Death as a measure of duration of conflict

This paper argues that the number of deaths in a conflict can be a measure of duration and demonstrates that there are information gains to this approach. The well-known conflict database of International Peace Research Institute is compared with the database of Center for Systemic Peace which includes data on the number of deaths in addition to length of conflict. There are differences in the two datasets. The number and distribution of conflicts vary however duration analysis yields results that are robust over datasets. We also show that the number of deaths, as a measure of duration, challenges some of the results based on the number of years as a measure of duration. We can also argue that economic integration coincides with shorter duration of conflicts in terms of both deaths and years. International conflicts have a shorter life and have fewer deaths than domestic conflicts. Geographic or ethnic characteristics of conflicts are not robust over datasets nor over measure of duration. -- Duration models ; violence and wars