Case Studies and Theory Development in the Social Sciences - Alexander L. George [185]
58
A complex debate has emerged over whether, as Dion argues, claims of necessity should be tested only against cases that are positive on the outcome of interest, and claims of sufficiency should be tested only against cases that are positive on the independent variable of interest. Jason Seawright has used sophisticated Bayesian reasoning to argue, contra Dion, that studying diverse cases, in addition to looking for single cases that could disprove necessity or sufficiency, can yield stronger and more efficient tests of necessity and sufficiency; others dispute this. See Jason Seawright, “Testing for Necessary and/or Sufficient Causation: Which Cases Are Relevant?” Political Analysis, Vol. 10, No. 2 (Spring 2002), pp. 178-193; for the critiques of Seawright and his response see Kevin Clarke, “The Reverend and the Ravens,” Political Analysis, Vol. 10, No. 2 (Spring 2002), pp. 194-197; Bear F. Braumoeller and Gary Goertz, “Watching Your Posterior: Sampling Assumptions, Falsification, and Necessary Conditions,” pp. 198-203; and Seawright’s rejoinder, “What Counts as Evidence? Prior Probabilities, Posterior Distributions, and Causal Inference,” pp. 204-207. Without attempting to resolve this debate, we merely note that apart from Seawright’s arguments, there may be reasons for selecting diverse cases to test theories, including those of necessity or sufficiency. Researchers are often testing not just one necessary condition hypothesis, for example, but also alternative hypotheses that may require different cases to test. Also, the causal mechanisms being tested may be different from the independent and dependent variables used to identify and select cases. We return to this question of case selection for testing claims of necessity and sufficiency in Chapter 2, which looks at the literature on whether it is a sufficient condition for peace between two countries if both of them are democratic.
59
Collier and Levitsky, “Democracy with Adjectives,” pp. 430-451.
60
King, Keohane, and Verba, Designing Social Inquiry, p. 225.
61
Donald Campbell, “Degrees of Freedom and the Case Study,” Comparative Political Studies, Vol. 8, No. 8 (July 1975), pp. 179, 181-182.
62
King, Keohane, and Verba, Designing Social Inquiry, pp. 119-120. King, Keohane, and Verba also acknowledge here that generating process-tracing observations can mitigate the problem of indeterminacy.
63
Olav Njølstad, “Learning From History? Case Studies and the Limits to Theory-Building,” in Olav Njølstad, ed., Arms Races: Technological and Political Dynamics (Newbury Park, Calif.: Sage Publications, 1990), pp. 220-246. See also our discussion of this problem in Chapter 2.
64
Timothy McKeown, “Case Studies and the Statistical World View,” International Organization, Vol. 53, No. 1 (Winter 1999), pp. 161-190.
65
Stanley Lieberson, “Small N’s and Big Conclusions,” in Ragin and Becker, eds., What is a Case? Exploring the Foundations of Social Inquiry, pp. 108-109, 113, 116; Achen and Snidal, “Rational Deterrence Theory and Comparative Case Studies,” pp. 160-161.
66
Useful here is the reminder by King, Keohane, and Verba (Designing Social Inquiry, p. 20), that parsimony is not an unalloyed goal, and that “theory should be just as complicated as all our evidence suggests.”
67
Alexander L. George and Timothy McKeown, “Case Studies and Theories of Organizational Decision Making,” in Robert Coulam and Richard Smith, eds., Advances in Information Processing in Organizations, Vol. 2 (Greenwich, Conn.: JAI Press, 1985), pp. 43-68; McKeown, “Case Studies and the Statistical World View.” goal of providing the strongest possible inferences