Berk Özler| World Bank |
Today’s subject matter is depressing. But, it is also worth continuing to read because, for many, the numbers will be eye-opening, and because with an improved focus on the types of programs targeted to communities suffering from these problems, we may be able to substantially improve the welfare of children and women.
In a study just published in Social Science and Medicine – Population Health, we evaluated the impact of Girl Empower – an intervention that aimed to equip adolescent girls with the skills to make healthy, strategic life choices and to stay safe from sexual abuse using a cluster-randomized controlled trial with three arms in Nimba County, Liberia. Among the target population of 13-14-year-old females, one in five had had their sexual debut at baseline and approximately 30% of the first sexual encounters among this sexually active group had been nonconsensual. Over one-third (37.3%) reported ever experiencing sexual violence: 7.8% had been physically forced to have sex, 8.4% had been non-physically pressured (coerced/persuaded) to have sex, 24.7% had had someone unsuccessfully attempt to have sex with them, and 28.9% had been touched in a sexual way.
There are a number of recent studies that examine the effectiveness of interventions to reduce sexual violence against women and children – gender-based and intimate partner. Such interventions include safety net programs, cash or in-kind transfers combined with behavior change communication (BCC) interventions, and girls’ clubs. See Buller et al. (2016), Peterman et al. (2017), Yount et al. (2017) for reviews and systematic review of reviews; see Bandiera et al. (2020) on girls’ clubs; Buchmann et al. (2018) for empowerment programs and incentives to delay marriage; and Roy et al. (2019) for BCC (for child nutrition) and cash/food transfers. [Also see these helpful slides by Hidrobo on cash transfers and IPV. You can find links to all the studies cited here in our paper that is linked in the previous paragraph.]
Our study differed slightly from some of these studies above. First, the target population was females in early adolescence (13-14 years old with mean age 13.7 at baseline). The mentoring program, designed and run by the International Rescue Committee (IRC), consisted of (i) Girl Empower life skills curriculum, facilitated by local female mentors weekly for approximately 40 weeks; (ii) Caregiver discussion groups, facilitated by IRC staff; (iii) Individual savings start-up for the girls; and (iv) Capacity building for local health and psychosocial service providers. Second, we wanted to assess the marginal gains from making participation in Girl Empower conditional on a small transfer to the caregiver of the girls. At $1.25 per session attended (or a maximum of $40 for perfect attendance to 32 sessions), the transfer amounts were less than 10% of the cost of the program per person. Our hypothesis was that these transfers may enhance program effectiveness at a small cost, by boosting program participation among early adolescents. We did not have a 2×2 factorial design: instead we had a control group, GE, and GE + cash – with 28 communities randomly allocated to each arm.
[My co-author Marie-France Guimond (IRC) wrote a very useful piece that summarizes the program and its impacts, accompanied with some stories of the participants, photos, and a video. Here, I talk a little bit more about some details and mechanisms.]
The first interesting finding was that attendance in both GE and GE+ were quite high – around 28 out of 32 possible sessions. Based on the previous experience of our collaborators from the IRC and Population Council running similar interventions elsewhere (as well as the attendance rates in ELA clubs in Uganda reported by Bandiera et al. 2020), we had anticipated much lower attendance rates in the GE arm, leaving room for plenty of improvement in GE+. We do not know whether it was the diligent work of IRC Liberia in rolling out the intervention or high demand in Nimba County, Liberia (a post-conflict/post-Ebola setting) that was responsible for the 88% participation rate. Whatever it was, this was not differential between GE and GE+, meaning that any differences in impacts between these two variants of the mentoring program cannot be due to increased participation in the combined intervention.
Figure 2 in the paper summarizes the impacts on seven pre-specified domains at the 24-month follow-up, i.e. more than one year after the mentoring interventions and the small cash transfers ended (Tables 1-7 then unpack each of these domains, so that the readers can see the impacts on sub-components. All indices are constructed such that positive effects are desirable outcomes):