The purpose of this study was to assess the utility of using maternal assessments of infant birth size as proxy measures for birth weight in Ecuador, a country in which a sizeable proportion of births take place at home, where birth weight is typically not recorded. Four thousand and seventy-eight women who experienced a live singleton birth between January 1992 and August 1994 were interviewed in the Ecuador Demographic and Maternal–Child Health Survey. All women were asked if their child was weighed at birth, his/her weight, and what they considered to be his/her birth size relative to other newborns. The consistency between birth size and birth weight measures was assessed, and the differences between infants with and without reported birth weights were explored. The authors conclude that maternal assessments of birth size are poor proxy indicators of birth weight. Estimates of low birth weight based on maternal assessments of birth size as very small should be recognized as underestimates of the actual prevalence of low birth weight. Moreover, infants for whom birth weights are missing should not be considered similar to those for whom weight was reported. Those without reported birth weights are more likely to be low birth weight. Thus, relying solely on reports of numeric birth weight will underestimate the prevalence of low birth weight.