Abstract
Previous research on the effect of infant feeding methods in the Third World has utilized univariate statistical methods. This research has ignored or simplified the many factors inherent in infant feeding and focused on a comparison of breast‐ and bottle‐feeding. As the use of an inappropriate statistic may yield spurious correlations or inflated results, this article reexamines the effect of infant feeding methods on infant weight gain using a multivariate model. Data from 1,610 Fiji‐Indian infants (at birth and at six weeks of age) and 860 mothers showed that there was no difference in infant weight gain between the breast‐ and bottle‐fed infants. Assuming that infant weight gain is an index of health status in the neonate, this research implies that the breast/bottle feeding controversy is of no clinical significance in this setting. It is recommended that future research utilize multivariate statistical methods to test the effect of infant feeding methods on infant health.

This publication has 3 references indexed in Scilit: