1 See for example, Horelick, Arnold, “The Cuban Missile Crisis: An Analysis of Soviet Calculations and Behaviour,” World Politics 16 (1964), 363–89CrossRefGoogle Scholar, and Allison, Graham, Essence of Decision: Explaining the Cuban Missile Crisis (Boston: Little, Brown and Company, 1971), 39–66Google Scholar.

2 Illustrative of this approach is Steinbruner, John, The Cybernetic Theory of Decision (Princeton: Princeton University Press, 1974)Google Scholar; Allison, Essence of Decision, 101–43; and Nathan, James A., “The Missile Crisis: His Finest Hour Now,” World Politics 27 (1975), 256–81CrossRefGoogle Scholar.

3 See for example, Allison, Essence of Decision, 185–244.

4 See for example, Stone, I. F., “The Brink,” New York Review of Books, April 14, 1966, 1–4Google Scholar; Dewart, Leslie, “The Cuban Crisis Revisited,” Studies on the Left 5 (1965), 15–38Google Scholar; and Steel, Ronald, “The Kennedys and the Missile Crisis,” New York Review of Books, March 13, 1969, 15–22Google Scholar.

5 The work of Jack L. Snyder is illustrative of the application of this approach to the Cuban Missile Crisis. “Rationality at the Brink: The Role of Cognitive Processes in the Failure of Deterrence,” World Politics 30 (1978), 345–65CrossRefGoogle Scholar.

6 See, for example, Mongar, Thomas, “Personality and Decision-making: John F. Kennedy in Four Crisis Decisions,” this Journal 2 (1969), 200–25Google Scholar.

7 See Holsti, Ole, Brody, Richard A. and North, Robert C., “Measuring Affect and Action in International Reaction Models: Empirical Materials from the 1962 Cuban Crisis,” Journal of Peace Research (1964), 170–89CrossRefGoogle Scholar.

8 See for example, Horelick, “The Cuban Missile Crisis.” For a detailed analysis and critique of this approach, see Allison, Essence of Decision, esp. chaps. 1, 2 and 7.

9 See for example, Stein, Janice and Tanter, Raymond, Rational Decision Making: Israel's Security Choices, 1967(Columbus, Ohio: Ohio State University Press, 1980)Google Scholar.

10 This study focusses on the six months, or twenty-six weeks, prior to the crisis. A pre-crisis phase of six months was considered sufficiently long for meaningful patterns and relationships between objectives and behaviour to emerge. A more serious problem arose from our efforts to determine the timing of the crisis. The CADIC project initially attempted to determine the onset of the crisis in quantitative terms, that is, as a significant increase (more than two standard deviations) from the normal range of relations (NRR) of the previous two years of coercive interaction between the protagonists. The data bases on which to establish such an NRR and the deviation scores did not exist and could not be easily created. The project turned therefore to a qualitative definition of crisis as that period which begins and ends with a threshold event perceived by both parties. Most analysts would agree that the onset of the Missile Crisis—the critical threshold event—was the national television address by President Kennedy on October 22, 1962, when the discovery of the Soviet missiles as well as measures of mobilization and quarantine were announced.

11 All verbatim statements of objectives by leading Soviet and American decision makers were extracted from The Current Digest of the Soviet Press, The New York Times, The Department of State Bulletin, The Public Papers of the Presidents and United Nations General Assembly and Security Council debates. For a complete account of the procedures used for identifying objectives that are articulated, as well as the sources and categories used to code them, see Legault, Albert, Stein, Janice, Sigler, John and Steinberg, Blema, Comparative Analysis of Dyadic Inter-state Conflict: Coding Actor Objectives, CADIC Technical Report No. 2 (Centre québécois de rélations internationales, June 1974)Google Scholar, available on request from the authors.

12 It is the responsibility of the case analyst to determine the nature of the impact of objectives on the dyadic opponent.

13 Incompatible objectives are those deemed to have a negative impact on the dyadic opponent; compatible objectives are assumed to have a positive impact, and mixed objectives are those deemed to contain both positive and negative elements. The terms incompatible and conflict, and compatible and cooperative are used interchangeably in this article.

14 Increases and decreases in incompatible objectives and coercive behaviour are measured in the same way. An increase is coded if there is a rise from one week to the next or if there is a continuation at a level greater than zero. A decrease is coded if there is a drop from one week to the next or if there is a continuation of zero cell entries.

15 For coding purposes, each objective is first categorized in terms of one of the following functions: protection, restoration, acquisition or denial. Objectives are then re-classified into a challenging and defending dichotomy.

16 CADIC defines a challenging state as one whose ratio of challenging to defending objectives is greater than 3:1. A ratio of less than 3:1 but more than 1:1 indicates an intermediate position of “challenger-defender.”

17 A defending state is defined as one whose ratio of defending to challenging objectives is greater than 3:1. A ratio of less than 3:1 is but greater than 1:1 indicates an intermediate position of “defender-challenger.”

18 The CADIC case study of the Egyptian-Israeli Conflict of 1973 reveals that the challenger, Egypt, articulates almost no defending objectives. Thus, the opportunity for bargaining and trade-offs was significantly reduced.

