A data set on participants in wars and whether they won, lost, or fought to a stalemate. Useful for teaching multinomial models.
Format
A data frame with 590 observations on the following 10 variables.
micnuma numeric confrontation identifier
confnamea name for the confrontation
ccode1a Correlates of War state code for a participant
ccode2a Correlates of War state code for an adversary
outcomea character vector for the war outcome for the participant, one of "Win", "Lose", or "Draw"
lowfatrat1the minimum estimated fatalities the participant imposed on the adversary (divided by its military personnel), divided over the maximum fatalities the participant incurred (divided by its military personnel) plus the minimum estimated fatalities the participant imposed on the adversary (divided by its military personnel).
relpow1the highest composite index of national capabilities (CINC) score of the participant, divided over the max of the participant and adversary over the course of the war.
sidea1a dummy variable identifying whether
ccode1is on the side of the confrontation responsible for the first event that led to the war.maxquds1the maximum level of democracy for
ccode1observed over the course of the war.bilata dummy variable identifying whether the war was bilateral.
Details
I created these data for a lecture in my international conflict course at Clemson University. That lecture was on March 8, 2022. The data are basically "as is" from that date. Use with caution. The goal for these data is to provide an example data set for which a multinomial model would be useful.
Data omit World War 2 (MIC#0258) and the Lithuanian War of Independence (MIC#2604). These two cases are conspicuous in the data as the only instances in which a state switched sides during war.
Confrontation data and fatalities come by way of what would eventually become v. 1.0 of the Militarized Interstate Confrontation data set. We no doubt revised a lot of these figures before that publication came out. Definitely use with caution.
War outcomes hand-coded by my discretion. Feel free to disagree with me. Again, use these data with caution and don't ascribe too much inferential value to them.
peacesciencer is responsible for the capabilities, military personnel,
and democracy data. Democracy data were originally in the ccode_democracy
data and came by way of Xavier Marquez' QuickUDS package. Capabilities and
personnel data are version 6.0 of the Correlates of War National Material
Capabilities data.
To use the Iran-Iraq War (MIC#2115) as an illustration, consider that at least 95,147 Iraqis died during the war (in the version of these data). At its peak, Correlates of War says Iraq had an estimated million soldiers. At most, Iran lost 255,824 troops in the war. At its peak, Iran had around 654,000 soldiers. Iran's fatality ratio comes out to around .196.
References
Gibler, Douglas M., and Steven V. Miller. 2023. “The Militarized Interstate Confrontation Dataset, 1816-2014.” Journal of Conflict Resolution 68(2–3): 562–86
Marquez, Xavier, "A Quick Method for Extending the Unified Democracy Scores" (March 23, 2016). doi:10.2139/ssrn.2753830
Miller Steven V. 2022. “peacesciencer: An R Package for Quantitative Peace Science Research.” Conflict Management and Peace Science, 39(6), 755–779.
Pemstein, Daniel, Stephen Meserve, and James Melton. 2010. "Democratic Compromise: A Latent Variable Analysis of Ten Measures of Regime Type." Political Analysis 18(4): 426-449.
Singer, J. David, Stuart Bremer, and John Stuckey. (1972). "Capability Distribution, Uncertainty, and Major Power War, 1820-1965." in Bruce Russett (ed) Peace, War, and Numbers, Beverly Hills: Sage, 19-48.
Singer, J. David. 1987. "Reconstructing the Correlates of War Dataset on Material Capabilities of States, 1816-1985" International Interactions, 14: 115-32.