Causal Identification When Nothing Is Certain: Working with the UN in Darfur

• In Event: Bias in Data Collection and Interpretation

Abstract

This paper proposes a theoretical framework to understand the challenges of conducting mixed-method, causally-identified research in extremely fragile, conflict-affected settings, then draws on a case study from Darfur to present a set of tools and approaches to overcome those challenges. In both academic and policy circles, demand exceeds supply for rigorous research with project-specific quantitative and qualitative data collected in low-capacity and data poor areas where conflicts are active or nearby. The paucity of research is particularly striking with respect to work that seeks to understand the effects of complex large scale interventions, such as those undertaken by multilateral organizations. Empirical methods exist to conduct such research, yet their application remains rare, even when compared with other contexts that present logistical challenges. We propose that this gap can be explained by the higher levels of uncertainty and the scarcity of information available to research teams. Both of these factors – greater uncertainty and limited information – are exacerbated by 1) the nature of fragile conflict-affected contexts and 2) the nature of projects funded and implemented by large multilateral organizations. Through the case of a research project investigating a large-scale UN-supported peacebuilding project in Darfur, we present a set of tools to overcome these challenges.

Authors

• Sara Abdus Salam Ansari, Brown University

• Paul Thissen, International Initiative for Impact Evaluation (3ie)