![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
Abstract
Dealing with questions of war and peace and understanding the causes of interstate conflict is a primary goal of the field of international relations. In order to study interstate conflict in a rigorous manner, scholars have relied on established rules and procedures for gathering information into coherent data sets. Among those data sets is the Militarized Interstate Dispute (MID) data. In this paper we first outline the data-collection process for the MID3 data. Second, we introduce two new data sets emerging from the project, “MID-I” and “MID-IP.” Third, we present relatively small changes in coding rules for the new MID3 data and some descriptive statistics. The statistics indicate that the MID3 data are remarkably similar to the MID2.1 version, varying in some minor and predictable ways.
An earlier version of this paper was presented at the Annual Meeting of the American Political Science Association, August 2003. We are grateful to Resat Bayer, Doug Bond, Dennis Foster, Nils Petter Gleditsch, Kyle Joyce, Paul Senese, and John Vasquez for comments on this paper. All the data sets described in this paper are available on the COW2 website, http://cow2.la.psu.edu
Notes
1 It is impossible to exaggerate Stuart Bremer's importance to the MID3 project. It is fair to say that without his careful preparation of the project, direction in the construction of the coding rules, creation of the computer software that managed the data, guidance in the processing of the data, and leadership in virtually all phases, the project would never have been attempted, much less completed. Bremer's death in October 2002, as the project was reaching its conclusion, deprived him of the opportunity of seeing his central professional concern of several years come to fruition. All users of the MID3 data set are in his debt.
2 This forum was open to the public during the project. An abbreviated version of the forum will be made available at a later date.
3 Bennett was assisted by Hilde Ravlo and Faten Ghosn. The updated and corrected set of disputes from 1816–2001 is available on the COW2 website in version 3.021.
4 The PIs were more than ably aided by their research assistants. They were: J. Michael Greig (for Diehl); Zhang Wanfa, Randolph Jones, Samuel Nailling, and James Carruth (for Geller); John Cotter (for Gibler); Burcu Savun, Liying Ren, and June Park (for Gochman); Michael Allison and Kursad Turan (for Hensel); Lesley Terris (for Maoz); Dennis Foster, Diane Dutka, and David Sobek (for Palmer); Kevin Sweeney and Amy Oakes (for Pollins); Greg Casteel, Brandon Valeriano, Karen Petersen, and Hilde Ravlo (for Ray); Odul Celep (for Regan); and Navin Bapat, Xiang Zhou, Alistair Baldwin, and Eric Carr (for Stoll).
5 A few pairs of states account for a good portion of the interstate disputes in these regions. The majority of the Asian disputes are between North Korea, South Korea, and the United States; China, Taiwan, and the United States; Japan and North Korea; and India and Pakistan. European disputes mainly involved NATO allies and Yugoslavia, as well as disputes between Greece and Turkey. In the Middle East, disputes involving Iraq, Kuwait, and the Western allies represent the majority of the disputes, followed by disputes between Turkey and Iraq. As for Africa, Sudan and Uganda as well as Rwanda and Uganda witnessed the largest number of disputes. The majority of disputes in the Western Hemisphere (about 85%) were in Latin America; disputes between Honduras and Nicaragua were particularly numerous.
6 We refer to the Incident Coding Manual as “ICM,” and the Dispute Coding Manual as “DCM” hereafter.
7 There are numerous examples of stalemates, but a typical one is MID 4332. In February 1994, Hungary put its border guards on alter, fearing a reprisal from neighboring Yugoslavia in reaction to the anticipated bombing campaign against that country by Western forces. No action by Yugoslavia against Hungary occurred, however, thus leading to the dispute being coded a stalemate.
8 “Initators” are commonly identified as being on Side A on a dispute's first day, though the stipulation that Side A be the revisionist side is at times added.
9 We have included in the Appendix a list of candidate disputes that for one reason or another were not included in the final set of MIDs. That list is meant to provide a better understanding of the implementation of the coding rules.
10 It is, of course, possible, that conflict in the 1993–2001 period differs from that in the previous era for a variety of reasons. For instance, the collapse of the Soviet Union conceivably altered the nature of some interstate conflict sufficiently for there to be a change in the descriptive statistics as we have presented them. Investigating such hypotheses goes beyond the scope of the current article.
11 In the 1816–1992 data, about 16.2% of the states involved in MIDs had “territorial” goals, while 22.3% had “policy” goals. In the 1993–2001 data, about 13.6% of the participants had revision types that were territorial, and 22.1% were coded as having policy goals.
12 Cases in which concrete numbers were not available for both sides an estimate of the range was provided.
13 The seven categories are: (1) No fatalities; (2) 1–25 fatalities; (3) 26–100 fatalities; (4) 101–250 fatalities; (5) 251–500 fatalities; (6) 501–999 fatalities; and (7) more than 999 fatalities. We excluded from analysis the cases where fatality estimates were missing.
14 To be consistent with MID2.1, fishing rights, oil rights, and other issues pertaining to water are classified as policy issues and not territorial.
15 It is important to bear in mind that each incident has only one target state, so that a single action directed against more than one state is coded as multiple incidents.
16 No incident participants were coded as suffering between 501 and 999 fatalities in one incident. We therefore did the analysis with six categories of fatalities (omitting 501–999) rather than seven, as we did in the prior analyses.
17 The excluded category, as before, contains those instances where the desired revision was either “not applicable,” “regime/government,” or “other.”
∗ Deceased.