Many of the historical and contemporary phenomena in which social scientists are interested are difficult to study using traditional methods of comparative analysis.  Since most cases are complex systems – marked by interdependence and operating at multiple levels of analysis at once – controlling comparisons to adjudicate causality is fraught with difficulty.  This chapter argues that scholars can use historical archival research to help disaggregate the temporal and spatial properties of the phenomena we hope to compare while also tracing connections among those disaggregated elements.  Specifically, practices associated with archival inquiry – classifying, contextualizing, layering, and linking – allow us to identify the boundaries around subsystems that can be treated as relatively independent while identifying the hierarchical connections tying those substemic activities together. The chapter concludes by showing how William Tuttle’s masterful history of the Chicago Race Riot of 1919 provides a template for comparing complex cases.

How does nationalization of operations causally increase government revenue? Using the case of oil politics in prerevolutionary Iran, this chapter analyzes production elasticity and reduced informational asymmetries after the Shah of Iran nationalized the oil operations of BP and its partners. The chapter draws on a combination of conversations documented in archival records and quantitative analysis of historical fiscal data from the BP Archive. It begins by explaining the Shah’s surprise decision in January 1973 to reconfigure the National Iranian Oil Company from its role as a passive observer to a fully operational oil company able to set production levels and prices. The chapter then presents findings using BP’s revenue projections as counterfactuals to estimate the causal effects of the Shah’s decision with respect to whether NOC reform increased government take of oil revenues in 1974–1975 and whether revenues collapsed after retaliation by international oil companies in 1976 to strip the NOC of its ability to sell oil on the global market. The chapter illustrates both the nuances and consequences of operational versus nonoperational NOCs for fiscal strength.

We assemble a novel dataset in order to test theoretical propositions we develop on how states intervene in the elections of others. We start off with a random or representative sample of about 10 per cent of all elections since the end of World War II. Each of these is a case for us. We add a set of potential interveners, powers and organizations that may have a stake in intervening. We scour primary and secondary sources to extract information on how the government and opposition in the target state view relations with the potential interveners. We also extract information on whether and how the intervener acted in support of processes, or of candidates. This chapter is a codebook of how we constructed these novel, theory relevant variables. In addition, we supply extensive case-study notes. In those notes, we connect each of our coding decision to specific strings of text in the sources we used. The resulting database allows us to test the key propositions we develop.