This paper investigates the effect of two pupil-to-teacher ratio policies on test scores for children with different achievement levels. Using data from a large randomized experiment in early childhood, we estimate unconditional quantile treatment effects of small class and teacher aide, as compared to regular classes. For the small class intervention, results show that pupils in the middle of the achievement distribution profit the most from being assigned to a smaller class, whereas pupils at the bottom or at the top of the achievement distribution experience almost no gain in test scores. For the teacher aide intervention, the analysis reveals positive and significant effects for students at the bottom of the achievement distribution, an effect stronger for boys and disadvantaged pupils. The findings suggest that the average effects reported in traditional empirical studies on pupil-to-teacher ratio interventions provide an incomplete characterization of the impact on the achievement distribution, thus constituting a weak guide for policymakers.