University of Southern California Dissertations and Theses

Prediction And Forecasting In International Relations: An Exploratory Analysis

(USC Thesis Other)

Prediction And Forecasting In International Relations: An Exploratory Analysis

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 70-25,078 YOUNG, Robert Arthur, 194-2- PREDICTION AND FORECASTING IN INTERNATIONAL RELATIONS: AN EXPLORATORY ANALYSIS. University of Southern California, Ph.D., 1970 Political Science, international law and relations University Microfilms, A X E R O X Company, Ann Arbor, Michigan THIS' DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED PREDICTION AND FORECASTING IN INTERNATIONAL RELATIONS: AN EXPLORATORY ANALYSIS by Robert Arthur Young A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (International Relations) June 1970 UNIVERSITY O F SO U TH ER N CA LIFO RN IA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES, CALIFO RNIA 9 0 0 0 7 This dissertation, written by Robert Arthur Young under the direction of h . D i s s e r t a t i o n Com mittee, and approved by all its members, has been presented to and accepted by The Gradu ate School, in partial fulfillment of require ments of the degree of D O C T O R OF P H I L O S O P H Y Dean ACKNOWLEDGEMENTS As is the case in most research projects of the kind reported upon in this paper, I am indebted to many people. The personnel at the University of Southern California's Computer Sciences Laboratory, particularly Hector Marin and Greg Carlton, who consistently went out of their way to expedite my numerous computer jobs. Shirley Singletary and Vivian Schwartz both helped com plete many coding, punching, data collecting and typing jobs for which I am very grateful. I am indebted to James Moore for his willingness to discuss and debate varying proposals, and not a few of his contributions are embodied in this research effort. Wayne Martin has been throughout our parallel graduate careers a deeply respected associate and friend. My debt is especially great to Professor Charles A. McClelland, who with patience and skill has guided me through my entire academic career. My thanks also to the other members of my committee, Professors Robert Berkov and Aurelius Morgner. To my wife, Nancy, it is impossible to express ii my appreciation satisfactorily. I can only say, thank you, realizing that as always, she will know and under stand. I, of course, must retain full responsibility for the results of the study. iii TABLE OF CONTENTS Page ACKNOWLEDGEMENTS..................................... ii LIST OF TABLES ..................................vi LIST OF FIGURES............... xi Chapter I. INTRODUCTION ............................. 1 Epistemological Considerations The Nature of International Phenomena Prediction and Forecasting in International Relations Normative Considerations Summary II. RESEARCH STRATEGY................ 39 III. AN INTERNATIONAL BEHAVIOR TAXONOMY: I . 6 1 Theoretical Considerations Data Source and Preliminary Variable Selection Statistical Techniques Preliminary Results IV. AN INTERNATIONAL BEHAVIOR TAXONOMY: II . 117 An Alternative Approach A Final Taxonomy: Results and Some Interpretations Other Considerations Summary and Conclusions iv Chapter Page V. AN INTERNATIONAL SITUATION TAXONOMY . . . 171 Theoretical Considerations Data Collection, Coding and Derivation Procedures A Preliminary Situational Taxonomy The Situation Data VI. FINAL ANALYSES, SUMMARY, AND CONCLUSIONS . 222 The Nation Behavior Taxonomy The Situational Taxonomy A Preliminary Event/Interaction Forecasting Paradigm Final Summary and Conclusions BIBLIOGRAPHY ...................................... 300 APPENDICES......................................... 312 A. WEIS Alphabetic and Numeric Actor C o d e s ............. 313 B. Raw and Ranked Data on Six Indicator Variables.............................320 C. Examples of Situation Coding . . . 337 D. Situation Data Across WEIS Combevents All Actors, USA and USR 1955, 1967, and 1955-1967 Combined ............. 356 v LIST OF TABLES Table Page 1. Rank Order of All WEIS Actors by Number of Acts Output— 1966-1969 ............. 73 i 2. Rank Order of 83 Nations by Number of Acts Output— 1966-1969 ................ 74 3. WEIS Event/Interaction Classification System............................... 76 4. 22 WEIS Combevent Categories for 83 Selected Actors— 1966-1969 . . . . 79 5. Factor Analysis of 22 Combevent Categories for 83 Selected Actors— 1966-1969 . . 98 6. Hierarchical Grouping Analysis of 22 Combevent Categories for 83 Selected Actors— 1966-1969 100 7. Q-Factor Analysis of 83 Nations for Four Sets of 22 Combevents, Year by Year— 1966-1969 . . . . . . . . . . 105 8. Hierarchical Grouping Analysis of 83 Nations for Four Sets of 22 Combevents Year-By-Year— 1966-1969 108 9. Nation Grouping (Preliminary) . . . . Ill 10. Q-Factor Analysis of 83 Nations for Six Indicator Variables by Three-Month Periods— 1966-1969 125 vi Table Page 11. Hierarchical Grouping Analysis of Six Indicator Variables by Three-Month Periods— 1966-1969 ...................... 127 12. Combined Grouping of 83 Nations— 7 Groups ............................... 128 13. Mean Group Ranks by Indicator Variable . 134 14. Group Rank Standard Deviations by Indicator Variable ............................... 145 15. Comparison of Young Behavioral and Russett Sociocultural Taxonomies ................ 154 16. Comparison of Young Behavioral and Banks and Gregg Political System Taxonomies 155 17. Comparison of Young Behavioral and Russett UN Voting Behavior Taxonomies . . . . 156 18. A Preliminary Situational Taxonomy 195 19. Total Acts Output by Major Situation Category ............................... 208 20. Total Number of Recorded Acts in Situation Data Collection . . . ................ 212 21. Total Acts Out of Seven Most-Active Nations by Situations ......................... 215 22. Verbal and Physical Conflict by Nation— 1966-1969 ............................ 225 23. Verbal and Physical Conflict by International Behavior Group— 1966-1969 ............. 228 24. Correlation of Verbal and Physical Conflict- 1966-1969 ............................ vii 231 Table Page 25. All Nations-Verbal and Physical Conflict by Month— 1966-1969 ................ 235 26. Type 1-Type 6 Nations-Physical and Verbal Conflict by Month— 1966-1969 .... 237 27. Correlations of Verbal Conflict, Physical Conflict, and Time— 1966-1969 . . . 239 28. Situation Correlation Coefficients— All Nations— 1955 and 1967 ............ 252 29 . Situation Correlation Coefficients— All Nations— 1955 ................ 255 30. Situation Correlation Coefficients— All Nations— 1967 ......................... 256 31. Situation Correlation Coefficients— United States— 1955 and 1967 . . . . 259 32. Situation Correlation Coefficients— United States— 1955 . .... . . 260 33. Situation Correlation Coefficients— United States— 1967 ................... 261 34. Situation Correlation Coefficients— Soviet Union— 1955 and 1967 .... 265 35. Situation Correlation Coefficients— Soviet Union— 1955 266 36. Situation Correlation Coefficients— Soviet Union— 1967 267 37. Group 1 Nations* Output by Combevent for Selected Situations— 1966-1969 . . . 276 viii Table Page 38. Group 1 Nations' Output by Combevent for Selected Situations in Percentages— 1966-1969 ............................... 278 39. Numerical Listing of WEIS Actors .... 314 40. Alphabetical Listing of WEIS Actors . . . 317 41. Raw Data on Six Indicator Variables 83 Selected Nations— 1966-1969 . . . . . 321 42. Ranked Data on Six Indicator Variables 83 Selected Nations— 1966-1969 ............. 324 43. Raw Data on Six Indicator Variables Group 1 Nations— 1966-1969 (Aggregated) . . . 327 44. Ranked Data on Six Indicator Variables Group 1 Nations— 1966-1969 (Aggregated) . 328 45. Raw Data on Six Indicator Variables Group 2 Nations— 1966-1969 (Aggregated) . 329 46. Ranked Data on Six Indicator Variables Group 2 Nations— 1966-1969 (Aggregated) . 330 47. Raw Data on Six Indicator Variables Group 3 Nations— 1966-1969 (Aggregated) . 331 48. Ranked Data on Six Indicator Variables Group 3 Nations— 1966-1969 (Aggregated) . 331 49. Raw Data on Six Indicator Variables Group 4 Nations— 1966-1969 (Aggregated) . 332 50. Ranked Data on Six Indicator Variables Group 4 Nations— 1966-1969 (Aggregated) . 333 51. Raw Data on Six Indicator Variables Group 5 Nations— 1966-1969 (Aggregated) . 334 ix Table Page 52. Ranked Data on Six Indicator Variables Group 5 Nations— 1966-1969 (Aggregated) . 334 53. Raw Data on Six Indicator Variables Group 6 Nations— 1966-1969 (Aggregated) . 335 54. Ranked Data on Six Indicator Variables Group 6 Nations— 1966-1969 (Aggregated) . 335 55. Raw Data on Six Indicator Variables Group 7 Nations— 1966-1969 (Aggregated) . 336 56. Ranked Data on Six Indicator Variables Group 7 Nations— 1966-1969 (Aggregated) . 336 57. All Actors— 1955 Situation Data Across WEIS Combevents............................357 58. All Actors— 1967 Situation Data Across WEIS Combevents............................358 59. All Actors— 1955 and 1967 Situation Data Across WEIS Combevents ................ 359 60. USA— 1955 Situation Data Across WEIS Combevents...............................360 61. USA— ”.967 Situation Data Across WEIS Combevents . . . . . 361 62. USA— 1955 and 1967 Situation Data Across WEIS Combevents............................362 63. USR— 1955 Situation Data Across WEIS Combevents...............................363 64. USR— 1967 Situation Data Across WEIS Combevents...............................364 65. USR— 1955 and 1967 Situation Data Across WEIS Combevents............................365 x LIST OF FIGURES Figure Page 1. A Factor Analysis........................87 2. A Q-Factor Analysis Matrix.............. 89 3. Matrix Configuration for Preliminary Q-Factor Analysis ...................... 103 4. Data Matrix for Six Indicator Variables on 83 Nations for Sixteen Three-Month Time Periods, 1966-1969 ............... ... . 123 5. Mean Group Ranks by Indicator Variable . . 135 6. The Form of a Possible Relationship Between Actor Type and Situation..............271 7. The Relationship Between Actor Type, Situation and Behavior............ ... 273 xi CHAPTER I INTRODUCTION In any investigation of prediction and forecasting in the social sciences, one must inevitably examine some very basic theoretical and philosophical questions.'1 ' In this introductory chapter, we shall be concerned first with some central epistemological issues, including the relationship between explanation and prediction and definitions of key terms of reference. We then shall review the opinions of several scholars regarding the basic nature of international phenomena, after which the conclusions reached will be examined within the context of the theoretical and practical problems encountered in prediction and forecasting. Finally, some of the normative aspects of the subject will be reviewed in an effort to determine whether or not one ought to forecast in inter national relations. ■^The terms "prediction," "projection," and "fore casting," will be defined and discussed at length below. 1 2 Epistemological Considerations The logical and practical relationship between prediction and explanation is of central importance to any study of prediction and forecasting. Simply stated, the question is whether it is possible to predict the occurrence of a given phenomenon without explaining fully why it will occur, or whether it is necessary to explain before one may predict. The position of philosophers of science on the matter usually depends upon their convic tion with regard to the more general question of the 2 purpose of science and knowledge. Some assert that the principal aim. of science is to produce predictive theory. Many of those who subscribe to this view do so because they believe that, on logical grounds, explanation is of little or no value unless it permits one to predict. Carl Hempel and Paul Oppenheim, prominent advocates of this point of view, argue that 2 This conclusion, and several others discussed below were drawn from Raymond Tanter's article on "Explanation, Prediction and Forecasting in International Relations" (1969). 3 . . . it is this potential predictive force which gives scientific explanation its importance: only to the extent that we are able to explain empirical facts can we attain the major objective of scientific research, namely not merely to record the phenomena of our experience, but to learn from them, by basing upon them theoretical generalizations which enable us to anticipate new occurrences and to control, at least to some extent, the changes in our environment.^ Other scholars, who also argue on logical grounds, reject the notion that a predictive capability is an essential attribute of an acceptable explanation, assert ing instead that the central aim of science is to explain the occurrence of observed phenomena, not to predict them. Thus, Israel Scheffler believes that unlike our concern with control in practical affairs, our concern in science is not directed particularly toward the future, but frequently involves employment of general principles in the effort to substantiate past events on the basis of later events . . .4 Since scholars who have devoted much of their academic lives to examining this question on purely logical grounds have been unable to reach agreement, we are not so 3 Hempel and Oppenheim, 1953: 323, as quoted m Tanter, 1969. See Lundberg, 1963: 66 for further support of. this position. ^Scheffler, 1963: 53, as quoted in Tanter, 1969. 