Teaching Mental Health: Evaluating India’s Saharsh “Happiness Curriculum”

The Happiness Curriculum consists of a daily 30 minute happiness class, delivered by class teachers at the start of each school day, made room for by shortening other classes. A typical class begins with a guided meditation, leading into a group activity. This daily activity rotates between categories including storytelling, reflective conversations, and a mindfulness-only day each week. This is followed by a class discussion, and finally, a mindfulness check-out. Teachers are trained by master trainers employed by the Labhya Foundation, and administer these sessions following handbooks. These were designed following the Clover Model of Youth Development created by Dr. Gil Noam at Harvard Medical School, and are adapted to the local context and the students’ specific grade. For example, grades 1-2 focus on the self including senses and emotions while grades 5-6 examine friendship, teamwork, trust, and communication.

2023-09-01

2025-04-01

Our survey’s primary outcomes of interest span mental health, wellness, academic performance, and social networks.

Mental health outcomes include standard measures of anxiety, depression, and loneliness, as well as beneficial traits such as self-control (which is linked to mindfulness), and grit. Specifically, we plan to incorporate the GAD-7 and MFQ to measure anxiety and depression respectively, as well as the UCLA Loneliness Scale, Grit Scale, and the Domain-Specific Impulsivity Scale for Children (DSIS-C). These are all widely used questionnaires for measuring mental health and have been tested extensively among adolescents and in development contexts. In addition, we include questions intended to measure subjective well being, empathy, motivation, classroom behavior, students’ relationships with their peers, parents, and teachers, and their social networks. Our surveys will also include brief measures of academic abilities, implementing leading Indian education NGO Pratham’s ASER evaluation of math and language, as well as Raven’s matrices to test cognitive skills. Survey questions were refined following piloting in April 2023. We will complement our surveys with administrative data on student academic outcomes. We will look at test scores and grades in other subjects to measure the aggregate impact on learning, as well as teacher reports of attendance.

We will conduct 800 teacher surveys to measure secondary outcomes. This will include teacher attitudes toward the curriculum, comprehension, and mental health.

We will measure how treatment effects on students vary with teacher buy-in, comprehension, and teacher mental health, and any effects of the program on teachers themselves. We seek to assess the role of skepticism in the outcome of teacher-led interventions in a highly-stigmatized area, as well as how variance in teachers’ ability to understand and deliver a socioemotional learning intervention impacts student outcomes. We further wish to document how teachers’ own mental health impacts classroom learning, which we believe to be an understudied area.

The experiment will take place with students in grades 1-8 in 100 randomly-selected primary schools in the West Tripura school district. Randomization will occur within schools at the grade level, where grades will be either assigned to: (1) a control, (2) the full Happiness Curriculum or (3) a mindfulness-only condition.

(1) Control group. The control group will receive no form of the Happiness Curriculum. Instead, the teacher will provide an additional 30 minutes of teaching on standard subjects such as math, science, and language. Because the Happiness Curriculum displaces 30 minutes of class time, this is the relevant comparison to understand the curriculum’s effects. On the one hand, the curriculum could come at a learning cost, as it crowds out teaching time spent on traditional subjects. On the other hand, it could increase focus throughout the day, and improve mental health, creating potential learning gains. This design allows us to evaluate these hypotheses.

(2) The full Happiness Curriculum. The Happiness Curriculum is a socioemotional learning intervention developed by psychologists and education experts. It has two main components: guided mindfulness meditation sessions, and group activities such as storytelling and reflective conversations that vary daily. All grades participate in daily guided meditation to start their day.

(3) Mindfulness Only. To separate the effects of the two components of the curriculum, a third group of students will complete only the mindfulness component. They will start each day with a guided meditation, and will complete the once-a-week mindfulness-only sessions of the full curriculum, but will not take part in the other activities. Instead, they will devote the remaining time to teaching as usual, as in the control group. This allows us to measure whether the activities have additional value beyond a mindfulness-only option that is both less intrusive and has documented benefits for anxiety, depression, and self-control (Davis and Hayes 2012).

Not available

Randomization will be done in office by a computer.

School-grade

Yes

4,000 students, with 5 students per grade per school (the cluster), for a total of 800 clusters.

4,000 students

400 control grades (2,000 students)
200 full Happiness Curriculum grades (1,000 students)
200 mindfulness only grades (1,000 students)

Grade-level randomization across grades 1-8 in 100 schools (800 clusters), with five students surveyed per grade, yields a minimum detectable effect of 0.1 standard deviations for survey outcomes. This is based on simulated power calculations, with stratification by school and standard errors clustered at the school-grade level of randomization. School-level randomization is infeasible in this setting, as the government requires that the curriculum be present in every school. Randomizing at the grade level, rather than the classroom or individual level, minimizes the likelihood of spillovers to control students, and maximizes the number of randomization clusters subject to the government’s constraint. Additionally, randomizing within a school allows us to control for school-level fixed effects, further increasing statistical power.