This study examines the existence and extent of gender bias favouring boys in the allocation of education expenditure in households using data from Andhra Pradesh in India collected by the Young Lives study in 2009/10. It attempts to identify the changing degree of gender bias with age. The estimation uses Engle curve demand analysis and a hurdle model, which separately considers gender bias in the decision of whether to incur education expenditure and the expenditure level conditional on investing in education. The study finds gender bias in school enrolment among children aged ten upward. In terms of expenditure, evidence of a pro-male bias is found in the decision of whether to invest in education and the level of expenditure on school fees and extra tuition. It is found that the degree of gender bias increases with age. The study also examines the existence of gender bias within households by incorporating household fixed effects and finds that there is significant gender bias in the intra-household allocation of education expenditure. Gender bias is also found to manifest itself via differential school choice with the probability of being enrolled in private schools being significantly higher for boys than for girls at all ages.