Abstract: For almost three decades the Correlates of War project has supplied the data used by most studies of external and internal armed conflict. The COW war datasets have a relatively high threshold for the number of battle deaths. For external conflict the COW project has collected an alternative measure with a lower threshold on violence, the Militarized Interstate Disputes dataset. For internal conflict, the COW project has not yet generated a measure with a lower violence threshold. The Conflict Data project at Uppsala University has collected data on low-level conflict, internal and external, from the end of the Cold War. Here, that dataset is extended back to the end of World War II. This dataset was generated in two steps: First, a candidate list was generated on the basis of 13 existing datasets. Sec ond, the candidate conflicts were checked systematically on the basis of the Uppsala criteria. The re sulting list contains 261 conflicts for the period 1946–99, including 180 intrastate conflicts, 21 intra state conflicts with external participation, 19 extra-systemic conflicts, and 41 interstate conflicts. Patterns across regions and over time are described and compared with the datasets from which the candidate list was derived. The criteria for measuring armed conflict are assessed.
A major revision of this paper was published as: Gleditsch, Nils Petter; Peter Wallensteen, Mikael Eriksson, Margareta Sollenberg & Håvard Strand (2002) Armed conflict 1946–2001: A new dataset. Journal of Peace Research 39(5): 615–637
