Enhancing awareness and reporting of domestic violence : a Video Documentary RCT in Madagascar

The intervention is a 15-minute video documentary targeted at beneficiaries, both men and women, of a Malagasy multi-faceted micro-credit institution. The intervention aims to raise awareness and understanding of all forms of domestic violence (physical, psychological, economic, and sexual), and to disseminate information that can challenge prevailing biased beliefs. Drawing on real-life instances from the NGO's beneficiaries and under the guidance of a local clinical psychologist, the script is formulated to mirror the genuine experiences and dynamics within this demographic.
The video's script, developed in tandem with a domestic violence-specialized clinical psychologist, is grounded in real-world experiences of beneficiaries. We developed the video's script in collaboration with a clinical psychologist specializing in domestic violence, grounding it in the beneficiaries' real-world experiences. The goal is to provide a faithful representation of their encounters with domestic violence.

The initial animated segment of the documentary uses narratives from three distinct domestic violence cases, employing tactful animations to represent each scenario. The recurring slogan, "I report," highlights the critical role that reporting plays in challenging and altering the prevailing norms of domestic violence. In its second half, the documentary uses a live-action format to present victims' testimonials along with a detailed description of the NGO's intervention mechanisms. The narrative is further reinforced by contributions from key institutions—medical, law enforcement, and legal—who elaborate on the medical, legal, and societal repercussions of domestic violence. As the documentary concludes, it strategically adopts an encouraging tone, fostering resilience among victims and underscoring their potential to utilize available resources. Fundamentally, the documentary aims to inspire victims to break the silence surrounding their abuse, emphasizing that reporting domestic violence is a crucial part of the discourse on economic and social development.

2023-04-01

2023-09-30

(1) Knowledge and Awareness, (2) Reporting, (3) Attitudes , (4) Psychological and emotional well-being

(1) Knowledge and Awareness : improved knowledge and awareness of domestic violence, encompassing an understanding of the different forms of violence—physical, psychological, sexual, and economic—and the appropriate actions to take in instances of such violence.
(2) Reporting : dummy variables whether the individual personally reported a case of domestic violence, whether the social worker identified and confirmed a case of domestic violence and whether help was sought from social workers, attitudes and barriers to non-reporting through the use of scenario-based vignettes.
(3) Attitudes : perception of prevalence using vignettes that present hypothetical scenarios involving a female microentrepreneur; acceptability of domestic violence index, inspired by the Demographic and Health Surveys (DHS); legitimization of violence through vignettes and responses to an open-ended question on legitimization (recoded into categories for quantitative analysis). Similarly, we employ responses to the open-ended question on attitudes towards law and victim resilience.
(4) Psychological and emotional well-being : emotional well-being index (modified version of the Patient Health Questionnaire-9 PHQ-9), substance use (frequency of alcohol consumption over the past two weeks, adverse consequences, of alcohol use), anxiety (General Anxiety Disorder-7 GAD-7), prevalence of domestic violence experienced by women (modified version of WHO Violence Against Women instrument), life satisfaction, the ability to trust one's social environment, and trust in the institutions involved in addressing domestic violence issues, resilience index based on the capacity to bounce back after traumatic events.

Our experiment involves randomizing the broadcast of a domestic violence video among seven agencies of a microfinance NGO in Antsirabe, Madagascar, over a period of five months (May to September 2023). The video is incorporated into the activities conducted during the "loan allocation day," a three-hour group meeting held weekly for approved loan applicants to sign loan and health insurance contracts and interact with social workers. In the treatment group, the video is shown during this meeting, while in the control group, the activities remain the same except for the absence of the video. We take precautions to eliminate any priming effect from the video by ensuring that both the treatment and control groups have already watched a two-hour economic training video as part of the loan application process. This enables us to isolate the direct impact of the domestic violence awareness video.

To minimize the risk of contamination, we randomize the month in which the video is shown rather than randomizing at the individual level. Randomizing at the individual level is impractical due to the logistical challenge of distinguishing treated beneficiaries (who watch the video) from control beneficiaries within each weekly cohort, as they are called on the same day. Randomizing the month of video diffusion helps overcome this challenge. Additionally, it mitigates potential biases arising from unintentional exposure of controlled individuals to the video or treated individuals discussing the video with control individuals, as well as concerns about social stigma and non-compliance. It also prevents burdening social workers with the task of identifying randomly selected individuals for treatment, which could lead to misclassification. Both the treatment and control groups achieve complete compliance since the agencies are closed to other beneficiaries during loan allocation activities. Misclassification is unlikely as all agencies have received their diffusion schedules, and the research team reminds each agency at the beginning of the month whether or not to show the video. While there may be spillover effects if members of the treated group transmit messages from the video to future control group beneficiaries, these effects are worth studying as they provide insights into the content being transmitted.

computer

month per agency

No

no cluster

4000 individuals (evenly divided into a treated group of 2,000 and a control group of 2,000)

2000 individuals