Prosocial behaviour (i.e., voluntary behaviour intended to benefit another) seems to be fully developed in children by the age of 6 years. However, questions about which factors modify behaviour at that age remain understudied. Here we used a resource allocation paradigm to test behaviour in 6-9-year-old school children. They could decide between a selfish (i.e., one sticker for themselves) and a prosocial option (i.e., one sticker for themselves and one for the receiver) and we tested whether friendship, social status and prenatal androgen exposure (approximated by the 2nd to 4th digit ratio; 2D:4D) influenced children's choices. We found that children behaved prosocially, and that their tendencies were negatively correlated with prenatal androgen exposure; i.e., children with high 2D:4D ratios (reflecting low prenatal androgen exposure) acted more prosocially than children with low 2D:4D ratios. Further, their social status in the classroom influenced their choices: children with fewer interaction partners chose the prosocial option more often than more 'popular' children. However, they did so irrespectively of whether they were paired with a recipient or not. Our results highlight the importance of considering social, as well as physiological factors when investigating behaviour in children.