Differential Communication: Information Pooling within the Household

How to cite

Mustafi, Priyoma. 2024. "Differential Communication: Information Pooling within the Household." AEA RCT Registry. AEARCTR-0012429. https://doi.org/10.1257/rct.12429-1.0

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

Priyoma Mustafi*

* Primary investigator

Registry ID
AEARCTR-0012429
DOI
https://doi.org/10.1257/rct.12429-1.0
First registered
Last updated
Status
On going
Trial dates
2023-08-22 – 2024-11-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/12429

Abstract

Different household members often have access to independent information through their personal experiences and social networks. Recent field evidence shows that husbands and wives don’t always learn from each other (Conlon et al., 2022; Fehr et al., 2022). Using a lab-in-the-field experiment in India, this project studies information diffusion between married couples when the information set is gendered and non-gendered. To study the consequences of (im)perfect information flow within married couples on household decision-making when deciding who to allocate decision-making power to, based on information aggregation. Gendered societies lead to gendered roles and decision-making power within households, leading to different comparative advantages. Thus this project also studies information diffusion when the information set is gendered i.e, how individuals treat their own and their spouse's information in gendered domains, i.e., does one over-correct or under-correct depending on whether they have a comparative advantage in a domain.

Intervention

Both husbands and wives have the same urn containing red and white balls, and they are individually asked to guess the total number of red balls in their urn, twice. Before the first guess, each spouse draws some balls from the urn (with replacement) privately, to learn about the contents about the urn. Based on this information, each spouse privately guess the number of red balls. Before making a second guess, each spouse can either draw balls themselves again or learn about their spouse's first set of draws via a discussion.

In the second part of the experiment, after participants finish the ball task, they are shown two baskets of goods- male type and female type. The spouses privately guess the prices for each basket type. This first private guess is without any consultation with their spouses. After a discussion with one's spouse, they again privately guess the prices for each basket.

Experimental design

We will recruit married couples in Kolkata, India. Each couple plays an experimental game in which their primary objective is to guess the number of red balls in an urn containing 20 balls, each of which is either red or white. Individuals are told that the number of red balls in the urn can between 4 and 16, thus the distribution is known.

Each spouse individually draws a set of balls to learn about the contents of the urn and then makes their first guess. Before making their second guess they are again able to learn about the contents of the urn. The source of this second signal varies across two between subject treatments as follows:
Individual Treatment: In this treatment, each spouse again draws some balls (with replacement) a second time privately.
Discussion Treatment: In this treatment, individuals discuss with their spouses, for however long they want. Thus, the source of a second signal will be one’s spouse, instead of a second set of draws. Comparing the weight one puts on information that they have gathered themselves or learn from their spouse, will allow me to test whether the information flow from wives to husbands is indeed worse than the information flow from husbands to wives. Additionally I can then test whether the underweighting or overweighting your spouse's information is related to actual or perceived relative ability difefrences.

To examine the consequences of asymmetric learning on decision-making power, after the second guess, individuals are presented with a surprise opportunity to choose a decision-maker (DM). Each participant is told that a DM is one whose second guess will be counted for additional payment for the couple. The treatment variation before second guess allows me to explore if individuals can account for asymmetric learning when deciding how to allocate decision-making power, i.e., does the failure of perfect information agregation also translate to how couples decides who makes decisions based off of the same information.

In the second part of the experiment, after participants finish the ball task, they are shown two baskets of goods- male type and female type. The spouses privately guess the prices for each basket type. This first private guess is without any consultation with their spouses. Performace or ability o this guess would establish the comparative advantage. After a discussion with one's spouse, they again privately guess the prices for each basket. I am again interested in how much weight you put on your spouse's information when its the gender incongrunt domain and additionally how one adjusts their guesses after a discussion. With this price task, the existing expertise gives rise to well defined beliefs about ability, which guides who you should listen to as you can actually learn something useful from that person. Thus I will also collect beliefs for the different product types, to test that one was indeed aware of the comparative adavantge.

Design details

Randomization method
Randomization is done on stata
Randomization unit
Married couple level
Clusters
400 couples
Observations
400
Arms
200 couples in each treatment
Primary outcomes
Second guesses made in ball-in-urn task and pricing task. The expected payoff of guesses in ball-in-urn task, calculated relative to a Bayesian benchmark DM choices by husband and wife
Secondary outcomes
Beliefs about ability on both ball-in-urn task and price task

Cited by

No citation data available for this registration yet.

Keywords