This paper presents an economic model of the relationship between time allocated to learning activities at home and achievement in school. The model is contrasted to four alternative models of the home-school relationship. Data from a 1975 survey of the parents of 887 elementary school children from one New York State school district are used to examine the relationship between 36 measures of family behavior and characteristics to student performance on standardized math and reading tests. Independent variables are organized into six groups representing socioeconomic status, family structure, maternal availability and education, parental expectations, after-school time use and parent-child interactions. Controlling for parental expectations, home time use and parent-child interactions, differences in family socioeconomic status, family structure (e.g., birth order and siblings) and maternal employment add little or nothing to predictions of school achievement. The analysis strongly suggests that it is differences in what families do rather than differences in their characteristics or incomes that affects school achievement. (Author)