The pandemic of violence against women and girls is a longstanding human rights violation that stifles development. Growing evidence shows that the COVID-19 pandemic is associated with an increase in domestic violence. Additional stress due to health risks and economic uncertainty is likely to trigger conflict within the household; and survivors, trapped in the same space as perpetrators, face a greater risk of abuse.
As policy makers, NGOs and academics are working to understand the magnitude of the problem and to develop solutions to prevent it, the COVID-19 emergency presents additional challenges: how to collect the data without jeopardizing the safety of the respondents? It becomes even more challenging when data collection turns to a remote survey format, and both survivor and perpetrator may be together in the same space due to COVID-19 lockdowns.
Following the 16 Days of Activism Against Gender-Based Violence (GBV), we look at an innovation in rephrasing GBV questions to minimize risks in phone survey data collection, and what the data tells us about policies and programs that may mitigate GBV.
The questions do not directly ask about GBV experiences. Rather, they present a series of proxies and vignettes, which are jointly aimed to capture experiences of victimization. Proxy questions ask about experience of injury, feeling of safety, frequency of arguments. Vignettes describe hypothetical situations (e.g. husband lashing out on his wife due to financial stress) and ask the respondents if the situations are common in their community. The data collection team adhered to the WHO guidelines in training of the enumerators, and developed a protocol to ensure that the respondent was alone when starting the interview, with code phrases to skip questions or terminate the interview if the respondent was no longer alone during the interview.
The World Bank’s East Asia & Pacific Gender Innovation Lab team used these questions to collect data from 866 women in Indonesia in late August 2020. Respondents were part of an ongoing impact evaluation of the Indonesian government’s program for prospective migrants. Thus, the survey is not nationally representative. However, the rich data collected provides some valuable insights on which factors are associated with greater exposure to GBV, as well as perceived worsening of the situation due to the COVID-19 pandemic.
The phone survey also collected extensive information on employment, non-agricultural enterprises, remittances, food security, social assistance, knowledge of COVID-19, domestic work, and health symptoms. We were able to link this information to earlier data collected on the same households back in 2018. With this rich dataset, we employ machine learning to identify indicators that are important in predicting experience of violence in the last six months prior to the survey, as well as perceived worsening of the situation. The machine learning algorithm, however, does not inform whether an indicator is associated with higher or lower incidence of violence. For that, we apply a regression analysis to the strongest predictors of GBV identified through machine learning.
This analysis leads to two notable observations which have bearing on policy responses during the COVID-19 pandemic:
1. Women in households marked by economic distress, such as food insecurity, are at a higher risk of violence
This is in line with the evidence suggesting that stress from economic hardship increases the likelihood of domestic violence. On the other hand, food insecurity may also capture poverty and the gender norms that—in Indonesia—vary with household income. As the graph below shows, women from wealthier households in Indonesia are less likely to agree that that beating by a husband may be justified—for any reason.
2. Having a job protects women from increase in violence due to the COVID-19 pandemic
This, again, is in line with the theory that women’s economic empowerment decreases the risk of domestic violence: as a woman’s potential options outside of marriage improve, her situation within marriage is expected to get better too. More economically empowered women are not solely reliant on their husbands for income. They have an option to leave an abusive relationship, which increases their bargaining power within the relationship, thus, decreasing violence. In our data, having a job was the strongest protective factor from worsening of GBV due to COVID-19.
These two findings have direct relevance to policy responses to the COVID-19 pandemic. First, many governments boosted social assistance. In addition to other issues that this bold action addressed, it likely contributed to reducing the risk of GBV. Second, as women are shouldering a greater share of childcare responsibility during the pandemic, it is important to remember the multitude of benefits associated with women’s work. Mitigating risks of GBV is one of them. Our findings add to the arguments in favor of bold policy actions to ensure that the COVID-19 pandemic does not negatively impact women’s gains in the labor market.
Download and read:
- Brief: What Factors Exacerbate and Mitigate the Risk of Gender-Based Violence During COVID-19? : Insights From a Phone Survey in Indonesia
- Measurement note with questionnaires: Can we capture exposure to Gender-based Violence (GBV) through Phone Surveys during a Pandemic?
Proxy and vignette questions on gender-based violence were developed by the East Asia & Pacific Gender Innovation Lab, in consultation with GBV experts: Diana Arango and Amber Peterman. A revision has been made to title of this blog post.
Join the Conversation