Beliefs about Beliefs About Gender in India: Implications for Female Labor Force Participation and Time Use

How to cite

Schofield, Heather; Bertrand, Marianne; Dizon-Ross, Rebecca. 2023. "Beliefs about Beliefs About Gender in India: Implications for Female Labor Force Participation and Time Use." AEA RCT Registry. AEARCTR-0009518. https://doi.org/10.1257/rct.9518-1.0

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

Heather Schofield* , Marianne Bertrand , Rebecca Dizon-Ross

* Primary investigator

Registry ID
AEARCTR-0009518
DOI
https://doi.org/10.1257/rct.9518-1.0
First registered
Last updated
Status
In development
Trial dates
2022-09-01 – 2023-06-30
Countries
Private
JEL codes
J16, J22
Registry record
http://www.socialscienceregistry.org/trials/9518

Abstract

In this survey-based field experiment, we will first measure whether men in India have inaccurate beliefs about what other men around them view as appropriate or status-enhancing when it comes to women’s engagement with labor market work. We will use information treatments to randomly correct married men’s beliefs and measure how much this changes their responses to questions about their wives’ labor force participation. We will further tailor our research design to the Indian context to better understand whose beliefs might be particularly relevant to own behavior. In particular, we will contrast the effect of changing beliefs about one’s own Jati (caste) members with the effect of changing beliefs about upper caste members to test a Sanskritization hypothesis under which economically upwardly mobile lower caste Indians might be trying to emulate the upper castes as way to further legitimize their elevated status in society.

Intervention

We will provide men with information about other's beliefs about whether female labor force participations is "right/appropriate" or whether it alters a man's social status. The information provided will be further randomized to pertain to specific social groups: scheduled caste, their own jati (caste group), or upper castes. These treatments will be compared to two "active controls" providing irrelevant information (e.g. the fraction of working women preferring coffee to tea).

Experimental design

We begin by collecting data on men's beliefs about whether it is "right" for women to work and whether it influences one's social standing if your wife works from a large number of households (roughly 2x the number of intervention households). We also collect "beliefs about beliefs" from these individuals for specific religion and caste-based groups including one's own jati, upper caste individuals, and scheduled caste individuals within their state (i.e. 6 specific beliefs about beliefs crossing "right" and "social standing" with "own jati", "upper caste", and "scheduled caste").

The personal beliefs data is then used to generate accurate information on beliefs by social group (caste x religion x state), requiring a minimum of 30 individuals per group. We return to a subset of these households four months later and provide information about beliefs as described in the "intervention" description above. We also survey these households to 1) confirm whether beliefs have changed in response to the information, 2) collect data on their preferences about women's labor force participation (see primary and secondary outcomes above).

Finally, after another four months we collect additional data on men's preferences about women's female labor force participation and women's labor force participation via survey questions.

Design details

Randomization method
Randomization done by computer.
Randomization unit
Household. If there is more than one participating male in the household, all will receive the same information.
Clusters
25,000 households
Observations
35,000 men
Arms
Approximately 3,125 households in each of the 8 arms (including two controls).
Primary outcomes
We will create an index of survey outcomes indicating greater willingness to have one's wife participate in the labor force. The underlying questions include: 1) whether the husband would like his wife to work outside of the home in the coming year, 2) willingness to pay (hypothetical) for job training opportunities for women, 3) choice (hypothetical) among training opportunities ranging from cooking to job-training, and 4) whether the individual would want their wife to take a job at different wages and with varied amenities.
Secondary outcomes
Our secondary measures include survey data on: 1) job search behaviors, 2) factors the respondent considers in deciding whether to allow his wife to work, 3) the respondent's preferences about their daughter working prior to marriage. In addition, we anticipate collecting data on actual labor market behavior (e.g. time spent working) and survey measures of willingness to work in a later survey wave conducted 4 months after the intervention. The registry will be updated as these survey measures are constructed.

Cited by

No citation data available for this registration yet.

Keywords