Scholars who write about comparative historical methods sometimes make it appear that the research tradition has a single basic approach for identifying patterns of causation. Yet, in fact, comparative historical analysts employ a wide range of strategies of causal assessment in their substantive research. These strategies encompass both methodologies for juxtaposing cases with one another and methodologies for analyzing processes that take place within individual cases. And they include both techniques of causal assessment designed to identify the necessary or sufficient causes of an outcome and tools for locating causal factors that covary with outcomes in linear patterns. Rather than narrowly limiting themselves to any one approach, then, comparative historical researchers are eclectic in their use of methods.

In this essay, I attempt to analyze systematically these different strategies of causal analysis. My main objectives are to specify the concrete procedures entailed in the strategies, discuss their underlying assumptions about causality, and assess their comparative strengths and weaknesses. Along the way, I engage the long-standing debate about small-N versus large-N research. I devote particular attention to the ways in which different comparative historical methods are or are not compatible with the assumptions that guide causal inference in conventional statistical methodologies. My hope is that this discussion will help clear up some of the misunderstandings that have developed between advocates of small-N and large-N research and refocus attention on the real points of contention between the traditions.