The Non-Working Wife as a Status Symbol: A Field Experiment in India

Please see the Experimental Design section.

2026-01-28

2026-08-31

Experiment 1: Gap in take-up of opportunities at home vs. outside the village (with co-travel)
Experiment 2: Take-up of the offered job opportunity

Take-up indicators are constructed directly from the willingness to take up questions in the survey (1 = “yes,” 0 = “no”).

In experiment 1, the respondent (husband) is asked whether he would choose one-week work or training opportunities for his wife, instead of an immediate cash payment of 50 INR, and we construct an indicator for choosing the opportunity instead of the cash. The work opportunity involves packaging oil bottles for a daily wage. The training opportunity involves learning healthy cooking practices, without any wages but instead involving free food ingredients which are valued similarly as the wages. All activities are conducted on a one-on-one basis with one female supervisor.

In experiment 2, respondents are offered the opportunity to take up a job, again instead of an immediate cash payment of 50 INR, under different circumstances (randomized by treatment) and again we will construct an indicator for choosing the opportunity instead of the cash.

Experiment 1: Gap in take-up indicators across all of the locations and/or conditions (e.g., an indicator for accepting the job opportunity at home minus an indicator for accepting the job opportunity at the clubhouse, or an indicator for accepting the job opportunity at the clubhouse minus an indicator for accepting the job opportunity outside the village.)

Constructed analogously to the primary outcomes.

We implement an incentivized choice experiment. Married men (husbands of women aged 18–45) are presented with a one-week job and training opportunities for their wives, which may be randomly implemented with some chance according to their responses.
The design systematically varies key dimensions of the opportunities:
Activity type: Paid work vs. training (with in-kind gifts instead of wages)
Location: At home, at a village clubhouse, or outside the village (with safe transportation provided)
Travel arrangement: Solo travel vs. travel with female companion(s)
Coordination: Job may be offered only if more than five households in the village agree to taking up the offer

Experiment 1: We will analyze how take-up of work and training vary across the following conditions.
a. Home: activity takes place at the respondent’s home, with female supervisors visiting.
b. Home with poster: Similar to the above. In addition, there will be posters put up in the village, describing the woman’s participation in the training or work activity at home.
c. Village clubhouse: activity takes place at a location inside the village, such as the village clubhouse where village meetings typically take place.
d. Outside village with co-travel: activity takes place at a workplace outside the village. Auto rickshaw pick-ups will be provided, allowing the wife to travel together with her female supervisor. Furthermore, they will travel with another woman from the village, who would be visiting the bank or block office near the workplace.
e. Outside village: activity takes place at a workplace outside the village. Auto rickshaw pick-ups will be provided, allowing the wife to travel together with her female supervisor.

Primary Hypotheses:
H1. Take-up of paid work offers will be significantly higher at home than outside the village (even with co-travel).
H2. The decreases in take-up when going from at-home to out-of-village opportunities will be larger for paid work than for training.
Secondary Hypotheses:
H3. Takeup of paid work will be lower outside the village (with co-travel) than in the clubhouse, and the decrease will be larger for paid work than it is for training.
H4. Takeup of paid work will be lower in the clubhouse than at home, and the decrease will be larger for paid work than it is for training.
H5. Increasing the wage paid for work will not change the fact that takeup of paid work falls off more when leaving the home than training (i.e., H2 will hold for the high payment treatment)
H6. Take-up of paid work offers will NOT be significantly higher at home without posters compared to when there are posters.
H7. The above hypotheses regarding paid work outside the village (H1-H3 and H5) will also hold for paid work outside the village without co travel.

Experiment 2: We will explore how allowing participants to coordinate in different ways before taking up jobs affects take-up. All of the offers allow for co-travelling, so the wife can travel with her female supervisor and at least one other woman from the village.
i. No coordination: Only one household in the village will get the job offer, and other people in the village will be informed that the respondent’s wife is the only woman in the village working in this job.
ii. Coordination with multiple workers: The job will take place only if more than five respondents have chosen the job offer. In addition, other people in the village will be informed that more than five respondents have chosen the job offer. The women from the households receiving the offers will all work at the same workplace but in separate rooms.

Primary Hypothesis:
H1. Offering households the opportunity to coordinate on accepting the job increases take-up more than no coordination.

More details are provided in the attached Pre-Analysis Plan.
.

Not available

All random assignments are generated by a computer program in the project office using a seed for reproducibility; allocations are revealed in person to all the respondents privately after all surveys in the village are completed.

The Experiment 1 questions are randomized at the village level, to only describe training scenarios or paid work scenarios. Hence all the households in a random set of villages would receive questions only regarding take-up of training activities at different locations, and in the remaining villages, households will only discuss take-up of paid work activities at the corresponding locations.

The second main set of questions will also be randomized at the village level, into either the no coordination question (E1) or the full coordination question (E2). Hence, all the households in a random set of villages will respond to the no coordination question and the rest will respond to the full coordination question.

More details on the Randomization are available in the Pre Analysis Plan.

Yes

170 villages

1700 households

About 90 villages for paid jobs, 70 villages for training, 10 villages for wage variation

Power calculations assume 90 paid villages, 70 training villages and 10 high wage villages, with 10 respondents per village. The MDE for detecting differences across training and paid work for all outcomes that represent the gap in take-up across locations (such as the MDE for a variable equal to an indicator for taking up of the home opportunity minus an indicator for taking up of the outside-village opportunity) fall in a narrow range of 0.11-0.12 SD. These magnitudes are small relative to the pilot differences we observed, indicating that the planned sample provides sufficient power (80%) to detect effects of policy-relevant size for our primary difference-in-differences outcomes. Similarly, for the coordination arm, the planned sample provides over 80% power to detect a minimum detectable effect of 23.3 percentage points in take-up.