Case Studies and Theory Development in the Social Sciences - Alexander L. George [215]
447
For a formal definition of causal effects, see Keohane, King, and Verba, Designing Social Inquiry, pp. 76-82. For an illustration of how case studies and statistical studies contribute complementary kinds of knowledge to a research program, see Chapters 1 and 2 and Andrew Bennett and Alexander L. George, “An Alliance of Statistical and Case Study Methods: Research on the Interdemocratic Peace,” Newsletter of the APSA Organized Section in Comparative Politics, Vol. 9, No. 1 (Winter 1998), pp. 6-9.
448
Clayton Roberts, The Logic of Historical Explanation (University Park: Pennsylvania State University Press, 1996).
449
Ernest Nagel, The Structure of Science: Problems in the Logic of Scientific Explanation (New York: Harcourt, Brace and World, 1961), p. 574.
450
Roberts, The Logic of Historical Explanation, pp. 9, 15.
451
Ibid., p. 20. Roberts notes that “colligation” has also been used by some writers to refer to “the grouping of events under appropriate conceptions.” By this he evidently means subsuming single instances of a given type of phenomenon (e.g., revolution, deterrence) under a class of such events. Roberts prefers to refer to this second meaning of colligation as “classification” and drops it from his preferred definition of colligation. This point is worth noting here since use of case studies for theory development, as in structured, focused comparison, is based on studying one or several cases, each of which is an instance of a class of events.
452
Ibid., p. 66. Roberts’ discussion of microcorrelation is less clear than in the statement quoted here.
453
This type of complex theory is referred to by Abraham Kaplan as “concatenated theory.” See Kaplan, Conduct of Inquiry (San Francisco, Calif.: Chandler, 1964), p. 298: “A concatenated theory is one whose component laws enter into a network of relations so as to constitute an identifiable configuration or pattern. Most typically, they converge on some central point, each specifying one of the factors which okays a part in the phenomenon which the theory is to explain.”
454
Our construction of Skocpol’s analysis is provisional and subject to reconsideration. A somewhat different construction of the analytical structure of Skocpol’s study is suggested by James Mahoney (personal communication). Skocpol’s study has generated a great deal of critical comment, much of it questioning her reliance on Mill’s methods. However, as Jack Goldstone has pointed out in one of the most discerning and balanced of the evaluations of her study, Skocpol supplemented use of Mill’s methods with considerable use of process-tracing, a fact that she did not clearly convey. Compare Jack Goldstone, “Methodological Issues in Comparative Macrosociology” (forthcoming); see also Goldstone, “Revolution, War, and Security” (manuscript, 1997).
455
This calls to mind, of course, Giovanni Sartori’s well-known metaphor of “moving up and down a ladder of generality.” See Sartori’s seminal article, “Concept Misformation in Comparative Politics.”
456
Roberts, The Logic of Historical Explanation, pp. 66-67.
457
Ibid., pp. 87-88.
458
Ibid., pp. 145-159.
459
Ibid., p. 155.
460
The uses and limitations of counterfactual analysis are also discussed in Chapter 8. See also Kaplan, Conduct of Inquiry, pp. 21, 91, 160; and James D. Fearon, “Counterfactuals and Hypothesis Testing in Political Science,” World Politics, Vol. 43, No. 2 (January 1991), pp. 169-195. However frequently counterfactual analysis is employed, it lacks explicit criteria and standards for distinguishing good practice from often highly speculative, less disciplined uses. An important effort to explicate standards for counterfactual analysis is Philip E. Tetlock and Aaron Belkin, eds., Counterfactual Thought Experiments in World Politics (Princeton, N.J.: Princeton University Press, 1996).
461
See, for