This paper proposes a new method of measuring excess risk of child mortality in cross-sectional surveys, which is applied to DHS I data for Burundi, Uganda and Zimbabwe. The expected child mortality experience is estimated for each mother on the basis of child's age, mother's age at child's birth and her parity, and compared with her observed experience. Mothers who exceed their expected child mortality experience and also had more than one child die are considered to have excess child mortality. Zimbabwe had the greatest concentration of child deaths as measured by a simple ratio of mothers to deaths, but when observed experience was compared with expected it had less than half as many excess deaths as Uganda and Burundi. In all three countries mother's education had a strong negative association with the risk of excess child mortality, and in Zimbabwe and Burundi there were significant regional differences.