The Economics of Domestic Violence: Evidence From Bangladesh

Standardized attitudinal questions from the Demographic and Health Survey (DHS), which capture whether violence is considered justified in specific situations to measure changes in husbands' personal attitudes towards IPV. Measured by the average score on a justification index across treatments. This index is a normalized index of how justified men think it is for a husband to use IPV.

Hypothetical vignettes, where respondents evaluate the likely and socially acceptable behavior of a hypothetical husband whose wife transgressed a gender norm. These vignettes mitigate social desirability bias and improve the accuracy of responses on sensitive topics to measure changes in husbands' personal attitudes towards IPV. Measured by the average level of violence respondents across treatments believe hypothetical husband will and should assign if the hypothetical wife transgresses. The scale will be between 0 and 5.

Husbands' perceptions about others' attitudes towards IPV were measured using the hypothetical vignettes. Respondents estimate the attitudes of the other surveyed men in their community using visual aids and stickers, each representing a different respondent.

Hypothetical vignettes, where respondents evaluate the likely and socially acceptable behavior of a hypothetical husband whose wife transgressed a gender norm. These vignettes mitigate social desirability bias and improve the accuracy of responses on sensitive topics to measure changes in husbands' personal attitudes towards IPV. This will be measured by the average level of violence respondents across treatments believe other men in their community will assign if the hypothetical wife transgresses. The scale will be between 0 and 5.

Men's coping and emotional self-regulation strategies are measured using an index constructed from responses provided by their wives. This index includes selected questions on men's approaches to conflict management from the Revised Conflict Tactics Scale, adapted items from the Bengali version of the Brief Resilient Coping Scale that capture spouses' perceptions of men's behavior, and measures of substance use. IPV-related questions are intentionally excluded, as these behaviors represent a distinct conceptual category measured separately. To simplify interpretation, the index is standardized relative to the control group by subtracting the control group's mean and dividing by its standard deviation. Consequently, the resulting index has a mean of zero and a standard deviation of one within the control group. Higher values of the index indicate more positive coping styles and better emotional regulation.

Husbands' self-reported symptoms of depression measured using the PHQ-9 scale. PHQ-9 is a 9-item validated questionnaire used to screen for depression with a range of scores from 0-45. A cumulative score of ≥10 is considered positive with lower scores indicating no or mild anxiety.

Husbands' open-ended responses on daily stressors and positive experiences to measure husbands' resilience and emotional well-being. A summary index with two subcomponents is developed. First, a sub-index of daily events using an adapted version of the Daily Hassles and Uplifts Scale that includes items relevant to our target population. Second, a sub-index that aggregates men's appraisal of stressors and is based on the 10-item version of the Perceived Stress Scale validated in Bangladesh. Each subcomponent is standardized to have a mean of zero and a standard deviation of one for the control group. The summary index is defined as the equally weighted average of these standardized sub-indexes, with signs oriented so that higher values consistently indicate more favorable outcomes (fewer stressors and/or lower perceived stress).

Biometric stress indicators, captured by Empatica E4 wristbands during sensitive survey modules, measuring physiological markers of emotional regulation and stress to measure husbands' resilience and emotional well-being. Specifically, electrodermal activity (EDA) data, expressed in microsiemens (μS), is used to analyze Skin Conductance Responses (SCRs). SCRs are quantified using a trough-to-peak (TTP) analysis, where the amplitude of each response is measured as the difference between the peak and the preceding trough within a defined time window. To identify event-related SCRs, a response window of 1 to 5 seconds following a stimulus is applied, with a minimum amplitude threshold of 0.01 μS to detect significant responses. The primary outcome variable is the sum of SCR amplitudes for significant responses. The average sum of SCR amplitudes across treatment groups is compared to assess differences in physiological stress responses.

An adapted version of the DHS, covering physical, psychological, financial, and sexual violence. To enhance privacy and reduce misreporting, women use visual response tools to answer these questions privately to measure women's experiences of IPV. Measured by the average of a standardized IPV index. This is a normalized index which captures the frequency (never, very few times, sometimes, or very often) of women's reported experiences of IPV in the past six months. Higher values indicate higher frequency of IPV.

Enumerators' reports of visible IPV-related injuries, such as bruises, documented using a standardized reporting protocol. Enumerators receive specialized training to identify and record signs of physical violence, with a focus on facial bruises and scratches, which remain visible even for veiled women to measure women's experiences of IPV.

Health assessments by community health workers, who check for IPV-related injuries to measure women's experiences of IPV. Measured by the percentage of female respondents across treatments on whom community health workers identify IPV-related injuries.

Women's self-reports of IPV through rolling phone surveys six to twelve months post-intervention.

Measured by the average of a standardized IPV index. This is a normalized index which captures the frequency (never, very few times, sometimes, or very often) of women's reported experiences of IPV in the past six months.

Husbands' reports of the likelihood that they are a "respectable man". Investigators constructed the respectable man scale using a combination of qualitative and quantitative methods. Husbands answer ten embedded questions throughout the survey that contribute to their respectable man score. At the end of the interview, investigators inform men that the top five scorers in the community are considered "respectable men," but we do not reveal their actual scores to them. Husbands indicate how confident they are that they rank among the top five in their community. Measured by the mean likelihood that men across treatments think they rank among the top five respectable men in their community. The scale will be between 0 and 100.

Husbands' perceived probability that two randomly selected community members would consider them respectable, elicited using an identical ten-bean method. To reduce misreporting, husbands receive financial incentives based on the accuracy of their estimates. Measured by the mean likelihood that men across treatments think others rank them among the top five respectable men in their community. The scale will be between 0 and 100.

Reduction in the number of women exposed to IPV or the intensity of IPV in the last six months per USD 1,000.

Multiply the observed reduction in IPV (either in absolute prevalence or an intensity index) caused by each edutainment campaign by the total number of men assigned to that campaign. Investigators will calculate costs by summing both start-up and implementation costs. Start-up costs include intervention development (content production and pilot testing), adaptation (translation and cultural adjustments), and set-up (community engagement and facilitator training). Implementation costs include broadcasting and screening expenses, facilitator compensation, and routine monitoring and evaluation. Research-related expenditures will be excluded and only include costs related to activities expected at scale.

Benefit-to-cost ratio calculated by dividing the estimated benefits from reduced IPV by the total costs of intervention delivery. To estimate the benefits of IPV reduction, investigators will use the treatment effects on income and a web-based cost benefit analysis tool developed by one of the Principal Investigators. This tool calculates the lifetime economic returns to increased income.

Net present value calculated as the difference in discounted benefits and costs. To estimate the benefits of IPV reduction, investigators will use the treatment effects on income and a web-based cost-benefit analysis tool developed by one of the Principal Investigators. This tool calculates the lifetime economic returns to increased income. The unit of measure for NPV is 1000 USD.