Can at-scale parenting interventions effectively reduce violence towards children?


Published 12.02.24

A free intervention in El Salvador aimed at addressing stress-management and parenting for caregivers, and delivered through SMS, lowered women’s use of violence towards children but worsened male caregiver’s mental health and lowered their interactions with children

Good parenting is at the forefront of early-child development and a healthy family environment. Despite the importance of healthy parenting practices, a recent report reveals that nearly three out of every four children experience regular physical or psychological abuse at the hands of their caregivers (UNICEF 2017). Exposure to violence at a young age can lead to the development of risky behaviours and attitudes that are correlated with participating in criminal activities (Sviatschi 2022). Parental mental health and good parenting practices play a key role in children’s development of socioemotional skills, thus supporting their future socio-economic success. Despite its importance and the fact that poor practices are widespread, there is limited understanding of the effects of highly scalable policies to address the problem in low-income settings (Justino et al. 2023).  

In our recent research (Amaral et al. 2023) we test the effect of a stress management and positive parenting intervention delivered to caregivers in El Salvador. Our intervention used information technologies and digital tools to reach caregivers during the onset of stay-at-home orders related to the COVID-19 pandemic. This was a time of high stress and economic volatility, when populations were hard to reach. As a result, in lieu of in-person policies, the promise of digital and remote interventions relies on its accessibility, affordability, and scalability. However, it is unclear whether such remotely delivered curriculum are as effective as when it’s delivered in-person.

Intervention and experimental design

Our study tests the effect of a free, remotely delivered parenting intervention on caregivers’ mental well-being and the quality of their interactions with their children. To test the effect of the intervention we collected information from a pool of 3,103 caregivers of children aged 0 to 8 years and randomly assigned half of the sample to a treatment condition. We compare the effect of the intervention of these treated caregivers with those that were randomly assigned to a control condition, and therefore did not have access to the intervention. Our experimental design also aimed at testing for gender differences. We hypothesised that since men and women have extensively different roles in parenting and mental health status, the effects may be ambiguous depending on the gender of the caregiver targeted. To test for this gender difference, we ensured a 50% split in participants of both genders in each treatment and control group.

The curriculum we tested was created by Glasswing International – an NGO with extensive experience working with parents and at-risk children in this setting. The intervention was a digital format adaptation of an in-person stress management and parenting curriculum. The objective was to support caregivers’ during this period of stay-at-home orders, since stress, anxiety, and frustration can affect the quality of interactions between parents and children (Cluver et al. 2020). The intervention involved content to enable participants to improve their mental well-being by helping them identify daily life stressors and teaching them skills to manage these stressors effectively. The intervention also involved challenging the use of bad parental practices and offering positive parenting practice alternatives. The content was delivered via WhatsApp, and participants received three messages per week over the course of nine weeks. Participants watched videos, read simple and engaging brochures, comic strips, and had access to blogs.

Key findings

  1. We find that exposure to the intervention had detrimental effects for men, but not for women. For men, stress and anxiety increased by 0.108 SD and 0.095 SD relative to the males in the control group. Women’s mental health did not change. Exposure to the intervention led to 0.137 SD fewer interactions between male caregivers and children, while it had no significant impact for women.
  2. We find that the intervention reduced female caregivers’ use of physical violence by 0.098 SD and had no significant effect among men. This result aligns with the result in Diaz et al. (2023).
  3. Importantly, we show that participants did consume the information provided by this intervention - over 70% of participants opened the SMS/WhatsApp messages, and this rate remained stable throughout the intervention. We also detected improvements in the knowledge of parenting and stress management techniques that were only taught via text messages.
  4. The detrimental effects of the interventions on males stems from negative impacts among those who live in poorer households and those who cohabitate with a partner. These findings show that remote-delivery parenting interventions can have heterogenous impacts, particularly when it comes to gender and how gender interacts with family conditions, for example through income levels and cohabitation status.
  5. The negative response to the intervention is moderated by caregiver sex, family structure, and economic deprivation. Communication expectations between spouses, the quality of interactions and the psychological consequences of economic shocks may help explain our results. This discussion of the mechanisms driving the intervention’s effects is exploratory and serves as potential directions for future academic and policy-oriented research.

Policy takeaway

Our research evaluates the effects of a policy that is highly scalable and affordable - the appeal of such policies in low-income environments is enormous. With widespread mobile phone access, significant social disparities in healthcare access, and pronounced stigma associated with mental health issues, adapting in-person interventions holds a lot of promise.

However, we find that the impacts of a light-touch and low-cost digital health intervention depend critically on the participant’s characteristics and, in our context, on family structure. More evidence is needed to understand the heterogenous impacts of parenting interventions, so that targeting these policies may be more effective. Future applications of parenting interventions should understand in what ways remote vs. in-person delivery modes differ and how these interact with the process of knowledge assimilation and behavioral change.

From a policy perspective, while remote delivery programmes may be cheaper, lower accessibility barriers, and provide more privacy to participants, it is important to first assess the ways in which remote-delivery curricula may require additional tailoring and targeting so that the benefits of in-person modes of delivery remain intact.


Amaral, S, L Dinarte-Diaz, P Dominguez and S M Perez-Vincent (2024), "Helping families help themselves: The (Un) intended impacts of a digital parenting program." Journal of Development Economics, 166: 103181.

Cluver, L D, F Meinck, J I Steinert, Y Shenderovich, J Doubt, R H Romero, C J Lombard, A Redfern, C L Ward, S Tsoanyane and D Nzima (2018), “Parenting for Lifelong Health: a pragmatic cluster randomised controlled trial of a non-commercialised parenting programme for adolescents and their families in South Africa”, BMJ global health, 3(1): p.e000539.

Diaz, L I D, S Ravindran, M Shah, S M Powers and H Baker-Henningham (2023), “Violent Discipline and Parental Behavior: Short-and Medium-term Effects of Virtual Parenting Support to Caregivers” NBER Working Paper w31338.

Justino, P, M Leone, P Rolla, M Abimpaye, C Dusabe, D Uwamahoro and R Germond (2023), “Improving parenting practices for early child development: Experimental evidence from Rwanda”, Journal of the European Economic Association, 21(4): 1510-1550.

Sviatschi, M M (2022), “Making a Narco: Childhood exposure to illegal labor markets and criminal life paths.” Econometrica, 90(4): 1835-1878.