Charles Ragin, The Comparative Method. (1987), Chapters 1, 2, 4, 5, 8.

Summary: Ragin proposes a model for comparative analysis that draws from qualitative and quantitative methodologies – this synthetic method, he believes, overcomes many of the problems inherent in both approaches. Ragin argues (against Lijphart) that the comparative method is superior to statistical methodologies in several ways (see below). Ragin’s Qualitative Comparative Analysis (QCA, or Boolean approach) combines case-oriented and variable oriented approaches into a single framework. The idea is to create a truth-table of cases, variables and outcomes, and then reduce down (algebraically) until you have the explanatory variables (i.e. F=ABc and F=ABd become F=AB, because the variable on which they differ is deemed unimportant).

Critique: As Rueda notes, the only real critique of Mill’s methodology that QCA overcomes is the problem of assessing multi-causality, other than that there’s no real advantage.


  • Case AGAINST Lijphart – the comparative method is good – better than statistical method

Chapter 1 – Distinctiveness of Comparative Social Science

  • Comparison is central to empirical social science
  • Virtually all social science is comparative
  • Emphasis on gulf between quantitative and qualitative work
  • Case analysis tends to be holistic and interpretive
  • Przeworsk and Teune focus on systemic-level variables – this is restrictive
  • “The boundaries of comparative social science, therefore, must be coterminus with a specific usage of macrosocial units” (5)
  • Comparativists do use macro-social units, so they must be operationalized
    • Macro-social units are central to comparative social science
  • Observational Unit: a unit used in data collection and data analysis
  • Explanatory Unit: the unit that is used to account for the pattern of results obtained
  • Comparativists are constrained by their subject matter – generally speaking, they are small-N

Comparative Method = the method of systematic comparatic illustionrations

  • Possible to argue that it is only useful when N is too small for statistical analysis

Comparative method is superior to statistical methods in several ways:

  • Stats are not combinatorial – they are piecemeal
  • Comparative Method produces explanations that account for every instance of a certain phenomenon
  • No need to pretend that there is a homogenous population
  • Comparative method forces case familiarity


Chapter 2 – Heterogeneity and Causal Complexity

The problem of finding order in complexity

  • Simplifying the complexity among combinations of cases and of characteristics of cases and then constructing a model of all the types that exist
  • Simplifying the complexity among combinations of causes of an outcome and then constructing a model of these causal combinations

Interests, Simplicity and Complexity

  • The degree to which a set of observations or cases is on population or many depends on the interests of the investigator and those of the intended audience (selection bias)

Causal Complexity

  • Importance of context to understanding causality
  • Interaction of causal effects – conjunction/combination is a key feature of causal complexity
    • Once complexity is admitted, the question becomes how the pieces fit together
  • Causation is complexly combinatorial in the social world, making the problem of order in complexity demanding

Analysis of Causal Complexity

  • Multiple conjectural causation can only be assessed with experimenst
  • This doesn’t work well for social scientists…
  • Two options:
    • Case-Oriented Approach
    • Variables and Relationships


Chapter 4 – The Variable Oriented Approach

Goals of Variable Oriented Comparative Research

  • Generality is given precedence over complexity
  • Study begins with hypotheses, then delineates the widest possible population of relevant observations
  • Implicit goal of parsimony

Importance of Statistical Control

  • A key feature of the variable oriented approach
  • Statistical control is qualitatively different than experimental control
    • DV is not examined under all possible combinations of IVs
    • Number of cases in most nonempty cells is likely small
    • Assumes that the meaning of the scores of the IVs is the same across all cases (regardless of the other IVs)
    • Problem of specifying relevant observations
    • Multi-variate statistical techniques contradict notions of multiple conjunctural causation
  • First two problems plague all non-experimental research
  • Remaining problems can be addressed through more sophisticated statistics
    • However, the data used by comparativists is often too weak
    • Also, interaction requires a strong set of assumptions for statistical assessments to work, and most social science data does not meet these requirements
  • The case oriented approach is hampered by its inability to handle increased number of cases
  • The variable oriented approach is hampered by its inability to handle increased complexity.





Chapter 5 – Combined Versus Synthetic Comparative Strategies

  • Weakness of case-oriented approach – Tendency toward particularizing
  • Weakness of variable-oriented approach – Tendency toward abstraction
  • Combined strategy = Apply both strategies to a problem
  • Synthetic strategy = Integrating several features of both approaches
  • Ideally, a combined strategy should allow the investigator to consider both structural factors and factors reflecting historical processes and human agency
  • The two ends of the methodological continuum have clear theoretical (structure vs. agency) biases

Three Combined Strategies

  • Time series analysis is not good for structural explanation, as key structural elements tend not to change over time

Elements of a Synthetic Strategy

  • Should be capable of addressing a large number of cases
  • Should embody as much of the strict comparative logic of experimental design as possible
  • Should allow investigators to formulate parsimonious explanations
  • Should be analytic and should allow holistic investigation through its constitutive parts
  • Should allow consideration of alternative explanations


Chapter 8 – Applications of Boolean Methods of Qualitative Comparison

  • Ragin runs his QCA method to three cases – nation building, subnations, and the empirical typologies of organizations
  • The method itself relies on creating truth tables of all variables, cases and outcomes which produce at least one case per possible combination