Some evidence suggests that breast feeding is weakly but positively associated with cognitive function. This association has been robust to adjustment for various confounders. The aim of this paper is to determine if duration of breast feeding is associated with cognitive function in late childhood. Data was abstracted from the 1970 British Cohort Study. 11004 liveborn white singletons born during 5–11 April 1970 in the United Kingdom were followed from birth to 10 years. Cognitive function at 10 years is the dependent variable, a latent construct composed of one ability test and three performance measures. Estimates derived from multiple linear regression and structural equation modeling were compared. Effect sizes were estimated using standardized coefficients (SC). Differences in cognitive function according to breast feeding duration were estimated to be small by multiple linear regression (SC = 0.07) and much smaller and non-significant as estimated by structural equation modeling (SC = 0.02) after adjusting for parental socioeconomic status (SES), birth weight, parity, gestational age, maternal age and maternal smoking. Differences in cognitive function according to duration of breast feeding appear to be small and of little clinical importance as estimated by structural equation modeling.