CNN reported today that between 63,000 and 100,000 people have died as a result of the May 3 cyclone in Burma. According to the Washington Post, the death toll from Monday's earthquake in China has exceeded 12,000 and is expected to rise. Where do natural-disaster death estimates come from?
Eyewitnesses and guesswork. Government relief workers and agents from NGOs assess stricken neighborhoods for casualties. They literally count bodies, take down reports from district officials or locals who have lost family members, and make estimates based on damage to infrastructure. (If there are 20 people missing and they all worked in a building that collapsed due to a tremor, the relief workers might count those 20 people as dead.) The workers then report back either to a government agency in charge of emergency assistance or to the United Nations Office for the Coordination of Humanitarian Affairs. An aggregate figure filters down to national media outlets.
U.N. estimates are often higher than the local government's. (Burma's state television, for example, currently reports that around 34,000 people have died, while the U.N. claims the actual toll is closer to 100,000.) That's because the U.N. tries to account for regions that have not yet been assessed on the ground by using satellite footage of the wreckage and prior demographic information. If they know that 15 percent of the population perished in one village, they assume the same percentage died in a similarly affected village. Even without such projections, U.N. numbers are often higher because they err on the side of overestimation to ensure an adequate relief response. Local governments, on the other hand, might underreport to save face or to prevent international organizations from assisting opposition groups.
On Sunday, Oxfam warned that the death toll in Burma could reach 1.5 million without massive humanitarian intervention. To arrive at that figure, Oxfam used the U.N.'s 100,000 estimate as a base. Then they used research from previous natural disasters and demographic analysis (children and the elderly are less likely to survive, etc.) to predict that 15 times that many people could die from typhoid, malaria, dengue, cholera, and other diseases.
Newspapers and wire services don't have the resources to verify mortality statistics independently, so you'll often see two or more numbers cited in the same article. A recent article on Burma in the Canadian Globe and Mail, for example, gives the Burmese government's official number of dead and missing; the United Nations' far higher, unofficial number; and Oxfam's 1.5 million "in danger of dying" estimate. In the first couple days after a disaster, death tolls as reported by the media are often low and then creep upward. (On May 6, news sources were reporting just 10,000 dead.) That's because the first numbers usually come from government agencies that 1) may be underreporting or 2) are pressured to make approximations before relief workers have gauged the scale of destruction. Body counts also rise as victims die from indirect causes.
Neither eyewitness reports nor the projections done by the U.N. are thought to be entirely accurate. In the aftermath of a disaster, it's difficult to prevent double- or triple-reporting of individual deaths or to independently verify estimates from district officials. Predictions about inaccessible villages based on similar surveyed areas aren't entirely reliable, either. After the tsunami, for example, relief workers assumed devastation in remote islands that turned out to be relatively unscathed because the locals had escaped to higher ground.
Got a question about today's news? Ask the Explainer.
Explainer thanks Stephanie Bunker of the United Nations Office for the Coordination of Humanitarian Affairs, Mike Kiernan of Save the Children, Liz Lucas of Oxfam, and Court Robinson of the Johns Hopkins Bloomberg School of Public Health.