Media, institutional reports, and research reviews from all around the world are giving voice to what has been called the silent or shadow pandemic: the increase in violence against women during the COVID-19 pandemic. It has been attributed to, among other things, the intersection of confinement during lockdowns, increased stress at home, and a decline in access to services. Here we look at new evidence on the impact of a program in Bolivia, which was not designed to directly address gender-based violence. Rather, it was intended to improve income opportunities for young persons by developing their soft and hard skills.
The Youth Empowerment Program was implemented by Save the Children with funding from Bulgari. The program combined training in soft skills and technical skills with sex education, mentoring, and job-finding assistance. The program started in September 2019, and ran for about four months in the form of 4-hour weekly meetings. Most activities were completed by March 2020 when a strict lockdown was implemented in these cities. Diego (our guest blogger), Selim Gulesci from Trinity College Dublin, and Manuela Puentes Beccar from Bocconi University partnered with Save the Children to evaluate the program. The evaluation is a randomized control trial with 600 vulnerable girls and boys aged 15-18 who applied to the program in four cities in Bolivia, a random half of them assigned into the program.
The baseline survey consisted of face-to-face in-depth interviews conducted in 2019 before the program began. Violence in the survey can be classified into 3 categories; physical, psychological, and sexual. The baseline data document a high level of violence experienced by both boys and girls. Over half (55%) of these youth reported having ever experienced some type of violence and there are no statistical differences in violence between treatment and control groups. Such high levels of violence are consistent with other data from Bolivia. The survey does not deconstruct the source of violence, specifically intrahousehold versus, for example, violence at schools (including bullying). Boys are more likely to report being the victims of violence than girls (57% and 53% respectively), with especially higher levels of psychological violence, but with lower levels of sexual violence than girls.
The follow-up survey took place in September 2020, seven months after the end of the program and six months into the lockdown. The survey was conducted by phone. Enumerators were carefully trained following stringent ethical guidelines with regards to questions on violence by phone, aimed at guaranteeing respondents’ safety and offering support services in case of need. Due to limitations associated with a phone survey, the data on violence are not comparable to the data from the baseline survey. Thus, the follow-up can’t tell us if overall levels of violence have gone up to down. The analysis of the follow-up data and a comparison of youth in the program to the control sample can inform whether the program served to mitigate or protect participants from violence.
In the follow-up survey, unlike at baseline, girls report higher levels of violence than boys (here is the paper). The largest difference is in the rate of psychological violence, which “includes cases when someone insults, threatens, verbally abuses, ridicules or makes fun of you”. In terms of the program effects offsetting or mitigating the experience of violence, the figure below shows that the program significantly reduced violence experienced by girls. The prevalence of any violence in the last three months reported by the girls who participated in the program six months earlier was 10 percentage points lower (from a mean of 21 percent) than girls not in the program. Lower levels were found for all three types -- physical, psychological, and sexual violence -- although only the last two differences are statistically significant. On the other hand, there was no mitigation effect for boys; there is no significant difference in rates of violence experienced by boys in the treatment from the 7 percent prevalence reported in the control group.
Of course, there are concerns about self-reporting bias affecting these results. The survey also elicited the prevalence of violence using item list experiments, a strategy adopted in the literature to elicit sensitive topics that relies on indirect elicitation techniques. The list experiments confirmed the main findings of a significant reduction in violence experienced by girls who were participants in the youth empowerment program in comparison to that experienced by girls who were not, and of no effects of the program on violence experienced by boys.
What explains the program’s effect at lowering violence?
The program was designed to strengthen youth income-generating capacity by developing their soft and hard skills, and by offering job-finding assistance. It could have resulted in less violence either through a direct effect of the program on soft skills such as expressiveness and self-confidence or through an increase in earnings and bargaining power within the household. With the limited scope of the phone survey, it is difficult to unpack these alternative pathways.
The program did increase girls’ earnings but had no effect on boys’ earnings. The increase in earnings was not generated through jobs obtained with the help of the program; instead the program was associated with (imprecisely measured) increases in self-employment and transfers. The increase in girls’ earnings may have contributed to reducing the prevalence of violence against them through two mechanisms. First, a change in girls’ access to economic opportunities may have improved their outside options and enabled them to leave abusive relationships. Second, it may have mitigated the higher stress levels linked to economic insecurity observed during the COVID-19 pandemic. But without more detailed data than that collected by the phone survey, it is hard to distinguish between the two.
There are no significant effects of the program on girls’ soft skills, such as self-confidence and expressiveness. However, this null effect could be due to the fact that soft skills are, in general, difficult to measure, and collecting the data by phone makes this harder.
While the results here are promising and expand the scope of how we think about the value of youth empowerment programs, more work needs to be done to understand what particular mechanisms explain the decrease in violence, and whether they can be replicated in other contexts. The results also point to how a quick pivot in the data collection approach of an RCT, in light of the new realities of COVID-19, can bring additional insights to an evaluation.
Join the Conversation