Calendar versus Analysis Time: Reanalyzing the Relationship between Humanitarian Aid and Civil Conflict Duration

Previous work in International Studies Quarterly shows higher levels of humanitarian aid prolong civil conflicts. It also finds, among conflict–years in which aid is received, that this conflict-prolonging effect is more acute in insurgency-based civil conflicts, albeit with weaker supporting evidence. However, I show this work accidentally generated its conflict duration variable incorrectly, with the duration measuring time since January 1, 1960, not time since civil conflict onset. The duration values also exclude the first at-risk day for the first observation in each conflict, which drops true one-day durations from the estimation sample. I rerun the original analysis with the corrected duration coding and find evidence that supports the opposite of the author's main hypothesis: higher levels of humanitarian aid either have no effect on or shorten civil conflict duration. Additionally, the weak evidence for the author's second hypothesis mostly disappears, depending on the conflict's duration.