow us to recover any learning effects on the voter side without relying on survey-based methods. We will run our experiment in the weeks just prior to this year’s October elections in Brazil, at a geographical scale sufficiently large to detect effects on aggregate vote shares.

External Link(s)

Registration Citation

Citation

de Albuquerque , Amanda et al. 2024. "Reducing the Gender GAP towards a More Representative Democracy ." AEA RCT Registry. September 26. https://doi.org/10.1257/rct.14405-1.0

Sponsors & Partners

Experimental Details

Interventions

Intervention(s)

Intervention (Hidden)

Intervention Start Date

2024-09-29

Intervention End Date

2024-10-06

Primary Outcomes

Primary Outcomes (end points)

Vote share at the level of the municipality

Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)

Secondary Outcomes (explanation)

Experimental Design

Our experiment will consist of a control group and 5 treatments assigned at random across 1000 geographical units. The assignment of our treatments will target independently both male voters and female voters. The size of the geographical units (i.e., a municipality) will depend on Instagram’s penetration rate and the level of confidence of Instagram’s users’ locations. The specific treatments are as follows:

1. Gender identity: Did you know that women make up more than 50% the population, but they represent only 16% of the National Congress? Around the world only 27% of parliamentarians are women.* Who deserves your vote in these elections? A female candidate or a male candidate? Think carefully about this.
* Inter-Parliamentary Union. Women in national parliaments, as of 1st January 2024

2. Policy noninformative: What is important to you in this election? Education, health care, child welfare? Vote for candidates who truly defend what is important for you every day.

3. Policy informative: What is important to you in this election? Education, health care, child welfare? Did you know that studies show that female parliamentarians invest 77% more on childcare*, education, and health care** than male politicians? Vote for candidates who truly defend what is important for you every day.
* K. A. Bratton and L. P. Ray. 2002. “Descriptive representation: Policy outcomes and municipal day-care coverage in Norway,” American Journal of Political Science, 46(2), pp. 428–437.
** R. Chattopadhyay and E. Duflo (2004). “Women as policy makers: Evidence from a randomized policy experiment in India,” Econometrica 72(5), pp. 1409–1443;
**Gerrity JC, Osborn T, Mendez JM. 2007. Women and representation: a different view of the district? Polit. Gender 3:2179–200
Note on Bratton and Ray: +77% comes from the sentence: “The model predicts that 16.10 gender representation to predict departures from the level of child-care coverage at the beginning of the electoral percent of children will be covered if the council is entirely male; 28.53 percent of children will be covered in the most gender diverse councils.” (p.434)

4. Ability noninformative: What is important to you in this election? Competent and qualified politicians who work hard to improve your local government and community. Vote for candidates who meet the quality you demand.

5. Ability informative: What is important to you in this election? Competent and qualified politicians who work hard to improve your local government and community. Did you know that studies show that, on average, female politicians are of higher quality*, more competent, and work harder** than their male counterparts? Vote for candidates who meet the quality you demand.
* Baltrunaite, Audinga, Piera Bello, Alessandra Casarico, and Paola Profeta. "Gender quotas and the quality of politicians." Journal of Public Economics 118 (2014): 62-74.
**Anzia, Sarah F., and Christopher R. Berry. "The Jackie (and Jill) Robinson effect: Why do congresswomen outperform congressmen?" American Journal of Political Science 55, no. 3 (2011): 478-493.

Experimental Design Details

Randomization Method

Randomization done in office by computer

Randomization Unit

These municipios were divided into 10 stratums. These stratums were created using the covariates: "population","literacy rate","worst case penetration","small smpl flag", "lagged share of votes for females in congress", "proportion of pardos", "proportion of pretos", "proportion of indigena", "proportion covered by internet", "gdp per capita","schooling","urban population","proportion aged 20 to 44","proportion aged 45 to 64","proportion aged 65plus".
For each stratum, we randomly assigned treatment status. There are 10 blocks and 7 treatments, each pair block X treatment roughly have 14-15 observations.

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

1000

Sample size: planned number of observations

1000

Sample size (or number of clusters) by treatment arms

T1 (Gender Male)= 143
T2 (Gender Female)=142
T3 (Policy Uninformative, Male and Female) =142
T4 (Policy Informative, Male and Female)=142
T5 (Ability Uninformative, Male and Female) =145
T6 (Ability Informative, Male and Female) =145
Control= 141

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Supporting Documents and Materials

IRB