19 Soviet challenges are concentrated in weeks 1 (April 23–29), 7 (June 4–10), 12 (July 9–15), 13 (July 16–22), and 18 (August 20–26); only in week 7 is there no mention of the Berlin-Germany issue.

20 These are weeks 4 (May 14–20), 7 (June 4–10), 21 (September 10–16), and 26 (October 15–21).

21 Weeks 21 (September 10–16) and 26 (October 15–21).

22 Many of these Soviet goals do not appear in the profile of incompatible objectives since they replicate similar statements of American objectives. They are therefore coded as “compatible” or “mixed.”

23 Soviet perceptions of American objectives were drawn from the speeches of senior Soviet military and political decision makers.

24 See for example, Kahan, Jerome H., Security in the Nuclear Age: Developing U.S. Strategic Arms Policy (Washington: The Brookings Institute, 1975), 109–11Google Scholar; Allison, Essence of Decision, 52–54; and Horelick, Arnold L. and Rush, Myron, Strategic Power and Soviet Foreign Policy (Chicago: University of Chicago Press, 1965), 127Google Scholar.

25 See Horelick and Rush, Strategic Power and Soviet Foreign Policy, 136–38.

26 The Washington Post, December 18, 1962Google Scholar.

27 ”The main thing was that the installation of our missiles in Cuba would, I thought, restrain the United States from precipitous military action against Castro's government.” Khrushchev, Nikita, Krushchev Remembers, trans, and ed. by Talbott, Strobe (Boston: Little, Brown, 1970), 494Google Scholar.

28 For this analysis see Dinerstein, Herbert S., The Making of a Missile Crisis: October 1962 (Baltimore: Johns Hopkins University Press, 1976), esp. 232–36Google Scholar.

29 See for example, Leslie Dewart, “The Cuban Crisis Revisited,” 19; and Chase, Stuart, “Two Worlds,” Bulletin of the Atomic Scientists 19 (1963), 20CrossRefGoogle Scholar.

30 See for example, Horelick and Rush, Strategic Power and Soviet Foreign Policy, 127–36: Allison, Essence of Decision, 47–56; Hilsman, Roger, To Move a Nation (New York: Doubleday, 1967), 161–65Google Scholar; Ulam, Adam, Expansion and Coexistence: The History of Soviet Foreign Policy, 1917–1967 (New York: Praeger, 1968), 201–02Google Scholar; and George, Alexander and Smoke, Richard, Deterrence in American Foreign Policy: Theory and Practice (New York: Columbia University Press, 1964), 462Google Scholar.

31 All Soviet and American foreign policy events as reported in The Current Digest of the Soviet Press, The New Times of Moscow and The New York Times, were extracted. The data were then cleaned to eliminate duplicate reporting of events. For a complete account of the procedures used to identify actor behaviour reported in the source materials as well as the categories and techniques used to code that behaviour, see Legault, Albert, Stein, Janice, Sigler, John, and Steinberg, Blema, Comparative Analysis of Dyadic Inter-State Conflict: Coding Actor Behaviour, CADIC Technical Report No. 3 (Centre québécois rélations internationales June 1975)Google Scholar, available upon request from the authors.

32 The Soviet Union commits 154 conflict actions; the United States commits 58.

33 Punishments are defined as those acts involving punitive measures by an actor against a target; threats are defined as actions that describe negative sanctions against the target; acts of interference are defined as actions deliberately designed to be punitive, but without involving the direct use of force.

34 Acts of mobilization are those acts of unusual governmental nature that are directed toward the mobilization of domestic resources and/or the initiation of sudden change in the environment.

35 Charles A. McClelland suggests that marked increases in a state's event flow above its average score provides an important warning of approaching crisis (“Warnings in the International Events Flow: EFI and Roz as Threat Indicators,” International Interactions 5, Nos. 2 and 3 [1978], 135–204)Google Scholar.

36 It must be emphasized that objectives were extracted from verbatim speeches; behaviour was extracted only from newspaper reports of foreign policy events. Great care was taken to employ different materials to collect data on objectives and behaviour in order to avoid overlap and the danger of inferring objectives from behaviour.

37 Increases or decreases in challenging and defending objectives, in acts of punishment, interference, threats and mobilizations for each actor were measured from one week to the next (an increase was coded as +, a decrease as -). To measure the changes in objectives and behaviour we used a delta measure of first differences. Each variable was then correlated with subsequent increases or decreases in the aggregate conflict behaviour of each actor. All relationships were explored using one- and two-week time lags—using one-week lags, we have an n of 25; using two-week lags we have an n of 24, based on a 26 week pre-crisis period.

Four models were tested with and without a control variable. These models were: The Open Model: A → B, B → A, A → A → B, and B → B → A; The Closed Model: B → B, A → A, B → B → B, A → A → A; The Modified Open Model: A → B → B, and B → A → A; and the Modified Closed Model: A → B → A, and B → A → B.

When bi-variate relationships were tested using a one-week time lag, T → T → T+1, changes in the control variable were measured at the same time as the independent variable; when two-week time lags were used T → T+1 → T+2, changes in the control variable were measured in the intervening time period.