4 arrogant as to suppose we might do so here. Fortunately, it is not necessary. It is possible instead to recognize the dichotomy, referred to by Scheffler, between conclu sions reached on the basis of practical arguments and those based upon purely abstract logic. Thus, Harold and Margaret Sprout, in an early attempt to examine the rela tionship between prediction and explanation in the field of international relations, avoided agonizing over the logical relationship between these two goals of inquiry. Rather, they recognized that "it is frequently stated that the logical principles involved in explanation of histori cal events and more complex states-of-affairs are the same as those involved in prediction," but nevertheless 11. . . there are certain practical differences" (Sprout and Sprout, 1961: 61). Abraham Kaplan also argues persuasively that while abstract philosophy may require an ideal explanation to have predictive power and an ideal prediction to con tain the elements necessary to explain why the specified event will occur, it is unreasonable to expect all science, particularly behavioral science, to conform rigidly to such requirements. He advances two reasons for this con- 5 elusion. First, he points out that "explanations provide understanding, but we can predict without being able to understand, and we can understand without necessarily being able to predict" (Kaplan, 1964: 350). Kaplan reasons that predictions can be and often are made even though we cannot explain what is being predicted. Con versely, many explanations of observed phenomena exist which do not permit prediction. The former case . . . is characteristic of well established empirical generalizations that have not yet been transformed into theoretical laws. Ancient astronomers made predictions of quality incompar ably better than their explaining theories; even today the periods of certain pulsating stars are very well known even though we do not have any firm explanations of their behavior. Similarly, the course of a mental illness or the outcome of an election campaign might be predictable with far better grounds than are available for any explanation of these phenomena. (Kaplan, 1964: 349) Kaplan admits that a complete explanation probably will nearly always permit prediction, but precious few explana tions, especially in the behavioral sciences, are as inclusive as one might wish: A certain neurosis might be explained as the result of a childhood trauma, but not all traumas produce a neurosis in later life; the neurosis is not pre dictable on the basis only of what serves to explain it. . . .An automobile accident is said to be due to defective brakes or to excessive speed or to a 6 drunken driver? none of these things allow us to predict the accident, but in particular cases they may have considerable explanatory force. . . . (Kaplan, 1964: 348) Kaplan's second reason for denying the necessity of requiring that theoretical statements have both explanatory and predictive capabilities rests on the refutation of an often-made comparison between the behav ioral and physical sciences: It is not of much help in the conduct of inquiry, especially in behavioral science, to hold up the ideal of explanations having the predictive power of celestial mechanics. Why should behavioral science make demands on itself that are so seldom imposed in physics or biology? (Kaplan, 1964: 348-349) Other scholars also have noted this unreasonable compari son (e.g., Machlup, 1963: 172-174, and Nagel, 1963: 208). Raymond Tanter, for‘example, feels that . . . perhaps the comparison of physical events in a laboratory setting of tight experimental controls should not be with the social worldt of foreign policy decision-makers. Why not compare the physi cal world of earthquakes, mountain slides and epidemics to the social world of war, revolution and political integration? Natural scientists do not come off so well predicting to physical events in the world as they do in laboratory environments. Even where considerable control is present as in the aerospace industry, there is often a period where there is an inability to predict performance of new machines. Why should foreign policy scholars be expected to forecast the social future perfectly when natural scientists do not always succeed? (Tanter. 1969: 8)______________________________________. 7 In sum, our position on this issue is that while there may be, in the abstract and on the basis of pure logic, a requirement that one be able to explain why a given event will occur if he predicts it will happen, we do not believe that this is either a necessary or desir able requirement. It is not necessary since, as discussed above, perfectly valid and useful predictions can be made without the concurrent ability to explain them, and it is undesirable because adherence to it effectively would prevent significant progress from being made in the development of new theories and methodologies essential to the improvement of our ability to understand and better control our international environment. The realistic separation of prediction and explanation becomes even clearer when considered in the context of the relationship between prediction, projec tion and forecasting. A forecast is a description of an event or trend which is expected to happen at some identified future time (Tanter, 1969: 10). It can be a result of intuition, prophecy, statistical analysis or 8 C any other method of gaining insight. In science, however, a forecast is a result of the projection of at least one element of a predictive hypothesis. A projec tion provides an estimate of the value of a variable at some future time, the estimate being based upon the observed change in that variable during some past time period.6 A prediction does not necessarily refer to the future at all, despite popular usage to the contrary. It involves only the form of the relationship between two or more variables, such as if variable "A" is present, variable "B" will be absent. An illustration should help to clarify these relationships. Suppose a student is sitting in a class room and, becoming bored, begins to gaze out the window. He observes that while it was bright and sunny when he entered the room, in the space of only a few minutes rather ominous-looking clouds have begun to form. He knows from past experience that if a lot of those black. 5 For a more complete discussion, see Meehan, 1968: 21-23. 6For an excellent presentation of their relation ships, including examples, see Rubin et al., 1967: 53-55. 9 heavy looking clouds (variable "A") get together to cover the sky, it will rain (variable "B") (his predictive hypothesis). He notices that they are continuing to form at the same rate they have been, and decides that enough will gather (his projection) for it to rain and on the way to his next class, to get him wet (his forecast). Notice that the student's inability to explain in precise, scientific terms why it is going to rain neither prevents him from making his forecast nor precludes it from being an accurate one. Similarly, it might be possible to discover a predictive relationship which permits the relatively accurate forecasting of impending international crises, perhaps simply on the basis of significant increases in the interaction between two nations as com- pared with their previous performances. The increases I in activity do not in themselves cause the crisis and at best they could only partially explain why it will take place but, given an established predictive relationship, the resulting forecast likely could be both accurate and 7 For examples of attempts to operationalize this notion, see McClelland etal., 1965, and Hoggard, 1969. 10 useful. A final distinction to be made in this section is among different types of forecasts. Saul Friedlander, in his extensive examination of forecasting in inter national relations, asserts there is a fundamental differ ence between short- and long-term forecasts. He defines each as follows: . . . a short-term forecast in international rela tions [is] any forecast which disregards the influ ence of the independent processes on the changing pattern of power relations between the States under observation. Forecasts of this kind can only be applied to a very short period of about two years at the most. During this period, the power rela tion between the States concerned, which is governed by the action of the independent processes, is regarded as being constant, and any forecast is entirely concentrated on the probable decisions of the actors and the interaction of those deci sions. . . . (Friedlander, 1965: 9-10) Long-term forecasting, which is concerned with periods from ten to fifteen years approximately, must primarily take account of changes in the power relation between the States under consideration arising from independent processes which determine the power quanta of these States. . . . (Friedlander, 1965: 11) As one can see, Friedlander believes that there are two important classes of variables to be considered in fore casting. The first, which is the "decisions of the actors and the interaction of those decisions" is the set of 11 information that is to be analyzed in short-term fore casts. The second set consists of the "processes which operate independently of the actors,1 1 by which Friedlander means those basic "demographic, social and economic processes" such as Gross National Product, Population, Natural Resources, etc. These are the major factors to be considered in long-range forecasts. We are not in total agreement with Friedlander on his definitions, but we believe his distinction between short- and long-term forecasting essentially is correct. The interaction of nations does seem to us to be the major consideration in short-term forecasts, although we would extend the range to three or possibly even four years out. Similarly, we would not quarrel with Friedlander's defi nition of long-range forecasts as being concerned with periods of from ten to fifteen years in the future, although in this case we can see no reason for an upper bound. Further, he probably is narrower than is desirable in his definition of what constitutes "long-term proc esses." We also found it curious that no mention was made of the range between two and ten years, to which we shall refer as the mid-range. Very little research has 12 been done in this area, perhaps because it may be the most difficult of all since it probably would require consideration of both sets of variables discussed above. To this point, we have examined three related epistemological considerations. The first was the rela tionship between explanation and prediction. There it was concluded that while there may be an abstract logical requirement that an acceptable explanation must also permit prediction, and vice-versa, on more practical and realistic grounds it probably is possible to predict without the necessity for explaining one’s prediction. Next, we distinguished between prediction, projection, and forecasting and used an example to illustrate these distinctions. Finally, definitions of short-, mid-, and long-term forecasting were provided. Before proceeding to an examination of the theoretical and practical possi bilities of prediction and forecasting in the field of international relations, it is necessary to consider the basic nature of international phenomena. 13 The Nature of International Phenomena The basic nature of international phenomena is a subject that is central to any study of prediction and forecasting. The controversy focuses principally on the question of whether or not a pattern exists in the inter actions between nations or whether each event must be considered as absolutely unique. Two extreme positions on the question can be cited. On the one hand, some historians contend that history (and therefore inter national relations) is a succession of unique events, none of which is exactly like any other. Any pattern perceived is imposed by the analyst. Thus, the phenomena are considered essentially random. H. A. L. Fisher, for example, asserts that Men wiser and more learned than I have discerned in history a plot, a rhythm, a predetermined pattern. These harmonies are concealed from me. I can see only one emergency following upon another as wave follows upon wave, only one great fact with respect to which, since it is unique, there can be no generalizations, only one safe_ rule for the historian: that he should recognize in the development of human destinies the play of the contingent and unforseen. (Fisher, 1939: xv) The other extreme, advocated by Marxists and Calvinists, among others, holds that all events are pre 14 determined and that it is consequently impossible for anyone or anything to affect the inexorable course of history. In a detailed examination and review of Marxist theory, C. Wright Mills concludes that Marx very definite ly believed that although men make their own history, given the cir cumstances of the economic foundation, the way they make it and the direction it takes are deter mined. The course of history is structurally limited to the point of being inevitable. (Mills, 1962: 90) Similarly, Calvin, in the Institutes of the Christian Religion, states that we hold that God is the dispenser and ruler of all things, that from the remotest eternity, accord ing to his own wisdom, he decreed what he was to do, and.now by his own power executes what he decreed. Hence we maintain that by his providence, not heaven and earth and inanimate creatures only, but also the counsels and wills of men are so governed as to move exactly in the course he has destined.8 History, according to both these latter schools of thought, is perfectly deterministic. 