The height-for-age of Indian children.

The height for age of children is used as an indicator of the prevalence of undernutrition, i.e., 'stunting'. This study uses unit-record data on over 20,000 rural children, from 16 states of India, to explain variations in their height-for-age. Previous studies of children's heights have focused exclusively on the mean of the distribution of heights-for-age using ordinary least squares (OLS) regression. Instead, this paper uses quantile regression--which analyses different parts of the height for age distribution--and permits a richer set of conclusions to be drawn. The analysis shows the importance of locating malnourished children in terms of their place in the distribution of nutritional outcomes and, then, studying the differential impact of the determining variables on outcomes for children in different locations of the distribution.