The interaction between family size and children quality has been a recurring topic in the economics of family. However, there is scarce evidence in Latin America, and the literature has not yet explored new mechanisms to explain either positive or null effects of an additional sibling found by different authors in the last ten years. This article addresses these two issues. On the one hand, I construct a simple theoretical model which rationalizes negative and positive effects of an additional sibling due to family interactions. On the other hand, I estimate the effect of family size in Colombia on school lag, school attendance, school dropout and child labor. I use data from the Demographic and Health survey and construct a set of instruments based on the report of the ideal number of children. The novelty of the instruments lies in that unlike most articles which can only estimate the effect from two siblings onwards, I can estimate the effect of a first sibling. I find that for first (second) born children a first (second) sibling generates null or positive effects on the four outcomes but there are negative effects from two (three) siblings onwards on the four outcomes.