This paper revisits the issue of the intra-household allocation of education expenditure with the recently available India Human Development Survey which refers to 2005 and covers both urban and rural areas. In addition to the traditional Engel method, the paper utilizes a Hurdle model to disentangle the decision to enroll (incur any educational expenditure) and the decision of how much to spend on education, conditional on enrolling. The paper finds that the traditional Engel method often fails to pick up gender bias where it exists not only because of the aggregation of data at the household-level but also because of aggregation of the two decisions in which gender can have opposite signs. There exists little evidence of gender bias in the primary school age group (age 5-9); however, pro-male bias in educational expenditure is observed in many Indian states in the age group 10-14. In contrast, gender discrimination in the secondary school age group 15-19 mainly takes place through the decision to enroll boys and not girls, and not through differential expenditure on girls and boys. Overall, the results reveal large regional disparities in the nature and existence of gender bias in education.