Socioemotional Skills in Preschool. Evidence from Mexico

The intervention we study is Educating for Well-Being (EW), a program developed by a team of psychologists and neuroscientists from the Mexican NGO AtentaMente in close collaboration with a team of researchers from the Center for Healthy Minds at the University of Wisconsin-Madison. EW is an innovative and scalable SEL program targeted to preschool students aged 3 to 5. The program combines an age-specific mindfulness curriculum with other selected evidence-based interventions from social and clinical psychology. EW aims to develop specific socioemotional skills—such as emotional self-regulation, attention capacity, and sociability—that are considered critical for school readiness and academic success, and for socio-emotional well-being in general.

Specifically, EW comprises two main components. First, the SEL curriculum combines elements from other internationally recognized programs, such as the Kindness curriculum. The EW’s basic foundation is the mindfulness practice, aimed at cultivating attention and emotional regulation. It is combined with a set of complementary activities and exercises that are intended to strengthen different dimensions of “kindness” and sociability. These soft skills are relevant since they help young children to adapt more easily to primary school and to strengthen the learning process in general.

The program includes lesson plans and sample activities that are designed to be embedded into the regular day-to-day curriculum, and usually last between 5 and 20 minutes. Second, teachers receive a set of four relatively short face-to-face training—a 6-hour introductory workshop plus three 3-hour “booster” workshops— which is complemented with a more intensive self-paced online course (up to 72 hours distributed over 21 weeks). This process started in November 2019. The materials and activities include videos, guided practices, worksheets, and a mobile app to support participants’ daily practice, as well as sample lesson plans for participants to use when facilitating sessions with preschool students.

By developing these soft skills in teachers, the program is expected to reinforce its impact via different channels, including the better modeling behavior of teachers. Interacting with more self-regulated adults is expected to improve the self-regulation of children—and expected to result in higher quality student-teacher interactions due to the teachers’ improved well-being. EW is designed to be scalable, since it can be embedded into the day-to-day curriculum, relies on the existing pool of teachers, and minimizes the length of the face-to-face training.

2019-11-30

2020-05-30

Our ultimate goal is to estimate treatment effects on students. To this end, we intend to study the impact of the program on children's socioemotional wellbeing in terms of inhibitory control and socioemotional development.

To study some of the mechanisms behind any treatment effect we might find, we intend to look at some intermediate outcomes at the teacher level. Specifically, we intend to study teachers’ wellbeing using measures of emotional wellbeing, stress, anxiety, depression, burnout.

This project constitutes the first one of our research agenda on SEL during early childhood and their mid- and long-term effects on development outcomes. We plan to track these kids (using administrative data and their unique ID) to measure the impact of the program on first-grade enrollment and performance (using educational census), primary school test-scores (using the nation-wide standardized PLANEA test), wages (using the IMSS database), and whether they vote or commit a crime (using the INE database). These mid- and long-term effects will be part of upcoming projects.

The study universe includes all preschools in the municipality of Escobedo in Nuevo Leon. Out of the total of 111 schools, 50 percent of them were randomly assigned to treatment (55 preschools) and the rest to control (56 preschools) groups. The randomization process used the type of system (In Spanish Sistema, which indicates whether the school was originally a State-ran school or whether it was transferred from the Federal to the State Government) and shift (whether it is a morning or an evening shift) as stratification variables. In some of these municipalities, some preschool centers are located in the same facility but have different shifts. These are completely separate schools that do not share principals, students, or teachers. For this reason, some treatment and control preschools are located in the same facility.

The study universe includes all preschools in the municipality of Escobedo in Nuevo Leon. Out of the total of 111 schools, 50 percent of them were randomly assigned to treatment (55 preschools) and the rest to control (56 preschools) groups. The randomization process used the type of system (In Spanish Sistema, which indicates whether the school was originally a State-ran school or whether it was transferred from the Federal to the State Government) and shift (whether it is a morning or an evening shift) as stratification variables.

In this municipality, some preschool centers are located in the same facility but have different shifts. These are completely separate schools that do not share principals, students, or teachers. For this reason, some treatment and control preschools are located in the same facility. The implementation of this research project will include two phases. First, schools assigned to the treatment group will receive the full intervention, including components 1 and 2. After one and a half years, schools in the control group will participate in the intervention.

The randomization was done in-office by the researchers using STATA

The randomization unit was the preschool

Yes

111 schools

We expect to collect data from around a total of 500 teachers and 1,500 caregivers (15 caregivers per preschool).

55 preschools are treated and the 56 preschools were assigned to the control group.

For the statistical power and sample size estimations, we assume (conservatively) an intra-school correlation of 0.3, an R2 of 30%, a type I Error rate of 0.05, and statistical power of 0.8. Using these values, the minimum detectable effect is 0.29SD. However, if the intra-school correlation is lower, say 0.15, then the MDE is 0.21. An additional complication is that 3 treatment schools are non-compliant, as their principals decided not to participate in the program (all control schools are compliant). Thus, the MDE with non-partial compliance is 0.30 and 0.22, with an ICC of 0.3 and 0.15, respectively. See estimations in Figure 2. These estimated MDE are not different from some found in existing studies for similar interventions.