With Marjan Hossain, Sheikh Touhidul Haque
International Journal of Educational Development, 2025, Online Appendix
This study investigates the impact of corporal punishment at home on children's foundational learning skills in Bangladesh, a country where nearly 89% of mothers/ primary caretakers report using physical or psychological punishment at home. Using data from the 2019 Multiple Indicator Cluster Survey (MICS), we employ a two-stage least squares (2SLS) instrumental variable approach to isolate the causal effects of corporal punishment, leveraging the implementation period of Bangladesh's "2011 Child Law" act as an instrument. We find that exposure to corporal punishment at home significantly reduces the likelihood of children possessing foundational reading and numeracy skills. Male children are disproportionately affected, highlighting potential gender-specific mechanisms through which corporal punishment may affect learning outcomes. Our findings point to two policy imperatives: (1) stronger enforcement of existing bans on corporal punishment and (2) early-childhood interventions to address its long-term consequences.
Belief Measurement: The Role of Survey Incentives
With Atiya Rahman
This study investigates whether monetary incentives affect the accuracy of reported beliefs about social norms surrounding girls’ education and marriage in rural Bangladesh. Using a randomized survey experiment with 2,914 fathers, we test whether informing respondents of a financial reward for accurate peer belief guesses alters their responses. Initial analyses suggest that incentives modestly reduce misperceptions on specific outcomes—such as beliefs about financial support from daughters and early marriage prospects. However, these effects lose statistical significance when adjusted for multiple hypothesis testing using Anderson’s sharpened q-values and Romano-Wolf Family-Wise Error Rate (FWER) corrections. Heterogeneity analyses reveal suggestive patterns among poorer, older, and less-educated respondents, but most of them too are not robust to multiple testing adjustments. Overall, the findings indicate that monetary incentives yield limited and statistically fragile improvements in belief accuracy, raising questions about their effectiveness and cost-efficiency in sensitive social norm research.
Perception Versus Reality: Social Media, Elopement Fears, and Marriage Decisions in Rural Bangladesh
With Munshi Sulaiman and Shaila Ahmed
Parental concern about the exposure of their children to social media is almost universal. This has an added social dimension in a country like Bangladesh, where it is sometimes linked with risks of elopement by adolescents that limit access to digital tools and possibly contribute to child marriage out of fear of shame. Using baseline and midline data from a randomized controlled trial (RCT) of the Shopno Sharothi (SS) program in rural Bangladesh, this study examines three interrelated questions. First, we test whether daughters’ social media usage is associated with parents’ perceived risk of elopement. Baseline results show a strong positive association: parents of social media users are about 45\% more likely to perceive a higher risk of elopement, though this relationship weakens over time, partly due to selection and treatment effects. Second, we explore the relationship between perceptions and actual outcomes one year later to find that social media use significantly increases the likelihood of a girl marrying her own chosen man. Finally, we assess whether program interventions mitigate these risks of elopement. The intervention substantially reduces overall marriage rates, particularly among girls who use social media and whose parents express elopement concern at baseline, while showing smaller effects for other subgroups. These findings highlight the dual nature of digital exposure—empowering girls yet intensifying parental fears—and emphasize the importance of interventions that integrate digital literacy, parental engagement, and women empowerment.
Optimizing Cash Transfers with Social Empowerment Interventions to Reduce Child Marriage in Bangladesh
With Munshi Sulaiman, Shaila Ahmed, Khandker Wahedur Rahman, Sakib Mahmood.
Midline is completed; Endline survey is scheduled in December 2025
Play-Based Learning in Crisis Contexts: A Randomized Controlled Trial on Educational and Psychosocial Outcomes in Host Community Children
With Shaila Ahmed, Justine Howard, Sakila Yesmin, Muhammad Towkir Hossain.
Baseline survey is completed; Endline survey is scheduled in September 2025
“Partnership for Research on Progress and Resilience in Education (PREPARE): Round 1 Policy Brief” with Marjan Hossain, Khandker Wahedur Rahman, and Shaila Ahmed. 2023. BRAC Institute of Governance and Development (BIGD) Report. Dhaka, Bangladesh. [PDF]
"Unravelling the Association between Shocks and Education in Bangladesh", with Marjan Hossain, Radhika Nagesh, Shaila Ahmed, and Khandker Wahedur Rahman. Center for Global Development, July 11, 2023.
The Learning Crisis Persists in Bangladesh: Findings from a Two-stage Study", with Marjan Hossain, Radhika Nagesh, Shaila Ahmed, and Khandker Wahedur Rahman. Center for Global Development, Nov 29, 2023.