8 Calvin, as quoted in Dakm, 1946: 23. Dakin concludes that "... his (Calvin's) teaching concerning providence would seem to lead to a determinism as rigid as any in philosophy. ..." (Dakin, 1946: 83-84) 15 It is of course impossible for these quite opposite contentions both to be true. Fortunately, theoreticians in the field of international relations have devoted considerable attention to precisely this question of the nature of the phenomena they are studying. As might be expected, the vast majority of authorities in the field subscribe to neither of the extremes noted. Rather, most agree with Kenneth Thompson's conceptualiza tion : The substance of theory is history, composed of unique events and occurrences. An episode in history and politics is in one sense never repeated. It happens as it does only once; it is nonrecurrent in that it has never happened before nor will it be repeated again. In this sense, history is beyond the reach of theory. Underlying all theory, however, is the assump tion that these same unique events are also con crete instances of more general propositions. The wholly unique, having nothing in common with anything else, is indescribable, for we know that all analysis (and even description) is made in terms of predicates, class concepts, and repeat- able relations. Both recurrence and uniqueness are present in history as in everything else, and the logical difficulty in reconciling them is neither greater nor smaller in international relations than in other fields. . . . The ordering of data can help the observer to distinguish uniformities and uniquenesses. . . . Theory holds out the tools whereby the observer can discover in the welter of events that which is recurrent and typical . . . [and] 16 . . . can be an instrument for understanding not only uniformities and regularities but contin gencies and irrationalities as well.9 This conclusion that history is composed of both "uniformities and uniquenesses" is of very great importance since if it were not possible to set forth general propositions, prediction and forecasting would by definition be impossible. Before considering the implications of this conceptual view, however, it will be useful in our later discussions to clarify with the presentation of examples precisely what is meant by theorists when they refer to "patterns" and "uniquenesses.1 1 The many types of pattern that exist in the system usually are labeled processes, structures, laws, and so on. It is almost universally accepted, for example, that many nations have gone to great lengths to formulate general foreign policy objectives and then to act quite consistently along those lines, in some cases for many years. Note that such observed behavior patterns often are presented in the form of a predictive hypothesis. Great Britain, for example, long conceived of herself as the "balancer" in a balance of power system. Organski observes that "the nation that is believed to have filled 17 this difficult role for the longest period and in the most successful manner in England" (Organski, 1961: 278- 279) . The predictive hypothesis is found in the classic statement on the balance of power and the British role therein written by Sir Eyre Crowe in 1906: The equilibrium established by such a grouping of forces is technically known as the balance of power, and it has become almost an historical truism to identify England's secular policy with the maintenance of this balance by throwing her weight now in this scale and now in that, but ever on the side opposed to the political dictator ship of the strongest single state or group at a given time. (Crowe, 1928: 402) Russia is another example of a nation which has persistently pursued a basic foreign policy objective, in this case, that of expanding her territory whenever possible to the sea. Robert J. Kerner ends his book on the subject by stating that the conclusion to be drawn from this analysis extending over a thousand years of recorded history is, I believe, obvious. Here is a process— the elements of which are the people, rivers, and portages, the ostrogs, monasteries, and furs— which may be traced in its action and development from tribal community to world empire. . . . Over these rivers and portages went repre sentative members and groups of several types. . . . Each, whatever its ideology, utilized them. These changing types could come and go, but the elements of the process always remained. The urge to the sea always dominated. (Kerner, 1946: 103-04)_________________________________________ 18 The United States apparently has several basic foreign policy tenets which she has supposedly adhered to during the past 200 years.'*'® One of the best known is the Monroe Doctrine. Though variously interpreted through the years, it has been affirmed "in principle" many times, most recently in a State Department release which asserted in part that "the principles of the Monroe Doctrine are as valid today as they were in 1823 when the Doctrine 11 was proclaimed." Besides this kind of consistency, one could cite many examples of regularities observed in processes such as arms races (Richardson, 1960), international negotia tions (Lall, 1966), or international crises (McClelland, 19681• The structure of the international system is still another example of the way in which pattern or 10 See Bailey, 1964: 1-2, for a listing of six long-standing foreign policy objectives. ■^U.S. Department of State, 1960: 170-171, as citec in Dozer, 1965: 185-187. Dexter Perkins, in his book, A History of the Monroe Doctrine, 1955, notes in his fore word that the principles of the Monroe Doctrine, in the summer of 1940, "were approved by the virtually unanimous vote of both houses of Congress. They are at the heart of our creed with regard to foreign policy." 19 regularity is hypothesized and demonstrated. The terms "bipolar" and "multipolar" are now familiar labels used to describe different patterns observed in the system (Kaplan, 1957). Alliance configurations (Singer and Small, 1968) , trade patterns (Russett, 1966), and airline connec tions (Gleditsch, 1967) are all further examples of system structure. Since we have established that international relations are essentially composed of two mutually exclu sive phenomena— those which fit a definable pattern and those which do not— we can assert that, by definition, any event which is not part of an identifiable pattern is unique. And, conceptually, it will remain unique or idiosyncratic until and unless it can be fitted into a (perhaps newly discovered) pattern. Should the Soviet Union, for example, cease its "urge to the sea," we would have to say that this was a unique event, unique in the sense that it deviated from a pattern that had been estab lished over perhaps a thousand years. Similarly, Cuba's becoming a Communist state and President Kennedy's assassination are both probably classified by most analysts as idiosyncratic events. Natural disasters', such as 20 floods, hurricanes, earthquakes and the like, usually also are put into this category. We can conclude, then, that generally speaking any event that is unexpected is labeled "irrational" or "unique." The discussion relating to the nature of the phenomena is well summarized by Charles A. McClelland in . his delineation of "parts of subject theory which appear to have achieved general acceptance in the profession. . ." There are regularities and recurrences which are reflected in the concrete data of international relations. We make the assumption that these regularities and recurrences exist, that we can find them in the welter of the data, and that we can explain them by means and methods exposed to the public view. The epistemological requirement is that we make this assumption if we profess to understand. Faith in this presupposition is, indeed, the foundation of the social sciences. (McClelland, 1960: 305) We have found, then, that despite opinion on the nature of international phenomena which ranges across a wide spectrum, consensus within the discipline of inter national relations itself can be well established to the effect that regularities in the relations of nations do indeed exist. Several examples of such patterns, the existence of which is critical in any discussion of prediction, were presented along with some clarification 21 of what is meant by a unique or idiosyncratic event. The implications of this consensual view of international phenomena must now be examined with regard to the possi bility of prediction and forecasting in the field. Prediction and Forecasting in International Relations It is possible to go one step beyond the concep tualization of the basic nature of international phenomena and still maintain reasonably strong consensus in the field. Most theorists agree that analysts should attempt to do their research in as systematic a fashion as possi ble, testing hypotheses and propositions with empirical data whenever possible. While there is normative agree ment on how we theoretically ought to proceed, however, one cannot cite such consensus with regard to what prac tically can be accomplished. The controversy here centers not so much on the existence of patterns as on their tremendous complexity and on the feasibility of using the tools of science (e.g., categorization, objective measure ment, statistical description and inference, etc.) on such seemingly elusive data. Henry Wriston, for example, feels 22 that the factors involved in historical events are so numerous and so complex that any explanation, however detailed, is an oversimplification— often so great as to be seriously misleading. . . So variable, so complex are the human factors that in deeply significant political matters the theory of probability breaks down. (Wriston, 1962: 377) Earlier in the same essay, Wriston had asserted, Further, it is impossible, even with modern computers, to put down all the relevant facts, for they are in number as the sands of the seas" (Wriston, 1962: 376). This is, as he goes on to point out in detail in his essay, a reference to the use of statistics or other "scientific" techniques to analyze international or foreign relations. The author uses these two arguments to explicitly reject the possibility of making predictions or forecasts of such phenomena: "... dependence upon the persistence of trends in human affairs is based on a fallacy. . . . When applied to questions of national policy, the extrapo lation process breaks down. ..." (Wriston, 1962: 375-76). Many others, both practitioners and scholars, echo this line of argument.It is important to note, however, 12 See for example, Hoffman, 1960. 23 that very few assert that trends or patterns do not exist. Rather, they will not accept that they can be projected or forecast. In sum, the major arguments against predicting international relations are (1) the phenomena, while undoubtedly containing some pattern, are too complex to permit forecasts; and (2) quantitative or scientific techniques are not appropriate to the analysis of international relations. Proponents of prediction rest their case on a variety of grounds. First, they contend that the very number and complexity of international relations make it necessary to employ techniques of analysis which have been designed to overcome just this problem. It is not possible, they assert, for the human brain to handle the myriad factors all agree are present in international relations. Thus, because of the first assertion above (the phenomena are too complex), the second (the use of quantitative techniques) becomes almost a matter of necessity if one holds out any hope for analyses that will be sufficiently well defined to ever lead into the 24 predictive mode.13 Further, they point to the increasing number of quantitative analyses that have identified pattern, structure, and process in the relations of nations and produced significant new insights. Much of the opposition to prediction of inter national relations may stem from a basic misunderstanding of the nature and form of prediction and forecasting efforts. Charles McClelland has suggested that one of the major reasons for such skepticism is that the official who confronts the response of other decision-makers really wants and needs what are called "point predictions." He is quite right in suspecting that science cannot calculate just when, where, and in what way a particular event will occur. A scientific prediction is based on contingency and probability. It can deal effec tively only with class phenomena and in terms of the probability of some specific kind of occur rence, given the presence of some set of specified conditions. Nothing better can be hoped for? the decision-maker is concerned about a particu lar future effect or response and the only way a scientific system can help is by reducing the degree of the decision-maker's uncertainty about future outcomes. Some possibilities can be ruled out and others might be rated according to their likelihood. (McClelland, 1969: 2-3) 13 It is not being held that "traditional" analyses are of no use, but that the use of quantitative techniques, where possible, can help to overcome the immense complex ity of the data. 25 As Professor McClelland points out later in his essay, it is unfortunate that some of the claims causing practi tioners and scholars in the field to oppose efforts in this area have been put forth by members of the profession itself (McClelland, 1969: 5-7). Thus, it is hoped that a modification of expectations along lines suggested above by McClelland may cause some to reconsider their opposi tion to prediction and forecasting efforts. Skepticism about forecasting in the social sciences certainly is not new. During the first few decades of the twentieth century economists were far from in agreement on the usefulness and feasibility of forecasting economic phenomena. Consider, for example, three propositions put forth by Oskar Morgenstern in 1928: I. Forecasting in economics by the methods of economic theory and statistics is 'in princi ple' impossible. II. Even if it were possible to develop a technique of economic forecasting, such a tech nique would be incomplete, by virtue of its necessary limitation to methods based on a knowl edge of economics alone; it would therefore be incapable of application in actual situations. III. Moreover, such forecasts can serve no useful purpose. All attempts to develop a formal 26 technique for forecast are therefore to be dis couraged. 14 Karl Karsten, an early proponent of business forecasts, assessed the climate of opinion in 1931 in the following way: In an age when it is popular to disparage fore casting of all kinds, the height of folly is considered the idea of statistical or "mechanical" forecasts of economic conditions. (Karsten, 1931: 3) He goes on to say, however, that his book . . . is a defense of the particular type of fore casting least in vogue. It is a challenge in that it seeks to show not only that this particular type of forecasting is scientific in nature, but also that it will command the endorsement and approval of everyone who desires the aid of science in the momentous problems of forecasting business and economic conditions. "A little knowledge is a dangerous thing," and in its early stages a new science is prone both to misunderstanding and misuse. Before chemistry there was alchemy; before astronomy there was astrology. The science of forecasting, if it may already be dignified by such a term, has been no exception, and been preceded by prac tices in its name which have fully merited the skepticism of every sensible person. We have had our share of mistaken predictions . . . (but) . . . in the last analysis, the thought that a 14 Morgenstern, 1928. The three points are a summary of the book by Marget, 1929: 312-339. (Footnotes omitted.) .. 