Captives in Interstate Wars
Because the focus of the book is on the treatment of prisoners in wars between states, the data for assessing the patterns and determinants of prisoner abuse consist of interstate wars from the widely used Correlates of War (COW) project.44 The time period of the study ranges from the end of the nineteenth century, starting with the Spanish-American War of 1898, to the 2003 interstate portion of the Iraq War. While references are sometimes still made to earlier conflicts, for the main analysis prior wars were excluded because of the lack of accessible and reliable data on prisoner treatment in many instances. Near the turn of the twentieth century a wider range and more comprehensive set of scholarly and other primary sources became available that allows for determining in a more comprehensive manner the ways prisoners were treated in a given war.
One of the benefits of selecting this time period is that it matches quite closely the development and codification of the laws of war. Admittedly, the first most commonly accepted international treaty dealing specifically with wartime conduct was negotiated much earlier in the 1864 Convention for the Amelioration of the Condition of the Wounded in Armies in the Field. As the title of the convention makes clear, however, it was fairly narrow in scope and dealt only tangentially with issues related to prisoners of war.45 In contrast, the 1899 Hague Conventions are widely considered to represent the foundation of the modern laws of war. The Hague treaties provide the basic definition and principles for the treatment of prisoners, which have largely continued in a similar form to the present, though with further inclusion and elaboration of certain rights and responsibilities. The wars considered in this data set thus involve belligerents who at least had an understanding of the general prevailing international rights and obligations concerning prisoners, even if they did not subscribe to them. The greater availability and quality of information on prisoner treatment is likely a further beneficial side effect of the establishment of international laws during this period, providing a focal point for governments and nonstate actors to pay greater attention to, and document the care of, captives during wartime.
Beyond the fact that some of the most notorious instances of prisoner abuse (along with praises of humane treatment) took place in wars between states, a more pragmatic reason for limiting the main analysis to interstate wars is the difficulty of reliably measuring prisoner treatment in each individual case. Data are also generally of a higher quality for interstate wars compared to other conflict types, which allows for greater accuracy and confidence when evaluating each episode of prisoner treatment.
The list of wars examined here includes a few notable changes from the standard COW interstate war data set. As has become increasingly common in the quantitative conflict literature, several long multiactor wars are divided into a series of separate military confrontations to more accurately reflect the actual fighting and political relations between the relevant belligerents. World War I was separated into four individual conflicts, World War II into nine separate conflicts, and the 1990-91 Persian Gulf War into the Iraq-Kuwait and U.S. Coalition-Iraq conflicts.46 The final data set contains seventy-nine wars in total, which are summarized in table A.4 in the appendix.
Each case of prisoner treatment is composed of a separate warring-directed- dyad. This means that in a given war there are two cases for any pair of opponents, where each is a potential violator or victim. For instance, in the 1982 Falklands War, one case examines British treatment of Argentinean prisoners while the other looks at Argentinean treatment of British prisoners. In wars involving more than two parties, there are two corresponding cases for each pair of opposing states.47 In theory, this potentially entails an enormous number of cases given the myriad combinations of opponents in large conflicts such as the world wars or the Korean War. In reality, however, most wars involve only a handful of states that are relevant when assessing the treatment of prisoners.