Back to History

Fields Changed

Registration

Field Before After
Last Published April 14, 2026 09:33 AM September 10, 2026 02:50 PM
Randomization Unit High school classes School-grade-shift
Planned Number of Clusters Approximately 400 clusters (high school classes) in total. Approximately 400 clusters per cohort, for a total of 1600 clusters across 4 cohorts.
Planned Number of Observations We expect 400 eligible applicants per cohort, for a total sample of 1,600 across four cohorts. We expect 650 eligible applicants per cohort, for a total sample of 2,600 across four cohorts.
Sample size (or number of clusters) by treatment arms We expect 400 eligible applicants per cohort. Across four cohorts, this implies a total of 1,600 eligible applicants. Assuming equal assignment across the three study groups, the full sample will include approximately 800 students in the control group and 800 students in the pooled mentorship treatment group. For comparisons involving each specific treatment arm, we expect approximately 400 students in the mentorship-only arm and 400 students in the mentorship-plus-financial-coaching arm, with the control group comprising 800 students. We expect 650 eligible applicants per cohort. Across four cohorts, this implies a total of 2,600 eligible applicants. Assuming equal assignment across the three study groups, the full sample will include approximately 1,300 students in the control group and 1,300 students in the pooled mentorship treatment group. For comparisons involving each specific treatment arm, we expect approximately 650 students in the mentorship-only arm and 650 students in the mentorship-plus-financial-coaching arm, with the control group comprising 1,300 students.
Power calculation: Minimum Detectable Effect Size for Main Outcomes We estimate Minimum Detectable Effects (MDEs) for selected final outcomes for which comparable baseline data are available and for which prior mentorship studies provide relevant benchmarks. MDEs are reported for both the pooled mentorship treatment and the combined mentorship and financial coaching treatment. The calculations account for the clustered design of the study, with randomization at the level of high school classes. We assume an average of 4 students per cluster and an intra-cluster correlation coefficient (ICC) of 0.12. We further assume a take-up rate of 90% and zero attrition, given that the main outcomes are measured using administrative data. For higher education enrollment, the minimum detectable effect is 4.6 percentage points for the pooled treatment and 5.2 percentage points for the combined treatment. For formal wages, the minimum detectable effect is 0.18 standard deviations for the pooled treatment and 0.21 standard deviations for the combined treatment.
Intervention (Hidden) We evaluate a free, online mentorship program for public high school students in São Paulo, Brazil. The intervention has two treatment arms and a control group, randomized at the classroom level. Mentorship arm: Students receive four one-on-one sessions with a volunteer mentor, a college student or young professional matched on shared characteristics and career interests. The structured sessions cover: self-awareness and identification of talents and interests; information on occupations, required qualifications, and mean wages by university-major combination; mapping of courses, internships, and educational pathways; and creation of a career plan and vision board. Combined arm: Students receive the four mentorship sessions plus an additional one-on-one financial coaching session. This session focuses on budgeting, savings and investment strategies aligned with students' career goals — a particularly relevant topic, as many participants are receiving substantial cash transfers through Brazil's Pé de Meia program for the first time. We evaluate a free, online mentorship program for public high school students in Brazil. The intervention has two treatment arms and a control group, randomized at the classroom level. Mentorship arm: Students receive four one-on-one sessions with a volunteer mentor, a college student or young professional matched on shared characteristics and career interests. The structured sessions cover: self-awareness and identification of talents and interests; information on occupations, required qualifications, and mean wages by university-major combination; mapping of courses, internships, and educational pathways; and creation of a career plan and vision board. Combined arm: Students receive the four mentorship sessions plus an additional one-on-one financial coaching session. This session focuses on budgeting, savings and investment strategies aligned with students' career goals — a particularly relevant topic, as many participants are receiving substantial cash transfers through Brazil's Pé de Meia program for the first time.
Back to top