27 "system" may exist which is more accurate in its appraisals of the future than the best opinion of the most experienced and intelligent human brain, is unflattering to the brain and is unconsciously but strongly resented by many who feel that they possess such brains. With such persons there is no useful escape from a dilemma in which the fore casters less successful than themselves are ridicu lous and those more successful than themselves intolerable, or incredible. For all persons, it seems difficult to.grasp the simple fact that the mind of man can make an instrument more accurate than the mind itself.^ During the post-World War II period the rapid development of economic theory and methodology and the consequent relatively great success and general acceptance of busi ness and economic forecasts have made the discussion of 16 "can we or can't we predict" a more or less moot point. What we are suggesting here is that the field of international relations may today be in a position somewhat analogous to that of economics in the early part of this century, and to the developmental periods of other sciences as well. Indeed, the issues discussed in the 15 Karsten, 1931: 3-5. See also Haney, 1931, espe cially the preface and Chapter 2. •^See a recent article in the New York Times by Heinemann entitled "Economics: Taking the Guess Out of Prediction," October 28, 1969, p. 16. 28 economic literature of the twenties and thirties are remarkably similar to those argued in the journals and books of international relations in the sixties. In this section, we have attempted to examine and discuss the cases for and against international rela tions predictions and forecasts solely from the standpoint of the possibility of predicting such phenomena. The mixed nature of international phenomena (i.e., historical events) was established and the need to discover and isolate pattern asserted. The divergence of opinion with regard to the feasibility of systematically establishing predictive relationships, projecting and forecasting then was examined and the sources of disagreement discussed. Finally, the experience of another social science, economics, was cited in support of the thesis that predic tion and forecasting in international relations is theoretically and, at least at the academic level, tech nically possible. As much as possible, normative con siderations of the form of should we predict have not been touched upon. It is to this question that we now turn. 29 Normative Considerations There are two major normative arguments favoring systematic prediction and forecasting in international relations. The first asserts the . . . inadequacy of personal experience for evaluating and planning international policy. I mean this statement to apply to the personal experience of officials, nongovernmental experts, and citizens. 'Personal experience' is short hand for values, private models, ill-defined samples of facts, and encountered incidents which each of us carry in our memory. The limited resources which personal experience affords are sanctified not multiplied when policy is made by even large numbers of persons with the same cognitive stock and value profile. As a consequence of reliance upon such "personal experi ence" in formulating foreign policy, Kenneth Boulding believes "appalling" are committed (Boulding, 1963: 77) . The second argument, put forward recently by Charles McClelland, refers more to basic research within the academic environment; nothing much is accomplished, whether the subject is flow of the weather or the flow of international 17 Bobrow, 1969: 1. See also Boulding, 1963. 30 events, until actual prediction attempts are made, consistently, periodically, and persistently. It is the discipline of calculating futures from observed data that serves theory and basic re search in such large and difficult fields. (McClelland, 1969: 7) These two complementary arguments constitute the main positions taken by prediction proponents but at the same time are in themselves a source of some confusion. If, as is usually the case, only one is considered or they are considered separately, it is possible to advance arguments against prediction that otherwise might be somewhat untenable. Consider, for example, the decision maker who becomes aware of predictions made solely for the purpose of theory generation and/or validation. Let us label these "academic" predictions. Usually, his opinion of them will be negative since, by definition, they will not help to solve his immediate problem which is choosing among alternative policies. If he is willing, however, to view them as necessary steps along the long road to theory construction and validation that may, several years downstream, be brought to bear (directly or indirectly) on his problems instead of purely in the context of his own preoccupations, he perhaps will be less inclined to be critical. 31 By the same token, the decision-maker continually makes forecasts of the probable effects of alternative policies before deciding which to implement, although he may not do so explicitly. Whenever such estimates are 18 formalized (we will label them "practical1 1 predictions) and exposed to outside, especially academic, review, again they generally are disparaged as being "simple-minded," "misleading," and so on, because of necessity they are usually rough approximations that ignore the "scientific approach." What is needed, of course, is greater under standing and communication between the two groups which would be greatly facilitated by the mutual recognition that academic and practical prediction have, at least in the short run, very different goals and objectives. The major objection to explicitly working toward attaining the ability to forecast the effects of alter native foreign policies is cited by Davis Bobrow in an article in which he proposes that a series of indicators be established that would be designed to monitor the state of our relations with other nations in the international 18 After McClelland. See McClelland, 1969. 32 system. Such an indicator system, if implemented, undoubtedly would play a central role in future forecast ing and system control efforts. This objection would be to . . . the possibility that whatever gains an establishment makes in its ability to evaluate and plan contribute to control and manipulation of citizens, but not to progress toward generally desired goals. Knowledge is power; restricted knowledge is power for a few. (Bobrow, 1969: 3) Bobrow's reply is succinct and appears to effectively answer such criticisms: This possibility is real but hardly inevitable. Explicit indicator systems can make the evalua tion and planning of international policies far more public and participatory than has been the case. Series of readings which index progress toward and regression from normative states en able meaningful broadening of participation. Without such readings, those outside an inner circle can only evaluate elite views on the basis of generalized trust and distrust. (Bobrow, 1969: 3-4) Here, as in all normative discussions, it is difficult to conclude that one set of arguments is wrong and the other right. Nevertheless, we believe that the weight of the arguments presented favor the conclusion that prediction and forecasting in international relations are important enterprises which ought to be pursued to a much greater extent than currently is the case. Harold 33 and Margaret Sprout make the point well: No one could believe more strongly than we that the historical study of any subject is intellec tually rewarding and socially useful. But for one who calls himself a social scientist to take refuge in history because he cannot "pin-point" the future is not only to forsake another reward ing and useful sphere of research and analysis, but is also to misconceive the nature of predic tion and the uses men make of it in the business of living. The issue is not whether to predict or not to predict. The issue is rather what one may profitably try to predict, and how to go about it. (Sprout and Sprout, 1961: 66) Summary In this introductory chapter we■have initiated our exploration of prediction and forecasting in inter national relations with an examination of some basic theoretical and philosophical issues. We began with an investigation into the relationship between explanation and prediction and concluded that although on purely abstract, logical grounds, a theoretical proposition should have both an explanatory and predictive capability, more practical and realistic considerations dictate that the presence of both attributes is not essential. After defining some key terms of reference, we examined the basic nature of international phenomena and concluded that 34 most scholars in the field agree that international relations are composed of both patterned and unique events. It then was asserted that the task at hand is to discover and isolate these recurrences and regulari ties. The arguments for and against the feasibility of prediction and forecasting were, reviewed in the next section and it was concluded that it probably is both theoretically and technically possible to obtain useful results from such research efforts. In the final section of the chapter we reviewed some of the normative factors to be considered in prediction and forecasting and decided that the possible benefits to be derived from such efforts outweigh the inherent risks. Now that the necessary theoretical and philosophi cal groundwork has been laid, it is possible to proceed to a discussion of the research strategy to be followed in this study. Our major purpose, in brief, shall be to investigate, principally through the analysis of empirical data, several related theoretical propositions which, when considered together, may permit us to gain some new insights into the problems of prediction and forecasting in international relations. 35 References for Chapter I Bailey, Thomas A. A Diplomatic History of the American People. New York: Appleton-Century-Croft. 7th ed. 1964. Bobrow, Davis B. "International Indicators." Paper prepared for delivery at the 1969 Annual Meeting of the APSA, September 1969, New York. Boulding, Kenneth. "Is Peace Researchable?" Background, Vol. 6, No. 4 (1963), 70-77. Crowe, Sir Eyre. "Memorandum on the Present State of British Relations with France and Germany." British Documents on the Origins of the War, 1898- 1914. Ed. by Gooch and Harold Termperley. Vol. 3. London: H.M.S.O., 1928. Dakin, A. Calvinism. Philadelphia: The Westminster Press, 1946. Dozer, Donald M. The Monroe Doctrine: Its Modern Signifi cance. New York: Alfred A. Knopf, 1965. Fisher, H. A. L. A History of Europe. Boston: Houghton Mifflin Co., 1939. Friedlander, Saul. "Forecasting in International Rela tions." Futuribles: Studies in Conjecture. Ed. by Bertrand de Jouvenal. Vol. II. Geneva: Librairie Droz, S.A., 1965, 1. Gleditsch, Nils Petter. "Trends in World Airline Pat terns." JPR, No. 4 (1967), pp. 366-408. Haney, Lewis H. Business Forecasting. New York: Ginn and Company, 1931. 36 Hempel, Carl G. and Paul Oppenheim. "The Logic of Explanation." Readings in the Philosophy of Science. Ed. by H. Feigl and M. Brodbeck. New York: Appleton-Century-Croft, Inc., 1953. P.'. 323. Herz, John H. International Politics in the Atomic Age. New York: Columbia University Press, 1959. Hoffman, Stanley, ed. Contemporary Theory in International Relations. Englewood Cliffs, N.J.: Prentice-Hall Inc., 1960. Hoggard, Gary D. "Indicators of International Inter action." University of Southern California, 1969. (Mimeographed) Kaplan, Abraham. The Conduct of Inquiry: Methodology for Behavioral Science. San Francisco: Chandler Publishing Co., 1964. Kaplan, Morton. System and Process in International Politics. New York: Wiley, 1957. Karsten, Karl. Scientific Forecasting. New York: Greenbore Publisher, Inc., 1931. Kerner, Robert S. The Urge to Sea. Berkeley: University of California Press, 1946. Lall, Arthur. Modern International Negotiations, Princi ples and Practice. New York: Columbia University Press, 1966. Lerche, Charles O., Jr. and Abdul A. Said. Concepts of International Politics. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1963. Lundberg, George A. "The Postulates of Science and Their Implications for Sociology." Philosophy of the Social Sciences: A Reader. Ed. by M. Natanson. New York: Random House, 1963. 37 Marget, Arthur W. "Morgenstern on the Methodology of Economic Forecasting." Journal of Political Economy, Vol. 37, No. 3 (1929), 312-339. McClelland, Charles A. "International Interaction Analysis in the Predictive Mode." University of Southern California, 1969, pp. 2-3. ________ . "The Access to Berlin." Quantitative Inter national Politics, Ed. by J. David Singer. 1968. ________, et al. Communist Chinese Performance in Crisis and Noncrisis; Quantitative Studies of the Taiwan Straits Confrontation, 1950-1964. China Lake, Calif.: Department of the Navy, U.S. Naval Ord nance Test Station, May 1967. ________ . "The Function of Theory in International Rela tions." Journal of Conflict Resolution, Vol. IX, No. 3 (1960. Meehan, Eugene. Explanation in Social Science: A System Paradigm. Homewood, 111.: The Dorsey Press, 1968. Mills, C. Wright. The Marxists. New York: Dell Pub lishing Co., 1962. Morgenstern, Oskar. Wirtschaftsprognose: Eine Unter- suchung ihrer Voraussetzungen und Moeglichkeiten. Vienna, 1928. Morgenthau, Hans. Politics Among Nations: The Struggle for Power and Peace. New York: Alfred A. Knopf, 1948. Organski, A. F. K. World Politics. New York: Alfred A. Knopf, 1961. Perkins, Dexter. A History of the Monroe Doctrine. Rev. ed. Boston: Little, Brown and Co., 1955. 