The role of social science is to explain social events. Explanation consists of applying general sentences or, more precisely, theories or sets of such general sentences, to particular events. If the explanation is to be general, parsimonious, and causal, then the accumulation of knowledge – confirmation and/or modification of theories – must involve comparative research. However, explanation in comparative research is possible if and only if particular social systems observed in time and space are not viewed as finite conjunctions of constituent elements, but rather as residua of theoretical variables. General lawlike sentences can be utilized for explanatory purposes. Only if the classes of social events are viewed as generalizable beyond the limits of any particular historical social system can general lawlike sentences be used for explanation. Therefore the role of comparative research in the process of theory-building and theory-testing consists of replacing proper names of social systems by the relevant variables.
Important Insights: Przeworski and Teune treat comparative analysis in an algebraic way. They propose the importance of substituting proper names for variables in order to develop general theories – this is what bridges historical observations and general theories.
Critique: Przeworski and Teune suggest that the four elements of explanation (accuracy, generality, parsimony and causality) exist in a sort of zero-sum relationship. They do not, however, offer any insight into the ordering or importance of these factors. They give a lengthy treatment of causality, though it isn’t clear (to me at least) that explanation must necessarily be causal in the sense that they describe (see Charles Ragin The Comparative Method (1987)) in particular in respect to causal explanation that does not rely on such a simplified description).
Przeworski and Teune articulate two research designs. The “Most Similar Systems” Design focuses on intersystemic similarities and differences. Common factors are seen as ‘controlled for’ and the differences constitute the explanatory variables. Belief that the more similar the systems, the better they are for comparative analysis. Conversely, the “Most Different Systems” Design does not operate at the systems level, and instead seeks to discredit systemic factors as the explanatory variables. The logic is if the same phenomena occurs in a homogeneous population (i.e. all auto workers), then systemic level factors are irrelevant, or at least not the most relevant.
Important Insights: The focus on the “Most Different Systems” Design is novel, especially since within this framework the authors allow for the role of factors in contributing to the explanation without putting any specific emphasis on a particular level of analysis.
Critique: There seems to be a tension present in the fact that “findings desirable in the most similar systems design are highly undesirable in the most different systems design and vice versa,” the implication being that there may be a \’correct\’ method for specific questions. What happens if both methods warrant an explanation?