We evaluate the impact of educational reforms starting from the mid-1990s in India on the school attendance rate of low-income rural children aged 6-14 compared to ineligible rural children, employing NSSO data from 1983 to 2004/2005. We estimate a triple difference model allowing for differential (linear) trends and find a positive causal effect of school reforms on the school attendance rate of rural low-income children, although somewhat stronger for girls than boys. For both girls and boys in these groups, the increase in attendance rate is driven by the 6-11 age category and by children of scheduled tribe or scheduled caste background.