Breastfeeding is promoted internationally as the preferred method of feeding for infants up to the age of four to six months. As a consequence there is much research interest in factors which may be predictors of breastfeeding. This paper presents a review of the recent literature identifying factors associated with the initiation and duration of breastfeeding among Western women. Attention was focused on studies which employed multivariate analysis of the data and breastfeeding initiation and duration were considered as two uniquely separate events. This review reveals that the associations between breastfeeding and many of the factors commonly investigated were not consistent. Many of the strong associations previously reported in univariate studies were not evident in more recent studies which employed multivariate analysis to control for covariates and potential confounders. While there was a strong and consistent association with demographic factors such as maternal age and level of education, there was a less consistent association with factors such as marital and socioeconomic status. Similarly, the previously reported univariate associations between breastfeeding and biomedical factors such as parity, method of delivery and infant health were less consistent in multivariate studies. However, a consistent negative association has been reported between maternal smoking habits and breastfeeding duration. More recent studies have investigated the association between breastfeeding and psychosocial factors. There is evidence to suggest that fathers play an important role in the breastfeeding decision and that intended duration is a strong predictor of actual duration. Despite the inconsistency of the reported associations, this review highlights the fact that breastfeeding is multifactorial in nature and future programs aimed at promoting breastfeeding must take this into consideration.