Background: Fewer girls than boys are born in India. Various hypotheses have been proposed to explain this low sex ratio. Our aim was to ascertain the contribution of prenatal sex determination and selective abortion as measured by previous birth sex.
Methods: We analysed data obtained for the Special Fertility and Mortality Survey undertaken in 1998. Ever-married women living in 1.1 million households in 6671 nationally-representative units were asked questions about their fertility history and children born in 1997.
Findings: For the 133 738 births studied for 1997, the adjusted sex ratio for the second birth when the preceding child was a girl was 759 per 1000 males (99% CI 731-787). The adjusted sex ratio for the third child was 719 (675-762) if the previous two children were girls. By contrast, adjusted sex ratios for second or third births if the previous children were boys were about equal (1102 and 1176, respectively). Mothers with grade 10 or higher education had a significantly lower adjusted sex ratio (683, 610-756) than did illiterate mothers (869, 820-917). Stillbirths and neonatal deaths were more commonly male, and the numbers of stillbirths were fewer than the numbers of missing births, suggesting that female infanticide does not account for the difference.
Interpretation: Prenatal sex determination followed by selective abortion of female fetuses is the most plausible explanation for the low sex ratio at birth in India. Women most clearly at risk are those who already have one or two female children. Based on conservative assumptions, the practice accounts for about 0.5 million missing female births yearly, translating over the past 2 decades into the abortion of some 10 million female fetuses.