A comparison of two U.S. surveys of infant feeding.

This study compares the 1988 National Maternal and Infant Health Survey (NMIHS) and the 1989 Ross Laboratories Mothers Survey with respect to sample characteristics and proportions of women who breastfed. Weighted proportions of women included in the two surveys were compared according to various characteristics to see how well they represented U.S. childbearing women. A z-statistic was produced to test for significant differences in the proportions who breastfed. In 1988-89, 52.2% of American women breastfed according to the Ross survey and 53.4% according to the NMIHS. Despite differences in sample size, sampling technique and the application of sample weights, the difference between the surveys was only 1 percentage point, and most levels of most variables examined were comparable. The Ross survey is a valuable source of data because it is ongoing, and it is generally corroborated by the NMIHS. The NMIHS is probably more representative of the U.S. population of childbearing women, and is a reliable source of data for setting U.S. objectives concerning infant feeding, and for the study of other issues vis-à-vis breastfeeding.