38 Richardson, Lewis F. The Statistics of Deadly Quarrels. Pittsburgh: The Boxwood Press, 1960. Rubin, T. J.; E. M. Kraus; and A. H. Schainblatt. Project ed International Patterns II. Santa Barbara, Calif.: General Electric-Tempo. September 1967. Pp. 53-55. Russett, Bruce M. Trends in World Politics. New York: Macmillan Company, 1965. Scheffler, Israel. The Anatomy of Inquiry. New York: Alfred A. Knopf, 1963. Singer, J. David and Melvin Small. "Alliance Aggregation and the Onset of War, 1815-1945." Quantitative International Politics. Ed. by J. David Singer. New York: The Free Press, 1968. Pp. 247-286. Sprout, Harold and Margaret S. Sprout. "Explanation and Prediction in International Politics." Inter national Politics and Foreign Policy: A Reader in Research and Theory. Ed. by James N. Rosenau. Glencoe, 111.: The Free Press of Glencoe, 1961. Pp. 60-72. Tanter, Raymond. "Explanation, Prediction, and Forecast ing in International Politics." Forthcoming in The Analysis of International Politics. Ed. by James N. Rosenau, Vincent Davis, and Maurice East. 1969. Thompson, Kenneth W. "Toward a Theory of International Politics." American Political Science Review, XLIX, No. 3 (September 1955), 733-746. Wriston, Henry. "Thoughts for Tomorrow." Foreign Affairs, Vol. 40, No. 3 (April 1962), 374-391. CHAPTER II RESEARCH STRATEGY In the preceding chapter, we attempted to establish that it is both possible and necessary to make forecasts of international phenomena. This study will attempt to investigate some areas of research that appear to offer promise of a modest improvement in our ability to locate predictive relationships, project them, and produce relevant and useful forecasts. In order to make the effort feasible and to clarify its objectives, several specific limitations will be placed upon the analysis. First, the study will be confined to the investi gation of short-term forecasting. The major reason for this limitation, as pointed out in detail in Chapter I, is that there are very significant differences between short-, mid-, and long-range forecasts. The focus of these differences is on the derivation of predictive relationships. It is quite possible for one to establish, for example, a high correlation between an independent 39 40 and a dependent variable on the basis of a few years of data that become insignificant when a long run of data is tested. As a consequence, one often discovers that com pletely different independent variables are required to predict a given dependent variable over the short, mid, or long term. One also finds that the same projection technique is not always equally suited to diverse ranges of forecasting. A linear least-squares regression line, projected out a few months or, in some cases, even a few years may prove quite satisfactory, but often leads to large errors if used in longer term projections. In the latter cases often it is necessary to resort to the use of nonlinear regressions, time series, or other more sophisticated (and complicated) methods which are able to better account for the subtler aspects of a predictive relationship that one must take into consideration in a longer-term projection. Another factor to be taken into consideration is the availability of the data required in making varying length forecasts. It seems almost tautological to state that the longer the desired forecast, the more data that are required to establish the predictive relationship. 41 It is an important consideration when one takes into account, however, the paucity of availability data which are (a) accurate enough to be of use, (b) capable of measuring the variables one considers relevant, and (c) can be found for extended periods of time. An addi tional factor to be taken into account is that one often wishes to (or must) make actual forecasts of the future as opposed to "post-dictions1 1 of events that in reality already have occurred. If the forecast is short-term— one month or one year or even three years— the feedback for eventual comparison of the expected to the actual is soon enough to permit reasonably continuous development of one's predictive hypothesis. Having to wait ten or fifteen years to obtain the results fairly effectively prevents such development. When all of these factors were considered together, it was decided that, given the resources at hand, we would have the greatest chance of success if we concentrated on the study of what probably is the easiest task— short-term forecasts of three years or less— thus leaving the mid- and long-range efforts to later studies. 42 A second decision was that our principal concern would be with academic forecasts. In actual practice, it probably is impossible to separate strictly the academic from the practical, but it was considered necessary that a choice of emphasis be made since the nature of the analysis undertaken largely would be determined by the option chosen. Thus, instead of concentrating, for exam ple, on a search for variables which predict the outbreak of military conflict, one might consider instead the nature of the interdependence between contiguity and international behavior or the relationship of salience to interaction. It probably is apparent that this con centration is an attempt to serve what was asserted in Chapter I to be a basic function of academic forecasts: the testing of (predictive) hypotheses for the purpose of assessing their relative merit under as wide a variety of conditions as possible. It should be pointed out also that our concentration on academic forecasting in no way precludes the possibility that our results could be utilized to help solve essentially practical problems. In point of fact, it is expected that many of our con clusions will be applicable, directly or indirectly, to 43 such questions. This study also is based upon a clear distinction between field theoretic and system analyses of inter national behavior. Field Theory originally was developed in other social science disciplines, particularly Psychology and Sociology, but was introduced to the field of international relations by Professor Quincy Wright in 1955.^ Wright presented a central theoretical proposition of his field theory of international relations in the following way: The state of international relations during a period of time may be defined in terms of the power and the policies of each of the principal govern ments of the world at each moment, and of trends of change in both power and policy.2 Rudolf Rummel, the major current proponent of this approach, disagrees with Wright in some ways but is in accord with Wright on this point: See Wright, 1955. For an extensive discussion of field theory in Psychology, see Cartwright, 1951. As will become apparent, we have depended very substantially on Professor McClelland's article, "Field Theory and System Theory in International Politics" (1968) , for our arguments in this section of the chapter. 2 Wright, 1955, p. 544, as quoted in McClelland, 1968, p. 9. 44 The notion is that the nature of the distance between two social units or their attributes is a force determining their behavior toward each other. Time is built into the theory in terms of the movement of dyads in behavior space. The attri butes of individuals are differentially undergoing constant change. As these changes affect the position of the social unit vectors in attribute space, the location of dyad vectors in behavior space will shift as a function of time to trace out a behavior path.^ Rummel has used a wide variety of variables to measure the attributes of nations including indicators of size, wealth, governmental structure, stability and so on. The overwhelmingly important point to be made in this dis cussion, however,— one on which both Wright and Rummel agree— is that a central concept of field theory is the notion that it is possible to predict the behavior of nations on the basis of their respective attributes. In an article initially published in 1955, Professor Charles McClelland first introduced the General System Theory concepts of Ludwig von Bertalanffy to the field of International Relations (McClelland, 1955).^ He ^Rummel, 1965, p. 185, as quoted in McClelland, 1968, p. 9. ^See also Morton Kaplan's book published in 1957 (Kaplan, 1957) for a later attempt to develop and expli- 45 since has pointed out in some detail the similarities and differences of System Theory and Field Theory in his paper, "Field Theory and System Theory in International Politics" (McClelland, 1968). In that exposition, Professor McClelland asserts that the two approaches have much in common. He cites ten philosophical points of agreement (McClelland, 1968: 6-7) before identifying three major theoretical convergences: First, both approaches conceive of international relations as one of the many specific cases that can be identified as part of the general theory. Second, both theories claim to be conceptually based on mathematical and statistical relationships. That is, the international relations aspect of them is manifest in statistical analyses of aggregate data. Third, they share many of the same conceptual terms (i.e., system: field, component: social unit, interaction: relation) and are predicated on the common assumption that "entities (usually nation-state connected . . .) each having their own distinctive characteristics, act on each other and produce relationships" (McClelland, 1968: 8) . cate some of these concepts. 46 Even considering all of these similarities, however, one can see that the two approaches are signifi cantly different (perhaps opposite or complementary) when viewed from the perspective of the nature of the predic tive relationships sought. The system theory point of view is grounded in the notion that one must know the structural characteristics of the system and the relevant operational processes in order to predict the behavior of the entities (nations) acting in that system. As has been discussed, however, field theorists are interested in which nation attributes most accurately predict inter national behavior. These positions thus are mutually exclusive— McClelland believes them to be somewhat "com plementary" (McClelland, 1968: 13)— and although benefit no doubt can be derived from research based on either approach, this study will confine itself to the investi gation of relationships that are System Theory oriented. As a result, our independent variables generally will be types or amounts of international behavior rather than size, wealth, or productivity and our inquiry will be oriented to the exploration of system and/or subsystem structure and process. 47 The fourth limitation on the study actually is a "nonlimitation.1 1 That is, where the other items were intended to place bounds on the analysis, this one is aimed at permitting greater freedom of inquiry. The issue is whether to take an inductive or a deductive approach in the analysis. It will be recalled from our discussion in Chapter I that conducting an investigation in either a purely inductive or purely deductive fashion probably is not only impossible but also is undesirable. In any event, we do not wish to place unnecessary restrictions on our analytic procedures and at the same time want to avoid the pitfalls of pursuing either method in its extreme form. As the reader will note in the following chapters, our tendency will be to pursue an approach that usually is more inductive than deductive, but always will be a mix of the two. Thus, we will not hesitate to change emphases when it is determined that doing so will improve the analysis in some manner. In the discussion so far, we have pointed out that the central goal of the study is to pursue a line of inquiry that seems to offer some promise of improving our ability to test predictive relationships and to make 48 forecasts of international behavior. We then established four guidelines for the inquiry. First, we will be con cerned only with short-term predictive relationships and forecasts of up to three or four years. Next, the inves tigation of academic questions will be our main preoccupa tion, although it is expected that many of our results will be applicable to more practical problems. The effort will proceed also within the framework of the system approach to the study of international politics. Thus, where we would be attempting to account for behavior in terms of a nation's attributes if working within a field theoretic framework, here we will be concentrating on the analysis of system structure and process. Finally, we will proceed in a generally inductive manner, although not in any rigid or extreme manner, thus permitting as much latitude as possible in the conduct of the study. The research design for "the analysis is based on the assumption that research in the field of international relations can be conducted in a scientific fashion. This question was examined in considerable detail in Chapter I so that it is unnecessary to repeat the discussion here. It probably will be sufficient to recall that we attempted 49 there to draw an analogy between the field of Economics as it was developing in the 1930's and the field of International Relations as it is evolving today. The major point is that International Relations probably is in one of the classic early stages in the development of a science. F. S. C. Northrop describes these natural stages of inquiry in the following passage: . . . the first stage of inquiry ends when the analysis of the problem has designated the facts which must be known in order to resolve the prob lem, (and) the task of the second stage and the scientific method appropriate to it are thereby determined. Its task is to inspect the relevant facts designated by the analyses in the first stage of inquiry. The scientific method appro priate for this is the inductive method of Bacon. As a rule, it involves not one method but three; namely, the method of observation, the method of description, and the method of classifi cation. . . . The second stage of inquiry comes to an end when the facts designated by the analysis of the problem in the first stage are immediately appre hended by observation, expressed in terms of con cepts with carefully controlled denotative mean ings by description, and systematized by classifi cation . . . . (Northrop, 1947: 35. Emphasis added.) Many scholars in the behavioral sciences have pointed out that in general there is a critical need for develop ing classification systems which are capable of operation alizing concepts: What makes a concept significant is that the classification it instituted is one into which 50 things fall, as it were, of themselves" (Kaplan, 1964: 10). Eugene Meehan, in his excellent book, The Theory and Method of Political Analysis, makes the point very suc cinctly when he states that science is systematic and logical. At the very lowest level of generality, scientific explana tion requires the systematic ordering and classifi cation of empirical data. Indeed, some of the physical sciences have not moved very far beyond this kind of taxonomic activity. . . . (Meehan, 1965: 40)5 Thus, if many of the social and even some of the physical sciences "have not moved very far beyond this kind of taxonomic activity," it is improbable that the field of international relations has reached this level of develop ment. Indeed, some scholars in the field have been of this same opinion. R. J. Rummel, for example, asserts that grouping nations, objects, individuals, or cases by types is a basic step in describing phenomenon and building science. The virtue of typing is that it enables parsimonious descriptions of 5 The phrase which follows is "and in political science even an agreed taxonomy has yet to appear." It was not included in the quotation because it was unclear what the author meant by "an agreed taxonomy in political science." It was felt that its inclusion tended to obscure what was interpreted as his main point. 