Those relationships that were statistically significant, using a Fisher's Exact Test, at .05 level or better, were then tested for their strength of association using a Somer's d measure. G. David Garson recommends the use of Somer's d as a measure of association for ordinal data based on probabilistic measures when the research hypothesis is directional (that is, asymmetric) (Handbook of Political Science Methods [Boston: Holbrook Press, 1971], 154)Google Scholar. See also Nie, Norman, et. al., Statistical Package for the Social Sciences (2nd ed.; New York: McGraw Hill Book Co., 1975), 229Google Scholar.

38 Recall that a challenger has been defined as a state whose ratio of challenging to defending objectives is greater than 3:1, while a defender exhibits a ratio of 3:1 or more defending to challenging objectives. Ratios of less than 3:1 but more than 1:1 indicate an intermediate position of either “challenger-defender” (when challenging goals are more numerous) or “defender-challenger” (when defending goals predominate).

39 To strengthen the validity of our results we tested the reverse relationships as well for both the Soviet Union (designated as A) and the United States (designated as B).

40 Soviet conflict behaviour increased in 16 cases out of a possible 25. This modified closed model explains 4 of the 16 cases. The results for cases when B's challenging objectives are increasing were not statistically significant and therefore are not reported.

41 This closed model explains 8 of the 16 cases in which Soviet conflict behaviour escalated. The results for cases when A's challenging objectives are decreasing were not statistically significant and therefore are not reported.

42 This open model explains 7 out of 16 cases of Soviet conflict escalation. The results for cases when B's coercive actions are increasing were not statistically significant and are not reported. Using a one-week time lag, this model and model 2A explain 11 out of 16 cases of Soviet conflict escalation. When weekly patterns were explored, only one case of overlap between the two models was observed.

43 This closed model explains 6 out of 16 cases of Soviet conflict escalation. The results for cases when B's conflict acts are increasing were not statistically significant and are not reported. Using a two-week time lag, this model and model 2B explain 14 out of 16 cases of Soviet conflict escalation. When weekly patterns were explored, only one case of overlap between the two models was observed.

44 American conflict behaviour increased in 16 cases out of a possible 25. This open model explains 10 of the 16 cases.

45 This modified-open model explains 6 out of 16 cases of American conflict escalation. Results for cases when B's challenging objectives are incireasing were not statistically significant and are not reported.

46 This modified-open model explains 6 out of 16 cases of American conflict escalation. Results for cases when B's coercive actions are increasing were not statistically significant and are not reported.

47 This modified-closed model explains 5 out of 16 cases of American conflict escalation. Results for cases when A's conflict behaviour is increasing were not statistically significant and are not reported.

48 There is only one case of overlap in the cases explained by any two of the three models.

49 This open model explains 9 out of 16 cases of American conflict escalation. The results for cases when A's defending objectives are increasing were not statistically significant and are not reported.

50 Among those analysts associated with an orthodox approach are Arnold Horelick, Myron Rush, Adam Ulam, Graham Allison, Elie Abel, Roger Hilsman, and Raymond Garthoff.

51 For example, Jerome Kahan and Anne Long argue: “The Kennedy administration's early emphasis on superiority can be said to have helped cause the Cuban Crisis by tilting the nuclear balance so far against the Soviets that they were ‘forced’ to emplace missiles in Cuba in order to rectify the strategic relationship. Had the United States been more sensitive to the Soviet need—both political and military—for equality, it might not have pressed its advantage as far as it did and, consequently, might have avoided the risks of the Cuban confrontation” (“The Cuban Missile Crisis: A Study of Its Strategic Content,” Political Science Quarterly 87 [1974], 490)Google Scholar.

52 See for example the works of I. F. Stone, Ronald Steel, and Leslie Dewart previously cited and Bernstein, Barton J., “The Cuban Missile Crisis,” in Miller, Lynn and Pruesson, Ronald (eds.), Reflections on the Cold War (Philadelphia: Temple University Press, 1974), 108–37Google Scholar.

53 Glenn Snyder and Paul Diesing argue that preference structures are a principal determinant of the development and resolution of crises. However, as they admit in their analysis of the Cuban Missile Crisis, preference structures are deduced, after the fact, from outcomes (Conflict Among Nations: Bargaining, Decision Making, and System Structure in International Crises [Princeton: Princeton University Press, 1977]Google Scholar, esp. 116). By focussing on objectives, the present study provides important empirical evidence for the content of Soviet and American “preferences.”

54 According to the United States Senate report, major American efforts to overthrow the Castro regime were effectively abandoned after the Cuban Missile Crisis (Alleged Assassination Plots Involving Foreign Leaders: An Interim Report of the Select Committee to Study Governmental Operations With Respect to Intelligence Activities, [November, 20, 1975], 139–73Google Scholar).

55 In a more general context, Robert Jervis notes that” states often place different values on different objectives, thereby facilitating accords that grant to each state what it values most” (“Deterrence Theory Revisited,” World Politics 31 [1979], 318)Google Scholar.