Mente Amiga - School-based mental health (SBMH) support system for public schools in El Salvador

The study experimentally compares two delivery models for school-based mental health (SBMH) support. Both models target primary school students (grades 3–5, ages 9–11), run for two consecutive school years (six months per year), and follow a three-tiered approach consisting of universal social-emotional learning (SEL) sessions for all students (Tier 1) and targeted group or individual psychological support for students with greater needs (Tiers 2 and 3). The two treatment arms differ only in the provider of these services: in one arm, trained school psychologists deliver the intervention, while in the other, trained classroom teachers deliver the intervention. Control schools receive no intervention. Treatment is randomized at the school level across 300 randomly selected public schools, with 100 schools assigned to each study arm.

2027-04-01

2028-09-29

- Student behavior and mental health
- Student academic performance
- Student social dynamics
- Classroom climate
- Teacher mental health

- Student behavior and mental health: measured through the Strength and Difficulties Questionnaire (SDQ) (Goodman 1997, 1999). The primary outcomes are three scores computed according to the instrument’s scoring guidelines, including the overall SDQ score (sum of all items), an internalizing score (sum over all items capturing anxiety and depression symptoms), and an externalizing score (sum over all items capturing hyperactivity symptoms and conduct disorder).
- Student academic performance: assessed through a 40-minutes paper-and-pencil test in math an language. A weighted score and Item Response Theory (IRT) score will be computed.
- Student social dynamics: measured through friendship nominations. Outcomes will include network measures derived from these nominations, such as number of friends, homophily index, clustering, in-degree centrality, eigenvector centrality, betweenness centrality.
- Classroom climate: assessed through the World Bank's Classroom Snapshot Observation System. The primary outcome is the average number of students engaged in off-task behaviors (e.g., social interaction, inattentiveness, lack of participation).
- Teacher mental health: measured through a validated version of the Copenhagen Burnout Inventory adapted for teachers (Fiorilli et al., 2015). The primary outcome is the overall burnout score computed according to the instrument’s scoring guidelines.

Schools will be randomly assigned to one of three experimental conditions:
- T1: psychologists-delivered SBMH support (100 schools),
- T2: teachers-delivered SBMH support (100 schools),
- Control: a control group (no support, 100 schools).

Not available

Randomization will be stratified by school size (2 bins), rural/urban area (2 bins), and previous year test scores on the national standardized test (3 bins). The randomization itself will be conducted using the software RStudio.

School

Yes

300 schools

13,500 students, 900 teachers (900 classes).

100 schools per treatment arm

Based on our design with 100 clusters (schools) per treatment, 45 students per cluster, and assuming 80% power and a 5% significance level, we estimate a minimum detectable effect size of 0.10 standard deviations for student-level outcomes and 0.16 standard deviations for teacher- and classroom-level outcomes (formula based on Bloom, 2008).