The death toll from wars around the globe in the second half of the twentieth century should be increased by a factor of three, according to a recent mathematical study by Ziad Obermeyer and colleagues at the Institute for Health Metrics and Evaluation in Seattle, Washington.

Estimating a war death toll is usually done by a mixture of eyewitness reports and media coverage. This involves extrapolating large numbers from sketchy data — read more about how this is done in the Plus article Body count.

Uppsala University in Sweden, and the Peace Research Institute in Norway, both keep death toll records estimated from media coverage, but Obermeyer's study suggests that the recorded death toll for 20th century wars could have been up to three times higher than they record.

The researchers looked at the death toll estimates gathered by the World Health Organization (WHO) — these numbers are extrapolated from telephone interviews with individuals with family members who may have died. This method is considered to be more accurate than gathering information from media reports.

In most cases, the WHO surveys recorded much higher numbers of dead than the Norwegian and Swedish databases. For example, the WHO figures suggest that more than twice the number of people died in Vietnam than previously thought — currently recorded at two million. On average, across 12 countries, the WHO figures are three times bigger. If true, then the average annual death toll for wars between 1985 and 1994 was 378,000.

During the 50 years covered by the study, Obermeyer suggests that there were 269,000 deaths in Bangladesh and 141,000 in Zimbabwe — nearly five times more than previously thought — and conflicts in Sri Lanka, Bosnia, Georgia and Laos are also estimated to be more costly than previously thought. However, in other countries, such as Burma, Ethiopia, Guatemala, Namibia and The Philippines, the death tolls dropped.

The study also found that a controversial report in 2006, which estimated the death toll after the invasion of Iraq at 655,000, may have been an over-estimate. You can read more about this study in the Plus article Body count. Obermeyer revises this number down to 184,000.

One downside of the study is that it only counts conflict fatalities and not deaths that have arisen from infectious diseases, which often afflict poor countries after war.