Missing the War: Handling Data Absence in (and Due to) War

• In Event: Limitations and Advances in Conflict Data

Abstract

Scholars and policy-makers use a variety of techniques to overcome information problems resulting from biased or absent data in (and because of) war. While fixes such as imputation or listwise deletion may be warranted and not unduly bias all findings, in some instances they are highly problematic. Scholars have also shown that some measurements and data that seem completely unrelated, including rainfall, are in fact affected by war. In these instances, the missingness and bias often affects scholars’ work in fields outside of political violence who may be entirely unaware of the issues. This paper undertakes a comprehensive review of the common tools and techniques used to cut through the “fog of war” and assesses what they might be missing empirically and theoretically by doing so. It includes a number of new, alternative fixes such as satellite data and scraping and other big data solutions to missingness. It finds that many fixes are used without adequate attention to the theoretical or empirical blind spots that they might generate. There is no single solution to the problem of missing and biased data related to war. However, highlighting the limitations of common approaches helps to focus attention on finding better ways of managing it and generating creative alternatives.

Author

• Bridget Coggins, University of California, Santa Barbara