This paper estimates production functions of child cognitive and social development using a panel data of nine-year old children each with over two hundred home and school inputs as well as family background variables. A tree regression method is used to conduct estimation under various speci.cations. A small subset of inputs is found consistently important in explaining variances of child development results, including the number of books a child has at various ages and how often a mother reads to child by age .ve, while the eects of race and maternal employment are negligible when detailed inputs are controlled.
What Matter for Child Development?
Paper No. 24-2006