51 objects and facilitates reliable predictions about them based upon their tendency to group.b In the context of the system approach, Northrop's first stage probably was completed when the actual behav ior of nations— what they do and say to one another on a daily basis— was designated as the set of facts which must be known and systematically recorded. The appearance of several coding systems designed to provide "carefully controlled denotative meanings" for the observed phenomena marked the entrance of analytic efforts in this area into the second stage.^ Our conviction which has influenced strongly the research strategy to be followed in the study, is that we probably still are in this stage. On that basis, it was decided that the most promising route to improving our ability to test for predictive relation ships, and subsequently to use our findings to make g R. J. Rummel, as quoted in Phillips and Hall, p. 3, n.d. (emphasis added). See first section of Chapter III for additional support for this position. 7 Category systems which have been used to date include those used in the WEIS Project (Fitzsimmons et al., 1969) , the DON Project (Rummel, 1966), and by Walter Corson (Corson, 1967) and Donald Puchala (Puchala, 1967). See Martin and Young, 1968, for a discussion and compari son of these classification systems. 52 forecasts, would be in the area of developing taxonomies which are capable of describing international behavior in more accurate and relevant terms than currently is possible. Such an effort will constitute the main thrust of the study. Within the parameters established above, two classification systems will be developed. The first will be a taxonomy of nations in the international system which will group polities on the basis of their inter actions with other actors in the world. Data from the World Event/Interaction Survey (WEIS) at the University of Southern California will be used in this analysis. These data constitute all of the international acts initiated by every actor in the world, as recorded in a single data source for the period 1966-1969. They are collected and coded in the form "who did what to whom" and number over 25,000 discrete items.8 A final attempt will be made to guard against the possibility that the final taxonomy represents an artifact of the statistical 0 See the second section of Chapter III and WEIS Technical Report #1 (1969) for a more detailed descrip tion of these data. 53 technique used by our becoming as familiar with the data as is feasible and by using two different multivariate techniques in the analysis. After a classification system is developed each of the groups of nations will be described in some detail. The final analytic section of this portion of the study will examine the temporal aspects of the taxonomy and compare our results with those of other researchers in the field who have grouped nations on the basis of different kinds of data, e.g., sociocultural attributes. Such a comparison will permit us to draw some preliminary conclusions regarding differ ences between the type of phenomena being used to group nations here and those used by other analysts. Where the taxonomy of nations is actor-centered, the second classification system will be event/interaction- centered. That is, a system will be developed which is intended to group events rather than nations. The idea is that nations, in going about the daily business of attempting to carry out their foreign policy objectives, g launch many different "undertakings'* which in turn g The idea of an "undertaking" originally was con ceived by H. and M. Sprout in Sprout and Sprout (1962), 54 involve them in a wide variety of situations. The notion of "situation" is new neither to political science nor some of the other social sciences. The systematic cate gorization of the specific notion of situation we have in mind, however, probably is somewhat novel.10 On a mostly intuitive basis, we know that some of these inter national situations recur in a fairly regular manner. The best known examples, perhaps only because they have been the subject of the largest amount of research, are international crises and international bargaining or negotiations. There are, however, many other common situations for which it would be useful to have not only labels but operational definitions as well. In order to derive such an exhaustive taxonomy, data will be collected from the front page of the New York Times for the two years 1955 and 1967. Where the WEIS data used in the nation classification system were of the form "who did what to whom," these data items will contain this same information pp. 105-107 as quoted in McClelland, 1970. 10 Other concepts of situation in the social sciences will be discussed.in detail in the first section of Chapter V. 55 but also (and most importantly) will include information on the situational context, i.e., who did what to whom about what. - * ' 1 With these data in hand, we will proceed to a derivation of the category system in a somewhat different fashion than was the case for the nation taxonomy. There the system was derived in a purely empirical fashion by statistically manipulating the data from the WEIS collec tion which had been coded according to the WEIS category system. The twenty-two mutually exclusive and exhaustive categories used there had been derived over several years by Charles McClelland and his associates from systematic observation of daily event interactions.-^ The intention here is to develop, on a first-cut, pilot basis, a cate gory system comparable to that used in the WEIS project to code event/interactions only in this case the purpose will be to gain the ability to code international See second section of Chapter V, for a more detailed discussion of the data collection, coding sources, possible bias, etc. 12 See McClelland, 1970, pp. 30ff, for a discussion of this process. 56 situations. In brief, the procedure will be to go through each of the 3,000 or so data items collected and attempt to label each of them with an appropriate descriptor. A list of labels then will be compiled, with categories and subcategories created from the noted redundancies. After checking to be sure the categories all are on the same level of abstraction, all of the data will be coded according to the newly-created category system and the relative frequency of occurrence of each determined for the time periods analyzed. The reader probably has noted that although we attempted to establish the importance of taxonomic systems in the development of the field, up to this point no reason for the selection of these two specific classifica tion systems has been provided. Although a detailed rationale will be provided in the introductory section to each of the two analyses, the following quotation from a recent paper by Professor Charles McClelland provides a general motivation for our selection: . . . we conceive that it is important to be able to classify international actors according to 'state descriptions 1 of their actions. . . . Earlier it was said that the task of international system analysis is to discover which interaction processes occurring repetitively are associated 57 with which types of reappearing international situations. If we reject the idea that polities are alike, we can round out the requirements of international system analysis to a complete form by defining the task as that of discovering which types of actors participate in which interaction processes associated with which types of inter national situations. . . . (McClelland, 1970: 27) While we are not so ambitious to "round out the require ments of international system analysis to a complete form," it is felt that the derivation of classification systems in these two particular areas— nations and situations— may make a modest contribution to such an effort. More important to the specific interests of this study is, however, the use of the two derived taxonomies in the attempt to improve our ability to postulate pre dictive relationships and subsequently test their validity as well as trying to ascertain whether or not, or when, they are able to improve our forecasting capabilities. It is for this reason that the final section of the study will be devoted to preliminary empirical tests of the two taxonomies. The central purpose of this effort will be an attempt to demonstrate their usefulness in ascertaining the existence of predictive relationships and making forecosts. Concurrently, we will endeavor to demonstrate the operational differences between 58 prediction, projection, and forecasting, a crucial distinction discussed at some length in Chapter I. Ini tially, we will test a simple predictive relationship first without differentiating between nation types and then using that distinction so that the results can be compared and the apparent effect of the taxonomy evaluated. A projection and forecast then will be made which is based upon the established predictive relationship. Next, data categorized according to the situational taxonomy derived will be analyzed in an attempt to test our hypothesis that nations tend to exhibit different patterns of behavior in varying situational contexts. Finally, a summary of the major findings of this study will be presented. 59 References for Chapter II Corson, Walter H. “Measuring Conflict and Cooperation Intensity in East-West Relations: A Manual and Codebook." Ann Arbor: Institute for Social Research, University of Michigan. Second Draft, November 19 69. Fitzsimmons, Barbara; Gary Hoggard; Charles McClelland; Wayne Martin; and Robert A. Young. “World Event/ Interaction Survey Handbook and Codebook." World Event/Interaction Survey Technical Report #1, School of International Relations, University of Southern California, January 1969. (Mimeographed) Kaplan, Abraham. "The Conduct of Inquiry. San Francisco: Chandler Publishing Co., 1964. Kaplan, Morton A. System and Process in International Politics. New York: John Wiley and Sons, Inc., 1957. McClelland, Charles A. "Applications of General Systems Theory in International Relations.“ Main Currents in Modern Thought, Vol. 12 (November 1955), 27-34, as reprinted in James N. Rosenau, ed., Inter national Politics and Foreign Policy. New York: The Free Press, 1961. Pp. 412-420. ________. "International Interaction Analysis: Basic Research and Some Practical Applications." World Event/Interaction Survey Technical Report #2, School of International Relations, University of Southern California, 1968. (Mimeographed) ________. "Some Effects on Theory from the International Event Analysis Movement." World Event/Interaction Survey, School of International Relations, Univer sity of Southern California, 1970. (Mimeographed) 60 Martin, Wayne R. and Robert A. Young. "A Review of Six International Event/Interaction Category and Scaling Methods." School of International Relations, University of Southern California, 1968. (Mimeographed) Meehan, Eugene J. Analysis. 1965. The Theory and Method of Political Homewood, 111.: The Dorsey Press, Northrop, F. S. C. The Logic of the Sciences and the Humanities. New York: Meridian Books, Inc., 1947, Phillips, Warren and Dennis Hall. "The Importance of Governmental Structure as a Taxonomic Scheme for Nations." The Dimensionality of Nations Project, Research Report No. 18, Department of Political Science, University of Hawaii (n.d.). (Mimeo graphed) Puchala, Donald J. "Operational Handbook for a Study of Intergovernmental Interaction During Regional Integration." New York: Columbia University, 1967. (Mimeographed) Rummel, Rudolph J. "Testing Some Possible Predictors Within and Between Nations." Peace Research Society Papers, 1964. Sprout, Harold and Margaret Sprout. Foundations of Inter- national Politics. Princeton: D. Van Nostrand Co, Inc., 1962. Wright, Quincy. The Study of International Relations, New York: Appleton-Century-Crofts, 1955. CHAPTER III AN INTERNATIONAL BEHAVIOR TAXONOMY: I The usefulness of a behavioral taxonomy of nations such as that to be derived below was discussed in some detail in Chapter II. There, however, our goal was to demonstrate the possible utility of such a classification in the first-cut prediction paradigm being developed in this study. While this use of course is paramount in our effort, there are other functions the taxonomy might fill which, although not relating directly to prediction, nevertheless may contribute in a modest way to the development of a more scientific approach to future research in the field.^ Theoretical Considerations Political scientists have been grouping political entities at least since the time of Plato, who classified While we reorganize the interrelated nature of systematic knowledge and prediction, it is useful at this juncture to distinguish between them. 61 62 o city-states as either agricultural or industrial. In more recent times we have come to recognize and accept readily the grouping of nations according to geographic A region, level of development, etc. Few problems arose until researchers began attempting to operationally define such concepts. R. J. Rummel clearly identifies the diffi culty: The problem with the prevailing types in political science is that the rationale underlying the cate gorization is often not explicit. It is not clear whether our types really divide different kinds of variance. If we are to deal in types, a clear and empirical basis for the distinctions must be made, such as the variance within each of our groups being less than the variance between our groups. The rapid development of the social sciences after World War II caused more and more scholars in political science to postulate the need for much greater conceptual clarity in the field, particularly in the area of operationally defined taxonomies. Snyder, Bruck and Sapin, in an article first published in 1954, stated that "... theoretical 2 Plato, The Republic. 