Gender Norms at Work: Impacts on Women’s Hiring and Workplace Experiences in Bangladesh

Our intervention is a gender norms-oriented hands-on curriculum that was developed by incorporating elements from BRAC’s current gender equality and skills development programming with elements from current best practices (CARE, 2014; Dhar et al., 2022). The activities include the following,
a) Workshops for employees and managers to cover key topics: gender equality and stereotyping, identifying and combatting workplace sexual harassment, promoting teamwork and cooperation, and fostering a woman-friendly workplace.
b) Workshop for managers on facilitating a gender-sensitive workplace, creating and communicating workplace protections for women and supporting women to increase job retention and career advancement.
c) Workshop with managers (and supervisory staff) to develop an “action plan” and targets to increase female representation in the workplace and promote a gender-inclusive environment.

2024-02-15

2024-08-30

Primary Outcomes (End Points): Gender attitudes, Gender productivity, Cooperation, workplace harassment / discrimination, employee turnover and recruitment of female workers

Primary Outcomes (Explanation): Our primary outcomes are gender attitudes, measured with three indices: a gender attitudes index ("gender index"), a gender productivity index ("productivity index"), and a workplace harassment/discrimination index ("discrimination index"). We use the swindex function in Stata (Schwab et al., 2021) to calculate indices, which calculates standardized weighted indices based on Anderson (2008), normalized to the control group mean. We calculate indices separately for owners/managers and workers.

Trust, sense of belonging, job satisfaction, women-friendly work environment, attitudes regarding violence against women, subjective expectations and mental health. Moreover, we will collect data on perspective-taking, empathy, manager-worker relationships, social connections, and gender segregation as our potential mechanisms

We have conducted a census across 27 districts in Bangladesh to identify eligible firms for our research. From the initial group of 4,754 firms, we selected 2,200 for the baseline survey which took place in October 2023, and the BIGD survey team completed the baseline survey with 1,888 firms. We randomized markets into a treatment and control group with equal probability, stratifying on division and the gender composition of the firms in the market. Specifically, whether in the market (1) all firms had only female employees, (2) all firms had only male employees, (3) all firms had both male and female employees; (4) firms had different compositions and/or no permanent employees. We only stratified by division in Dhaka and Chittagong because we had relatively few firms in those divisions. We will complement surveys with the lab-in-the-field experiments to measure manager and employee gender attitudes. We will implement several lab-in-the-field experiments (public goods game, trust game, ultimatum Game, prisoner’s dilemma game and a novel productivity experiment). We will then measure the impact of this intervention on gender attitudes and productivity in the short and longer term, 3 and 15 months after the end of the training.

Not available

Randomization was done in the office using the Stata software package

Market

Yes

807 Markets

1888 Firm owners / managers and 3207 workers

403 markets (971 firms and 1643 workers) in the control group and 404 markets (917 firms and 1564 workers) in the treatment group

We calculate the minimum detectable effect size with a 5% significance level at 80% power, based on an index of gender attitudes standardized to the control group, our primary outcomes. We calculate the Intra Cluster Correlation Coefficient (ICC) to be 0.1014 (for firm owners) and 0.1858 (for workers) based on the baseline gender attitudes index using the loneway command in Stata. We surveyed 807 markets and randomly allocated 403 markets to the control group and 404 markets to the treatment group. Based on the market-level randomization, we have 971 firms in the control group and 917 firms in the treatment group. Thus, on average we have 2.41 firms per market in the control group and 2.27 firms per market in the treatment group. Thus, we have an MDE of 0.14 standard deviations (s.d.) at 80% power for owners. Moreover, we have around 4.1 workers per market in the control group and 3.9 workers per market in the treatment group. Thus, we have an MDE of 0.12 Standard Deviations (s.d.) at 80% power for workers in our sample.