Comparing the morbidity of breast-fed and bottle-fed infants is confounded by inherent differences in breast-feeding and bottle-feeding mothers and their infants. Self-selection introduces complex variables encompassing much more than milk source used for infant feeding. Reasons for selecting breast or bottle feeding relate to demographic, socioeconomic, educational, ethnic, cultural, and psychological factors, as well as maternal and infant physical and emotional health. Many of the differences in the maternal populations may affect infant care practices, access to medical care, and infant health status. Studies published to date have not quantified these confounding effects and other potential biases in comparing morbidity to breast- and bottle-fed infants and the relationship between milk source and incidence of infantile disease remains in question. There is need for more cautious use of the available data and investigators must seek ways to design future studies to take into account the differences between breast-feeding and bottle-feeding mothers that affect both reported and actual infant morbidity.