3 R. J. Rummel, Dimensions of Nations, forthcoming, 1969, as quoted in Phillips and Hall (n.d.). ____________________________ 63 progress in the study of international politics will require eventually a typology of states ..." (Snyder, Bruck, and Sapin, 1954: 187) .4 In the quotation above, Rummel asks for a clear underlying rationale and derivation of types of nations. Brian J. L. Berry, a well-known economic geographer, expands upon this idea in asserting that too "many authors treat the (underdeveloped) world as essentially monolithic, paying only lip service to diversities and heterogeneity." Berry felt it essential that some "substantive flesh" be put on our intuitive image of given types of nations. He further suggested that it is not enough that we assert even a well-defined group of nations but such a group must be further examined to see how nations might "string out along the various continuum" by which we have derived our typology (Berry, 1961: 78). Bruce Russett suggests still another reason for the empirical development of nation taxonomies. He sees 4 The authors went on to say that this typology should be based on many factors including decision-making systems and types of foreign policy strategies employed. It is the latter aspect that most closely resembles what we have in mind here. 64 them as at least a step toward the solution of what he labels the "many roads" problem. In his words, the difficulty encountered in many quantitative analyses is that "the meaning of a variable will be different in different social or political contexts" (Russett, 1967: 8-9). This argument is quite similar to what we have come to call the "overaggregation problem." Despite the tradition in political science that suggests that all nations are basically the same— have the same underlying motives, employ the same methods of statecraft, and so on— it may be a major error to make this assumption. What we are suggesting, along with Russett, is that attempts to derive "global relationships" or relationships which purport to be equally applicable to all nations at all times under all circumstances may be an extremely dysfunctional research approach. Russett uses the example of a positive relationship that might exist between political violence and per capita income for poor coun tries which are in the early stages of development. The relationship between the two variables could become negative, however, when only rich nations are included in one's sample. An example more applicable to behavioral data such as that which concerns us here might posit a 65 relationship between the proportion of a nation's total activity which is classified as "nonmilitary conflict" and the amount of military conflict in which it engages. Were we to test for such a relationship for all nations in the system arranged in the same distribution, it is quite possible that the correlation would approach zero. If one controlled for type of nation (one type might engage in such behavior on a routine basis while for another it is unusual), however, it is possible that we would discover a very strong predictive relationship in the latter group and a weak or nonexistent one in the former. Thus, had we accepted the results of the analysis in which nations were undifferentiated, we would have had to conclude that no relationship between nonmilitary conflict behavior and military conflict behavior existed for any nation in the international system when in fact a strong one existed for a specific type of nation. Whatever the motivation, the response in terms of different typologies of nations appearing in the litera ture in recent years has not been insignificant. Un fortunately, very few have even considered the desirabil ity of grouping nations on the basis of some aspect of 66 their behavior rather than according only to nonbehavioral descriptors of nations such as political system, level of development, etc. Indeed, probably the only attempt at the former is that of Russett in which he grouped nations according to their voting behavior in the United Nations (Russett, 1966). Many of the latter, some of them quite innovative and useful, have appeared during the past several years, although for varying reasons no attempt usually was made to include all the nations in the international system. Examples of these nonbehavioral taxonomies include groupings of nations by Banks and Gregg by political system (Banks and Gregg, 1965), Russett by both socioeconomic level (Russett et_al., 1964) and sociocultural level (Russett, 1967),5 Cutright by politi cal development (Cutright, 1963), Berry by economic development (Berry, 1961), Riggs by governmental structure (Riggs, 1968), Peierabend and Feierabend by degree of 5 Russett also groups nations on the basis of several other criteria in Russett, 1967. g As presented in Phillips and Hall (n.d.). 67 7 political stability (Feierabend and Feierabend, 1966), and Lipset by the stability of democracies and dictator ships (Lipset, 1960). Given the lack of behavioral groupings and the apparent usefulness of such taxonomies, we feel justified in asserting that the availability of a taxonomy such as that derived below will be of some value not only in this study but perhaps in others as well. The research strategy to be followed will be straightforward. The first step will be to examine the data to be used, assess its source and select the actors to be included in the analysis. A basic, comprehensive data matrix then will be presented, its underlying dimen sions delineated, and a preliminary typology derived using two different statistical techniques. Thereafter, alternate subsets of data and different variables will be examined in an attempt to learn as much about the data, 7 The Feierabends also array nations on the basis of several other factors. For an example of some, a very early use of (and pioneering) Q-Factor Analysis to group some nations, see Cattell, 1949 and Cattell, et al., 1951. Other early groupings of nations include those by Eckstein (1962), Coleman in Almond and Coleman (1960), and Fitzgibbon and Johnson (1961), and Lipset (1960). 68 the actor groupings, and the relationships between them, as possible. These further analyses will be of the same basic body of data aggregated in different ways. After a final version of the taxonomy is derived, it will be compared with some of those currently in use in the field which are based on quite different data, e.g., Russett's sociocultural groupings. Finally, some of the temporal aspects of the taxonomy will be considered, particularly as they relate to the projection of the results to future time periods. Data Source and Preliminary Variable Selection The data used in this analysis are those collected in the World Event/Interaction Survey (WEIS) at the University of Southern California. When our study began, four years of data— January 1, 1966 through December 31, 1966— were available. The main WEIS data collection consists of all public, international acts reported in the New York Times. One of the principal aims of the project is to obtain information about what nations in the international system say and do to one another. In 69 g a simple and familiar phrase, "who does what to whom.1 1 The four-year data collection includes some 25,760 discrete event/interactions initiated by 159 different entities. Many criticisms have been leveled at the data collection regarding its validity and reliability. Many probably are justified. In an early WEIS Technical Report, Professor McClelland discussed this problem at great length. He concluded that we should expect some mistakes, some distortion, and some erroneous selection in the content of the international news. Some items will report events that never happened. Others will be products of governmental manipulation and deception and will have made their appearance on the record because they were "planted" there. How bad these defects are can be estimated only in very general terms. Several points can be made in support of the general reliability of the news: 1. There is no doubt that the volume, detail, and accuracy of international news reporting have greatly increased over the past half century. . . . 2. The professionalization of journalism and the maturing of the newsgathering organizations g We have not attempted to provide an extensive description of the WEIS data collection. Detailed infor mation about the collection and coding of the data is provided in WEIS Technical Report #1 (Fitzsimmons et al., 1969). The theoretical underpinnings of the project are delineated in WEIS Technical Report #2 (McClelland, 1968) and WEIS Technical Report #3 (McClelland, 1969) . 70 are important forces in increasing the effort toward accuracy. 3. Multiple sources and channels of informa tion provide some safeguard against the systematic falsifications of reports. The sovereignty of states and the diversity of national interests contribute some beneficial effects, globally, against news management. 4. The real source of large numbers of international news stories is the government agency. The news media are, essentially, pipelines between the regime and the public. One of the oldest maxims of statecraft,, stated by Machiavelli among others, is that it does not pay to lie too often. . . . 5. "Hard news" has a more straightforward status than the news that is explanatory, inter pretative, speculative, or predictive. . . . (Note that the latter type of data are not included in the WEIS collection.) 6. Although the public record of international events is defective by any strict standard, it serves well enough for many kinds of analysis. It is to be argued that public reporting can be used more effectively for statistical evaluations than for the purposes of historical reconstruction— its most common use. It is unlikely that a few erroneous reports will make a big difference in an aggregate of several hundred items. . . . (McClelland, 1968: 43-48) The issue of the American bias of the New York Times, or the particular national bias of any information source, particularly newspapers, also is important to consider. The hard fact for us to face is there probably is no source we can turn to that is not biased in some way. The Times (London) reflects the British view of the world, LeMonde the French, and La Prensa the Argentinian. 71 "When everybody has a bias, what test could there be for Q objectivity?" Under such circumstances, Professor McClelland and his associates chose to begin their data collection efforts using the New York Times exclusively, since there is very substantial agreement among scholars and practitioners that it is the single most comprehensive source of international events available.-1 - 0 Granting that the WEIS data collection is far from perfect, it provides, nevertheless, a large, machine-readable, and systematic body of information from which rough indicators may be constructed and used in the development of the behavioral typology such as that sought here. The single most important caveat is that the results of the analysis must (and will) be interpreted with the bias of the data source receiving full consideration. A possible result of such bias is the lack of a significant amount of data for many entities, especially 9 Charles A. McClelland, 1968, p. 46. See pp. 38- 49 for an excellent examination of this problem. ^■°See forthcoming WEIS Technical Report on this subject, Merrill, 1964; Gould, 1969. 72 in Latin America and Sub-Saharan Africa even when the scores for the total four-year period are aggregated. Whatever the reason, one can see from Table 1 that the total number of acts recorded falls off quite rapidly after the top twenty or so entities are listed, so that an attempt to.include all of them could seriously affect the validity of the analysis. For that reason, an arbitrary minimum number of acts of 32 was set as a criterion for selecting the nations to be included in the analysis. This would mean that when the data were arrayed by nations over the 4 8 months, every nation would have at least a theoretical chance of filling 75 percent of its cells in this one by forty-eight matrix. That is, it would be possible for 32 out of the 48 cells, i.e., months, to have one act recorded in them. A second criterion was that only sovereign nations were included since we had no theoretical interest in how others, such as the many Arab guerrilla organizations or NATO, might group. Table 2 contains a list of the 83 nations selected for inclusion An alternative explanation is that most of the nations in those regions actually participate very little in international affairs. For some preliminary conclu sions regarding this possibility, see Hoggard et al., 1970 73 TABLE 1 R A N K O R D E R O F ALL W EIS A C T O R S B Y N U M B E R O F A C T S O U T P U T 1966 - 1969* A C T O R F R E O . A C T O R FR E Q . A C T O R F R E Q . A C T O R F R E Q . U S A 4505 G R C 1 3 1 C H T 49 A F G 13 U S R 2224 A LG 131 Y E M 48 M LT 12 ISR 1445 T U R 121 SIN 45 E F T 12 C H N 967 L A O 121 C Y P 43 C H A 12 U A R 939 S W D 105 O A S 42 M A C 12 U N O 939 Z A M 101 FIN 4 1 C O N 1 1 V T N 925 R H O 97 O A U 41 BUI 1 1 U N K 907 M A L 94 T U N 40 NIC 9 G M W 725 IRN 93 K U W 40 B A R 8 F R N 685 B E L 90 B O L 39 LU X 8 JO R 497 S A U 89 V E N 36 E B E 8 V T S 496 S A F 87 G UI 36 S E A 8 M L G 420 H U N 86 N O R 34 P A R 7 C Z E 379 P O R 83 M O N 33 A F P 7 IND 341 B R A 80 E LS 32 C E N 7 S Y R 318 N A T 73 H O N 29 D A H 6 JAP 293 C O P 73 E C U 29 TRI 6 G M E 281 P E R 68 IRE 28 G A B 6 V C G 281 N T H 66 G U Y 27 G A M 6 C A N 272 G H A 65 B U R 25 R W A 6 P L O 256 ALB 62 U G A 24 ICE 5 NIG 232 B U L 59 S O M 23 SIE 5 V A T 232 E E C 59 H O K 23 JA M 4 C A M 228 S U D 59 U R U 22 S A N 4 Y U G 227 W A R 58 AR L 22 NIR 3 R U M 211 A U S 56 C E Y 22 U P P 3 K O N 209 D E N 55 LAP 22 T O G 3 K O S 207 K E N 54 P A N 19 B O T 3 ! P O L 198 C H L 53 D O M 17 M A G 3 C U B 175 TAZ 53 N E P 17 C A O 2 ! P A K 173 M E X 52 H A I 16 M A R 2 I BIA 167 M O R 52 S E N 16 M O C 1 ITA 165 S Y E 52 C O S 14 LIC 1 IRQ 160 N E W 52 M A U 14 A N D 1 1 PHI 156 A R G 5 1 G U E 14 LE S 1 ; INS 146 W B E 5 1 G U A 13 S W A 1 ; TAI 139 C O L 50 M L I 13 M A D 1 LEB 137 LB Y 50 IVO 13 M O M 0 i S P N 135 S W Z 49 LBR 13 N A U 0 A U L 133 E T H 49 M A W 13 TO TAL = 25 ,760 *See Appendix A fo r l i s t of actor abbreviations. 