Replicating the Resource Curse: Ross (2004) and Qualitative Replication

• In Event: Dealing with the Replication Crisis

Abstract

The ‘replication crisis’ in quantitative social science has motivated a re-assessment of previous findings and a re-thinking of approaches to research to increase transparency. Unfortunately, these changes have been slower in coming amongst scholars that use qualitative methods. Many have questioned whether applying the model of replicability to qualitative work is appropriate or, even, possible. While the question of appropriateness is outside of the scope of this paper, we do demonstrate that it is indeed possible and can provide illuminating results. We provide a model for qualitative replication using Michael Ross’ (2004) seminal paper on the connection between natural resources and civil war. The paper, which has been cited approximately 1000 times, uses information from thirteen cases to uncover the mechanisms that connect natural resource wealth to the onset, severity, and duration of civil wars. Using Ross’ sources and his own case notes, we sought to replicate the study and verify its findings. We find that roughly twenty percent of the case observations are mis-coded in some way (conceptual issues, measurement issues, missing data), calling in to question the validity of Ross’ findings. However, the replication exercise demonstrates a way forward—how we might design future research to further adjudicate to Ross’ claims and how we might apply lessons from this study to future qualitative replication efforts.

Authors

• Megan Becker, University of Southern California

• Jonathan Markowitz, University of Southern California