POLICYFORUM

he crisis of death and displacement in western Sudan began in February of 2003 and soon engulfed all three states of North, West, and South Darfur. To begin to comprehend the tragedy of the con-flict and the extent of the genocide (1, 2), it is necessary to have an accurate estimate of the number of deaths that occurred. Yet cur-rent estimates differ by more than an order of magnitude. This uncertainty results from difficulties inherent in surveying a war-torn region in Africa, as well as assumptions made by agen-cies trying to generate estimates. There is no way, as might happen in a natural disaster, to get an accurate body count; estimates must rely on interviews. Surveys from displace-ment camp samples must be substituted for unavailable population-based census data; extrapolating from limited samples to an entire population at risk is problematic. A quarter-century of famine and war has recon-figured nuclear families, making sampling units in surveys problematic. Current surveys also vary in recall periods and coverage. Finally, past estimates of Darfur mortality have been based on the dubious assumption of constant numbers of deaths per month