Secret Keeping and Jobs for Women

How to cite

Lowe, Matt; McKelway, Madeline. 2016. "Secret Keeping and Jobs for Women." AEA RCT Registry. AEARCTR-0001678. https://doi.org/10.1257/rct.1678-2.1

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

Matt Lowe* , Madeline McKelway

* Primary investigator

Registry ID
AEARCTR-0001678
DOI
https://doi.org/10.1257/rct.1678-2.1
First registered
Last updated
Status
Completed
Trial dates
2016-09-18 – 2017-08-31
Countries
India (Uttar Pradesh)
Registry record
http://www.socialscienceregistry.org/trials/1678

Abstract

Female labor force participation is low and falling in India. This pattern is concerning given the growing evidence that jobs empower women, and the particular need for such empowerment in India, where norms of patrilocality, patrilineality and dowry contribute to pervasive under-investment in women. In this project we explore two sides of the story, with a two-stage experiment. For both, we partner with a large carpet manufacturing firm that is recruiting rural women for weaving positions in a region where weaving is a male-dominated activity. We hypothesise that men often control the decisions of their wives, preventing them from enrolling against their wishes. In the first stage of our experiment, we explore whether enrollment decisions are affected by whether the husband or wife is directly canvassed, and in turn whether this is affected by amount of information known by the "non-canvassed" spouse. In the second stage, we randomise which women get positions amongst those that enroll. This way, we can also explore the effect of the job on longer-run income, labor supply and empowerment.

Experimental design

The first stage is a take-up experiment. Amongst eligible married women we randomise whether the husband or wife is given the job information and ticket required to enroll. To explore mechanisms behind any effect of who gets the ticket, we cross-randomise whether the "non-ticketed" spouse (a) doesn't know about the ticket (with the "ticketed" spouse having "plausible deniability"), (b) knows about the ticket but doesn't receive job information, (c) knows about the ticket and receives job information, or (d) meets together with the "ticketed" spouse to receive job information, and to have a three-minute discussion about the opportunity. This set-up permits us to explore both information with-holding and psychological mechanisms behind any effects of who gets the ticket.

Conditional on over-subscription in the first stage, we will have a second stage randomisation in which we randomise which enrolled women are actually offered the job training. These women will also be offered weaving work beyond the training period, conditional on good performance. We then plan to return in 1-2 years to evaluate effects on female (and spouse's) labor supply, income, empowerment and attitudes toward gender norms of each spouse.

Design details

Randomization method
Randomization done in office by a computer.
Randomization unit
Pair (either eligible woman and her husband, or eligible unmarried woman and her head of household.) This is equivalent to having randomisation at the eligible woman-level.
Clusters
For first stage: 1000 eligible women (will depend in particular on whether our timing fits that of our partner's plans to open women weaving centres. If our partner decides to stop opening centres (e.g. due to low interest), then we will also have to stop). For second stage: 300 women enrolled.
Observations
For first stage: 1000 eligible women (plus data for their "pairs"). For second stage: 300 women.
Arms
For first stage: 800 married women with 100 in each treatment group (husband vs. wife gets ticket, cross-randomised with (a)-(d)). 200 unmarried women with 100 in each treatment group (woman vs. head of hh gets ticket). For second stage: 140 offered jobs (7 training centres with 20 slots each), 160 control group (put on waiting list).
Primary outcomes
For the take-up experiment, the primary outcome variable is whether the eligible woman enrolled. Secondary outcome variables include those from a knowledge quiz carried out over the phone (to test directly for information flows/with-holding between spouses). For the second-stage jobs experiment our primary outcomes will be (a) labor supply and income of the eligible woman, (b) labor supply and income of her spouse (if married), (c) empowerment measures of the eligible woman, and (d) gender attitudes of both the eligible woman and her spouse (if married) (for example, on appropriateness of women working outside the home).

Cited by

No citation data available for this registration yet.

Keywords