This paper examines the effects of a massive salt iodization program on human capital formation of school-aged children in China. Exploiting province and time variation we find strong positive effects on cognition for girls and no effects for boys. For non-cognitive skills we find the opposite. We show in a simple model of parental investment that gender preferences can explain our findings. Analyses exploiting within province, village level variation in gender attitudes confirm the importance of parental gender preferences. Consequently, large scale programs can have positive (and possibly) unintended effects on gender equality in societies where boy preferences are important.