This thesis gathers research on three impact evaluations of interventions at the school and student level in Mexico. The first chapter evaluates the effects of a School Breakfast Program (SBP) on children's outcomes such as cognitive skills, illness, height and weight and grade repetition in the period 2002 to 2005. Quasi-experimental estimations provide evidence of positive effects on children's weight; however, such gains push children over their 'ideal' standardized average causing them overweight. This effect is significantly higher in the case of poorer children. The second chapter evaluates a Full-Time Primary Schools Program implemented in 2007, to work out if changing the time pupils spend at school can enhance skills in language and mathematics. Differences in Differences regressions point to a significant improvement of 0.11 standard deviations in mathematics and Spanish test scores after four years of treatment. These gains are three times higher in schools located in deprived areas and do not seem to be driven by students self-selection. The last chapter focuses on an exogenous policy change in Mexico which eliminates enforced grade repetition for all first to third grade students. This reform helped schools to reduce repetition rates from varying higher levels to almost zero in one academic year. Estimations coming from two-way fixed effects models using a panel of schools show an average reduction in dropout rates after reform implementation of 0.3% points along with no seeming effects on pupil's performance. General findings from the three chapters are of strong significance when placed into the broader debate about what works best in schools for improving children's academic performance and general education outcomes in Mexico.