This zip file provides all the necessary material to replicate all of the results in our article.
When putting together our replication material, we noticed two small typos in our STATA code. These typos do not affect any of the analyses that employ data from Mozaffar, Scarritt and Galaich (MSG); they only affect the analyses relating to the corrected data that we construct. As we illustrate in the article, our claims about the effect of either ethnic fragmentation and district magnitude on the size of African party systems does not depend on whether we use the MSG data or our own corrected data. As such, the claims that we make are qualitatively unaffected by the two typos that we found. However, we felt obliged to see exactly how our results change when we correct these two typos. The zip file that we provide here does precisely this. As we demonstrate, the claims that we made in our article remain supported.