Tech-Driven Parenting Support in Fragile Contexts

After baseline, caregiver–adolescent pairs will be individually randomized into one of three experimental groups:

Control group: Caregivers receive only unrelated WhatsApp messages and /or videos (to be determined in pilot) that are neutral in content and not connected to parenting, adolescent well-being, or mental health.
Treatment 1: Parent-only treatment: Caregivers receive content-related WhatsApp messages and/or short AI-generated video clips that reinforce positive parenting practices, emotional communication, and socioemotional support strategies (to be determined in pilot). Within this group, assignment is further randomized into two sub-arms:
A. Needs-based messages (50\% of the sample): tailored to the specific parenting problems and skill gaps identified at baseline (e.g., managing conflict, fostering adolescent self-regulation, reducing reliance on physical punishment, improving communication).
B. All messages (50\% of the sample): full set of intervention messages, regardless of baseline needs.

Treatment 2: Parent + adolescent treatment: Both caregivers and adolescents receive content-related WhatsApp messages and/or short AI-generated video clips (to be determined in pilot) . Within this group, assignment is further randomized into two sub-arms:
A. Needs-based messages (50\% of the sample): tailored to the specific problems and skills identified at baseline, focusing on strengthening both caregiver and adolescent capacities.
B. All messages (50\% of the sample: complete set of intervention messages, irrespective of needs.

2026-03-15

2026-04-30

1-Caregiver Mental Health: Caregiver mental health will be assessed using a combination of self-reported surveys and voice biomarkers. Data will be collected at baseline (self-reported surveys only) and at endline (surveys and voice biomarkers).
2-Caregiver Self-Efficacy: Data will be collected at both baseline and endline.
3-Adolescent Conduct, Emotional Regulation, and Mental Health: Adolescent outcomes will focus on conduct, emotional regulation, and depressive symptoms. Data on adolescent outcomes will be collected only at endline.

1-Parenting Practices on Physical Punishment: Parenting practices will be assessed with a particular focus on the use of physical punishment. Caregivers will self-report the frequency and types of disciplinary strategies employed with adolescents, capturing both punitive and non-punitive approaches. In addition, adolescents will provide self-reports on their experiences of physical punishment, offering an independent perspective on the prevalence and severity of these practices. Data will be collected at baseline and endline.

2-Family Relationships: Family relationships will be assessed through adolescent and caregiver surveys capturing the quality of communication, emotional closeness, and conflict within the household. Data will be collected at both baseline and endline.

3-Community Relationships: Community relationships will be measured using self-reported surveys from both adolescents and caregivers. Indicators will include perceived trust, mutual support, and participation in community activities. Data will be collected at baseline and endline.


4-Adolescent Academic Achievement: Adolescent academic achievement will be measured through a combination of self-reports and caregiver reports, complemented where possible by administrative school records. Data will be collected at baseline and endline.

The study will recruit 6,000 caregiver–adolescent pairs (2,000 per stratum: Venezuelan migrants, internally displaced persons, and non-displaced Colombians) and randomize them across three main experimental arms. The study will be implemented in Medellín (Antioquia) and will target households with at least one adolescent aged 10–17 living with a primary caregiver. We focus on three population groups: (i) Venezuelan migrants, (ii) internally displaced persons (IDPs), and (iii) non-displaced Colombian residents.

Not available

Computer randomization

caregiver–adolescent pairs

No

n/a

6000

2000 per arm

The study will recruit 6,000 caregiver–adolescent pairs (2,000 per stratum: Venezuelan migrants, internally displaced persons, and non-displaced Colombians) and randomize them across three main experimental arms. With a two-tailed test at $\alpha$ = 0.05, this sample provides high statistical power to detect modest but policy-relevant changes in caregiver and adolescent outcomes. Under standard assumptions of individual randomization with balanced allocation, no clustering, and covariate adjustment, the minimum detectable effect (MDE) for pooled treatment versus control contrasts is approximately 0.09 standard deviations with 80\% power. For pairwise comparisons between treatment arms (e.g., caregiver-only vs. caregiver+adolescent, needs-based vs. full content), the MDE increases slightly to 0.11–0.12 SD, reflecting reduced effective sample size. Subgroup analyses stratified by displacement status (2,000 pairs per stratum) yield MDEs of approximately 0.16–0.18 SD, which remain within the range of meaningful effects documented in the parenting and mental health literature. These calculations assume 10\% attrition between baseline and endline, consistent with prior survey experience in displaced and low-income populations in Colombia. If attrition is lower, detectable effects will be correspondingly smaller. The study also benefits from repeated measurement of several outcomes at baseline and endline, which further increases power by allowing for precision gains from covariate adjustment. Taken together, the sample size is sufficient to detect effect sizes comparable to those found in meta-analyses of parenting interventions in fragile and low-resource settings (0.2–0.5 SD), ensuring that even conservative estimates of program impacts will be detectable with high confidence.