This paper investigates expenditures of food away from home (FAFH) by husband-and-wife households with children, determines the differentiated impacts of economic and demographic variables on FAFH by type of meal and across household types. Using the 2008 and 2009 Consumer Expenditure Survey, the system of expenditures on breakfast, lunch, and dinner is estimated with a multivariate sample selection estimator and results are compared with the Tobit system estimates. Statistical significance of error correlations among equations justifies estimation of the sample selection systems. Differentiated effects of variables on probabilities and expenditure levels highlight advantages of the sample selection system over the Tobit estimator. The effects of demographic and socioeconomic factors are found to vary by type of meal. Income, work hours, race, education, geographical region and household composition are important determinants of FAFH.