Droughts in Ethiopia have commonly been associated with increased child mortality. Early indications were that the 2002/03 drought, which affected 13.2 million people, was no exception, despite a large relief operation. Humanitarian agencies reported sharp increases in child deaths and pockets of acute distress in some hard-hit localities. In response, the 2004 Ethiopia Child Survival Survey (ECSS) was designed to investigate the impact of the drought on child survival in the general population. The survey covered 4816 households in both drought-affected and non-drought affected, as well as rural and urban localities. Data from the ECSS indicate that child mortality was indeed higher in drought-affected areas. However, a closer analysis reveals that this differential is attributable to chronic conditions in those localities, rather than the immediate impact of the 2002/03 drought. Multivariate analysis was used to construct a model for the determinants of child survival in the sample population. Household-level demographic factors, household-level food and livelihood security, community-level economic production, and access to potable water, were predictive of child survival. Additionally, household receipt of food aid had a small but significant positive association with child survival, even though the ECSS cannot determine either the underlying causal mechanisms of this association or the role of confounding factors. Nonetheless, it is remarkable that the most extensive drought in the country's modern history passed without a measurable increase in child mortality among the general population. Yet Ethiopian children still suffer unacceptably high rates of chronic malnutrition and poor life chances, and large populations continue to live at the brink of destitution and calamity.