Experimental Design
Employer experiment
We recruit up to 600 "employers", individuals with hiring responsibilities from both businesses in Dhaka, Bangladesh, to be the primary subjects in our experiment. We also construct a worker pool consisting of up to 1,200 men and women aged 18 to 60 who responded to in-person recruitment drives on university campuses about an attractive one-day Excel workshop and job analyzing US stock market data live (7 pm-midnight in Bangladesh).
Each employer is assigned to one out of two transport conditions:
- Transport: Employers are informed that workers will remain in the office for the night shift and will be offered free safe transport home after the shift.
- No Transport: Employers are informed that workers will remain in the office for the night shift and are not informed about any transport.
In addition, employers are cross-randomized into one out of five subsidy conditions:
- No subsidy: Male and female workers receive BDT 1,500 (USD 15) for completing the shift. Employers receive BDT 500 (USD 5) if their hired worker is a man or a woman.
- Male worker subsidy: Male workers receive BDT 2,500 (USD 25) and female workers BDT 1,500 (USD 15) for completing the shift. Employers receive BDT 500 (USD 5) if their hired worker is a man or a woman.
- Female worker subsidy: Male workers receive BDT 1,500 (USD 15) and female workers BDT 2,500 (USD 25) for completing the shift. Employers receive BDT 500 (USD 5) if their hired worker is a man or a woman.
- Male employer subsidy: Male and female workers receive BDT 1,500 (USD 15) for completing the shift. Employers receive BDT 1,500 (USD 15) if their hired worker is a man and BDT 500 (USD 5) if their hired worker is a woman. This meant that employers would still earn more money from hiring male workers even if the male worker would complete zero tasks and the female worker would complete all tasks.
- Female employer subsidy: Male and female workers receive BDT 1,500 (USD 15) for completing the shift. Employers receive BDT 500 (USD 5) if their hired worker is a man and BDT 1,500 (USD 15) if their hired worker is a woman. This meant that employers would still earn more money from hiring female workers even if the female worker would complete zero tasks and the male worker would complete all tasks.
Employers make 10 hiring decisions between pairs of randomly selected potential workers and are informed that one of the decisions will be randomly selected to hire an Excel research assistant. For each potential worker, employers are provided the first name, gender, education, and Excel screening score.
Worker experiment
We recruit up to 1,000 "applicants", men and women aged 18 to 60 who responded to in-person recruitment drives on university campuses about an attractive one-day Excel workshop and job analyzing US stock market data live. To be able to measure selection into the night shift, applicants were only informed about the timing of the workshop and job (7 pm-midnight) after self-selecting into the experiment.
After completing an Excel screening test, applicants are informed that the workshop and job are from 7 pm-midnight and that hired workers receive an Excel certificate of completion and potentially a promotion benefit of BDT 500. To experimentally vary the perceived benefits and costs of the job to interested applicants, applicants are randomized into the following treatments:
1. Benefits:
- Low promotion rate: 10% of workers will be promoted.
- High promotion rate: 90% of workers will be promoted.
2. Costs:
- Transport: Applicants are informed about the safe transport home.
- No transport: Applicants are not informed about the safe transport home.
In both benefit treatment arms, the experimenters serve as the promoters and promote the best X% of workers.
After receiving detailed descriptions of the Excel workshop and job and completing the Excel screening test, we elicit applicants' reservation wages using the Becker–DeGroot–Marschak mechanism (Becker et al., 1964). We first ask applicants to choose their reservation wage and then randomly draw a wage between BDT 100 (USD 1) and BDT 5,000 (USD 5).