This section includes the projects that I worked or am working on only as a (Senior) Research Associate
Context
The education system in Bangladesh is highly susceptible to shocks, particularly affecting children in climate-vulnerable regions. These children face direct consequences such as the destruction or severe damage of school infrastructure during major calamities, as well as indirect consequences such as economic strain on their households, loss of livelihoods, and inadequate nutritional support. With schools being closed for 18 consecutive months, the COVID-19 pandemic only worsened the negative effects on children’s education and learning progression. Even before the pandemic, Bangladeshi children had poor educational outcomes. The prolonged disengagement from learning during the pandemic further raised concerns about learning gaps or even learning loss. Unfortunately, there is limited data on children’s learning status (foundational and grade-level) for both pre-and post-pandemic periods, with most assessments conducted solely within schools (only foundational tests were done at household levels). Data on children’s learning outcomes for the period of school closures are even scarcer.
This study utilized phone-based assessments to measure foundational literacy and numeracy skills among children aged 5-18. The findings reveal the negative effects of the pandemic on enrollment status and learning outcomes, exacerbated by the susceptibility of the education system to shocks in climate-vulnerable regions. Limited pre-pandemic data and scarce information on learning outcomes during school closures highlight the need for additional support and interventions to address learning gaps and ensure children reach their grade-level potential.
Objectives
The primary goal of the study is to assess the current state of foundational literacy (Bangla and English) and numeracy among children aged 5–18 years. PREPARE aims to understand the impact of shocks like COVID-19 and natural disasters on children’s education prospects and outcomes in Bangladesh.
Methodology
The study was conducted using the Annual Status of Education Report (ASER) tool, adapted for Bangladesh, over the phone. The sample used in this study was drawn from the Back to School (B2S) survey, a nationally representative household survey conducted in October 2021. The researchers randomly selected and surveyed a subsample of households and the findings of the study are based on the responses of 2,957 households with 4,002 children who were within the mentioned age range.
Findings and Recommendations
The study found that enrollment is higher for younger children and girls while non-enrolment is associated with children’s reduced learning motivation and the economic capacity of households. Additionally, it was found that enrollment is correlated with better foundational learning outcomes. A larger share of children has better foundational Bangla literacy skills compared to English and Mathematics. Girls and older children perform better. However, learning gaps persist in all three subjects as students progress to higher levels of education, particularly in English and Mathematics.
While It is not possible to determine whether COVID-19 has exacerbated the urban-rural divide in children’s learning outcomes, it was found that students from private schools demonstrate advanced foundational skills. Across all three subjects, rural children consistently underperform compared to their urban peers. The learning gap in English may persist into adulthood for rural children. The study also found that major catastrophes are associated with diminished educational outcomes and prospects for children. Children who have experienced shocks (natural disasters, poverty shocks, and the pandemic) have difficulty retaining their foundational skills as they progress through grade levels, especially in English and Mathematics.
The findings of the study emphasize the urgent need for comprehensive education quality reforms to accelerate children’s post-pandemic learning recovery. Further, the study highlights the need to build more resilient education systems that can withstand and recover from such shocks.
Context
Despite having high aspirations for their children’s education and viewing it as the key to better life prospects, interviews with the Ultra Poor Graduation Program (UPGP) beneficiaries in 2022 revealed that parents marry off their adolescent daughters in their early or late teenage years. The generation gap causing mistrust and conflict is particularly salient for parents and adolescent girls due to concerns about dishonorable behavior and the resulting reputation concerns. Further, religious, gender, and societal norms prevent girls from discussing their growth and agency with their parents, deepening the parent-adolescent gap. Girls’ voices are often excluded from marriage decisions, and asserting agency threatens patriarchal authority, leading to stigma. Hence, despite awareness of the dangers of child marriage and high education goals, parents consent to early marriage for adolescent girls out of fear of the child’s misconduct and consequential social ramifications.
Objectives
The main research objective is to investigate whether bridging the parent-adolescent gap through the provision of counseling can reduce their conflict and lead to a healthier dynamic, thus resulting in a decrease in the rate of early marriage and an increase in the rate of enrollment for girls in secondary and higher secondary levels. In addition, other outcomes that will be measured include perceived parental support and parent-adolescent conflict, conflict styles, and the proportion of female students receiving private tutoring
Methodology
The counseling sessions will be adapted from the CONNECT parenting support program, involving both parents and adolescents. UPG groups will be randomly assigned to the intervention. The study sample will consist of 200 groups with 100 groups randomly assigned into the treatment group, and the rest constituting the control group. The treatment group will receive this additional PFA intervention, while the control group will receive the standard UPG intervention.