74 TABLE 2 R A N K O R D E R O F 83 N ATIO NS B Y N U M B E R O F A C TS O U T P U T 1966 - 1969* U S A 4505 U S R 2224 ISR 1445 C H N 967 U A R 939 V TN 925 U N K 907 G M W 725 F R N 685 JO R 497 VTS 496 C ZE 379 IND 341 S Y R 318 JAP 293 G M E 281 C A N 272 NIG 232 C A M 228 Y U G 227 R U M 211 K O N 209 K O S 207 P O L 198 C U B 175 PAK 173 ITA 165 IRQ 160 PHI 156 INS 146 TAI 139 LEB 137 S P N 135 A U L 133 G R C 131 A LG 131 T U R 121 L A O 121 S W D 105 Z A M 101 R H O 97 M A L 94 IRN 93 B E L 90 S A U 89 S A F 87 H U N 86 P O R 83 B R A 80 C O P 73 P E R 68 N T H 66 G H A 65 ALB 62 B U L 59 S U D 59 A U S 56 D E N 55 K E N 54 C H L 53 TAZ 53 M E X 52 M O R 52 S Y E 52 N E W 52 A R G 5 1 C O L 50 LBY 50 S W Z 49 E T H 49 C H T 49 Y E M 48 SIN 45 C Y P 43 FIN 41 T U N 40 K U W 40 B O L 39 V E N 36 GUI 36 N O R 34 M O N 33 ELS 32 TOTAL = 22,435 *See Appendix A fo r l i s t of actor abbreviations. 75 in the analysis on the basis of the two criteria speci fied, along with the total number of acts each initiated during the four-year period. By comparing the total number of events initiated by all actors in the WEIS collection, as shown at the bottom of Table 1 (25,760), and the total number output by the 83 nations selected for inclusion in this analysis, indicated at the bottom of Table 2 (22,435), one can see that over 87 percent of all acts recorded during the four years are included in the data selected for analysis. The coding system used in the WEIS project con sists of 63 specific types of events. These events can be aggregated in any way one wishes, but the most fre quently used procedure is to merge them into 22 combined- event (combevent) categories. Table 3 illustrates this process. The category system was derived over a period of several years by Professor McClelland and his asso ciates and consequently plays a central role in the 12 theoretical understructure'of the data collection. For this reason, it was decided to use the 22 combevent 12 See McClelland, 1970, pp. 29-38, for a detailed account of this process. TABLE 3 W EIS EVENT/INTERACTION CLASSIFICATION S Y S T E M 1. YIELD 011 Surrender, y ie ld to order, submit to a rre st, etc. 012 Yield po sitio n; re tre a t; evacuate 013 Admit wrongdoing; re tra ct statement 2. C O M M E N T 021 E x p lic it decline to comment 022 Comment on situa tion -pessi m istic • 023 Comment on situation-neutral 024 Comment on s itu a tio n -o p ti m istic 025 Explain policy or future position 3. C O N S U LT 031 Meet w ith; at neutral s ite ; or send note 032 V is it; go to 033 Receive v is it ; host 4. A P P R O V E 041 Praise, h a il, applaud condolences 042 Endorse others policy or position give verbal support 5. P R O M IS E 051 Promise own policy support 052 Promise material support 053 Promise other future support action 054 Assure; reassure 6. G R A N T 061 Express regret; apologize 062 Give state in v ita tio n 063 Grant asylum 064 Grant p riv ile g e , d ip lo matic recognition; de facto re la tio n s, etc. 065 Suspend negative sanctions; truce 066 Release and/or return per sons or property 7. R E W A R D 071 Extend economic aid (fo r g if t and/or loan) 072 Extend m ilita ry assistance 073 Give other assistance 8. A G R E E 081 Make substantive agreement 082 Agree to fu rth e r action or program; agree to meet, to negotiate 9. R E Q U E S T j 091 Ask fo r information i 092 Ask fo r policy assistance;! seek 093 Ask fo r material assis tance 094 Request action; c a ll fo r 095 Entreat; plead fo r; appeal! to ; help j 10. P R O P O S E ! 101 O ffer proposal | 102 Urge or suggest action or j policy 77 TABLE 3 (Continued) 11. R EJEC T 111 Turn down proposal; reject protest demand, threat, etc. 112 Refuse; oppose; refuse to allow 12. A C C U S E 121 Charge; c r itic iz e ; blame; disapprove 122 Denounce; denigrate; abuse 13. P R O T E S T 131 Make complaint (not formal) 132 Make formal complaint or protest 14. D E N Y 141 Deny an accusation 142 Deny an attrib u te d policy, action, ro le , or position 15. D E M A N D 150 Issue order or command, in s is t; demand compliance, etc. 22. 16. W A R N 19. R E D U C E RELATIONSHIP (as Neg. Sanction) 191 Cancel or postpone planned event 192 Reduce routine in te r national a c tiv ity ; recall o ffic ia ls , etc. 193 Reduce or suspend aid or assistance 194 Halt negotiations 195 Break diplomatic relations 20. . EX P E L 201 Order personnel out of country 202 Expel organization or group 21. SEIZE 211 Seize position or possess ions 212 Detain or arrest person(s) 160 Give warning 17. TH R E A TE N 171 Threat without specific negative sanctions 172 Threat with specific non-m ilitary negative sanctions 173 Threat with force specified 174 Ultimatum; threat with negative sanctions and time lim it specified 18. D E M O N S T R A T E 181 Non-m ilitary demonstration; walk out on 182 Armed force m obilization, exer cise, and/or display F O R C E 221 Non-injury destructive act 222 Non-m ilitary injury-des- tru ctio n 223 M ilita ry engagement 78 categories as the variables for the initial analysis. This first data matrix thus contains data for the 48-month time period with the 83 selected nations in the rows and the 22 combevents in the columns. The data matrix is presented in Table 4. Before beginning the analysis of these data, however, it is necessary that we discuss in some detail the main techniques to be used. Statistical Techniques Two major techniques— Factor Analysis and Hier archical Grouping Analysis— will be used to derive the behavioral typology, although other statistical measures such as product-moment correlations or z-scores may be employed from time to time for specific cases. It is not necessary to go into a detailed discussion of these latter methods at this time since they seldom will be used and can be easily described at the point they are introduced into the analysis. Rather, we will concentrate on describing in some detail Factor Analysis and Hierarchical Grouping Analysis (hereafter H-group analysis), paying particular attention to those aspects of the two techniques which relate directly to our goal of delineating groups or clusters of nations in the international system. TABLE 4 22 W EIS C O M B E V E N T C A TE G O R IES F O R 83 S E LE C TE D A C T O R S 1966 - 1969 NATION 1 ----- 2 3 T" 5 6 7 8 9 ~ id 11 12 ^13"~l4 15 16 17 18 19 20 21 22 2 54 1491 693 172 157 168 251 227 152 268 103 228 56 157 19 89 2 1 27 87 9 13 63 20 1 72 60 9 1 6 16 17 10 24 12 15 a 6 3 3 0 1 8 1 2 1 40 0 22 17 10 3 18 3 2 1 4 3 2 43 4 1 0 3 1 7 3 3 6 1 70 0 6 14 3 0 2 0 13 3 4 2 3 0 0 0 0 0 0 1 1 0 0 92 1 3 2 0 1 0 0 2 2 0 2 9 0 1 0 0 0 2 1 0 0 6 100 0 4 15 2 0 4 0 9 3 3 0 1 0 0 1 0 0 1 1 1 2 3 101 0 6 8 1 0 3 0 3 1 2 3 4 0 1 0 0 0 0 1 0 2 1 135 0 7 15 1 1 7 0 9 1 1 5 4 1 1 2 1 1 1 1 0 8 1 140 0 9 22 3 1 8 2 6 0 4 7 3 1 1 1 0 0 3 1 1 4 3 145 0 10 7 3 1 0 0 5 7 1 0 2 1 0 0 0 0 1 0 0 1 0 155 0 10 10 1 0 1 0 8 3 3 1 4 1 1 1 1 0 3 1 3 0 1 160 0 6 7 2 2 4 2 3 0 2 3 3 1 4 0 1 0 2 4 2 0 3 200 7 209 193 30 27 33 30 62 40 62 39 48 19 1 1 6 16 6 15 36 3 6 9 210 0 7 13 0 2 1 5 7 1 5 4 3 2 2 1 0 0 5 0 2 2 4 211 0 19 18 1 1 4 1 5 8 5 2 9 4 4 0 2 1 1 5 0 0 0 220 4 178 129 28 24 16 19 64 31 46 3 1 45 7 11 6 13 1 8 20 1 2 1 225 0 13 10 6 3 3 0 5 1 0 3 1 2 0 0 0 0 0 1 1 0 0 230 1 2 1 2 1 2 1 12 5 14 3 8 5 13 4 1 4 3 1 2 7 1 4 2 235 0 12 8 8 4 2 1 2 2 5 9 10 1 10 0 4 0 0 3 0 2 0 255 1 244 133 30 20 22 17 62 22 60 3 1 26 4 24 2 6 0 4 6 0 8 3 265 3 32 30 8 4 15 5 14 13 26 7 52 8 0 1 1 9 7 12 1 1 0 12 2 290 0 25 37 7 0 12 3 15 13 5 6 40 2 1 1 6 0 1 1 7 3 2 2 305 0 17 15 1 0 2 1 8 4 2 2 0 1 1 0 0 0 1 0 1 0 0 310 1 12 22 6 0 2 1 9 4 2 1 14 0 0 0 3 0 1 2 0 4 2 315 5 76 84 17 11 14 8 24 17 10 13 32 8 17 7 2 0 14 8 7 3 2 325 0 24 39 8 5 3 6 20 12 8 11 8 1 2 3 0 0 5 2 2 3 3 TABLE 4 (Continued) JATION 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 339 0 4 11 3 2 1 1 2 0 4 3 26 0 0 0 3 0 2 0 0 0 0 345 0 37 79 11 3 11 8 14 4 13 4 22 10 1 0 5 0 2 1 0 1 1 350 1 19 30 4 3 5 2 13 2 3 2 7 3 6 3 6 4 0 8 3 6 1 352 0 10 18 0 1 2 0 4 1 0 0 3 1 0 0 0 0 0 0 1 2 0 355 0 6 14 6 0 2 1 4 2 3 3 8 0 1 0 1 0 2 1 1 3 1 360 0 47 62 9 7 8 3 24 3 10 14 8 0 2 3 4 2 4 1 0 0 0 365 0 357 419 79 68 55 77 180 52 126 67 405 52 34 38 70 19 33 23 17 30 15 375 1 7 16 4 0 2 0 4 2 4 1 0 0 0 0 0 0 0 0 0 0 0 380 1 16 18 3 1 20 9 5 3 11 1 6 2 1 0 1 0 5 0 . 1 1 0 385 0 7 6 0 0 3 3 4 1 1 1 2 0 0 0 1 0 3 0 0 1 1 390 0 11 10 1 0 1 0 3 6 11 1 3 1 0 0 1 0 4 0 0 0 2 438 0 1 11 0 0 6 0 2 1 2 1 3 0 0 1 0 0 3 3 2 0 0 452 0 6 11 1 1 11 2 3 8 1 0 6 0 1 0 1 0 0 6 3 4 0 475 0 47 38 3 2 13 1 15 2 16 14 16 1 5 0 4 1 1 6 4 13 30 490 0 7 7 1 2 6 0 10 10 0 0 12 1 1 1 1 2 0 6 1 2 3 501 0 8 8 3 0 1 2 5 0 1 0 5 6 0 2 1 1 2 5 4 0 0 510 0 4 13 1 0 4 0 9 3 1 2 6 0 0 2 0 1 1 4 1 1 0 530 0 2 23 0 0 4 1 3 4 5 1 2 0 0 0 0 1 1 1 0 0 1 551 0 22 12 1 1 5 0 5 4 6 2 18 0 2 4 0 3 3 6 3 1 3 552 0 23 20 0 0 1 0 1 0 4 7 8 0 4 1 3 0 0 11 7 3 4 560 0 15 11 2 1 0 3 5 7 1 9 18 0 3 0 6 2 0 2 0 2 0 600 0 5 17 7 2 3 0 2 3 5 0 4 1 1 1 0 0 0 1 0 0 0 615 0 13 31 1 4 13 3 4 5 5 2 10 1 1 3 0 1 3 13 0 17 1 616 0 3 13 2 0 3 0 3 2 1 1 6 0 0 0 1 0 1 2 0 0 2 620 0 13 7 1 4 1 0 4 4 2 1 2 0 1 2 2 0 0 3 1 1 1 625 0 7 13 1 2 6 1 6 1 3 1 6 2 0 2 0 0 0 4 1 2 1 630 0 20 37 2 2 1 3 14 4 1 3 1 0 1 0 0 1 0 1 0 0 2 640 0 17 36 3 2 5 1 13 5 4 2 3 1 0 2 1 2 19 1 0 1 3 645 1 21 33 5 8 11 2 11 5 9 5 16 1 0 0 3 1 4 12 0 4 8 651 5 181 130 27 8 25 6 38 27 35 35 153 3 19 11 12 11 16 18 4 8 167 TABLE 4 (Continued) NATION 1 2 3 4 b 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 652 1 38 46 9 4 20 0 16 8 11 10 74 3 4 3 7 8 4 12 1 3 36 660 0 18 15 2 4 6 0 8 4 4 7 21 5 7 1 2 0 2 1 3 2 25 663 2 69 72 11 2 10 1 16 36 16 10 78 5 9 7 13 3 5 4 0 2 126 666 5 325 82 17 8 35 2 39 45 30 46 199 10 50 9 29 10 8 9 5 30 452 670 1 14 23 2 5 1 3 5 4 5 3 10 0 2 1 1 2 1 1 0 1 4 678 0 10 9 0 0 3 0 4 2 0 1 4 1 1 0 1 0 1 1 2 5 3 681 0 6 16 0 0 5 0 0 4 1 2 9 0 0 0 0 0 1 4 1 0 3 690 0 4 6 2 2 3 3 3 1 5 1 1 0 0 0 3 0 2 4 0 0 0 710 2 76 83 33 29 39 22 38 6 20 38 320 50 4 19 28 12 44 22 15 26 41 712 0 5 5 2 0 0 0 3 0 1 0 6 4 0 1 0 0 3 0 1 1 1 713 0 11 12 0 1 4 1 2 0 1 1 8 2 0 0 0 1 1 0 0 1 3 731 0 17 24 3 3 8 1 7 3 4 11 39 1 3 6 5 2 1 1 0 5 65 732 1 37 42 6 1 5 5 5 6 10 3 13 0 2 2 4 2 6 0 0 5 52 740 0 99 57 15 6 9 8 34 6 9 11 8 4 3 1 1 0 16 3 0 0 3 750 5 64 75 12 1 7 1 29 14 19 11 39 8 6 3 5 3 10 12 5 2 10 770 4 28 48 9 0 9 1 29 4 6 3 13 2 2 1 1 0 5 2 1 0 5 800 0 40 33 2 3 2 4 16 4 8 0 13 2 6 0 1 0 1 0 1 0 3 811 1 60 19 7 0 15 0 Z 13 58 14 56 J 6 3 ! 6 0 4 1 2 8 812 1 36 22 1 1 2 0 1 4 5 1 16 1 5 1 1 0 0 0 3 4. 16 816 1 211 97 32 10 21 3 28 22 28 79 226 21 20 33 6 2 3 4 1 3 74 817 2 174 60 18 3 21 1 22 13 37 24 44 4 12 2 6 8 2 3 1 3 36 820 0 16 23 3 0 7 0 25 2 2 3 3 0 0 1 1 I 2 3 0 0 2 830 0 12 14 0 0 3 1 8 0 2 1 0 0 0 1 0 0 0 2 0 1 0 840 1 39 25 9 4 7 6 17 13 10 1 3 0 2 1 2 0 6 9 0 0 1 850 1 39 24 3 3 14 1 22 1 6 3 6 0 1 3 1 0 3 3 1 0 11 900 1 27 27 10 3 6 20 12 2 4 2 8 1 2 0 0 0 2 2 0 0 4 920 0 18 11 4 0 0 9 2 0 0 0 3 1 0 0 0 0 2 2 0 0 0 00 H 82 Before going on to the detailed technical discussion, a word about our reasons for selecting these two techniques out of the many available is in order. Many "cluster analysis" algorithms exist, each of which attempts to optimize the final derived clusters according to one of the many specific criteria of optimization that exists. One may wish for example to minimize within- cluster heterogeneity and sacrifice the maximization of between-cluster distance, or he may want to find an optimal solution which considers both aspects of the clustering criteria. Besides at least two dozen such algorithms, other techniques such as discriminant analysis are available. Many of them suffer, however, from one or more "fatal flaws." For example, the necessity for deciding a priori how many groups exist in the data or the number of cases to be assigned to each group or the requirement that each group have an equal number of members all seem to be unnecessarily stringent require ments when one is proceeding as much in the dark (i.e., 13 See Ball and Hall, 1968, for an excellent dis cussion of this subject. 83 with little or no theoretical guidance) as is the case here. Further, complications of a practical nature arise because few of the techniques, especially those loosely grouped under the term "cluster analysis" are far beyond the theoretical stage of development. Consequently, only a small number have been tested on social science data and even fewer are available in computer program form. The result is that while many different techniques are theoretically available, very few are actually available. As mentioned above, out of those available two have been chosen: Factor Analysis and H-Group analysis.^ They were selected because (1) neither required a priori judgments of the number of groups in the data or the 15 number of cases to be assigned any group; (2) both have been used in behavioral research so that not only has their utility been established, but problems of inter 14 BMD X72 and BMD 03M Factor Analysis, and Veld- man's H-Group programs were used in the analysis. See Dixon, 1968 and 1969, and Veldman, 1965. 1C . In the case of the H-Group program it is neces sary to specify the number of groups at which you wish it to begin printing out detailed information on the group ings, but this in no way affects the number of clusters the program yields. See Veldman, 1965. 84 pretation have been considerably eased; and (3) both have been programmed and are available in relatively easy to use form. In the case of the H-group, our choice was influenced also by Bruce Russett*s work in which he observed that "of the host of other methods available, the most appropriate seemed to be a hierarchical grouping technique. . . .1 1 (Russett, 1967: 49). The reader may be puzzled by our selection of two techniques. Why not just one? Or three or four? We decided to use more than one measure in order to have some assurance that the results obtained are a reflection of the data analyzed rather than an artifact of the statistical technique used. This is a potentially serious problem in the quantitative analysis of international phenomena that is magnified further here by our lack of long experience with behavioral data of the kind being used in this study. As will be seen below, the underlying principles of the H-group and factor analysis are suffi ciently different to make us reasonably confident that our results are not technique dependent, provided, of course, that convergence is actually obtained. The use of more than two techniques was ruled out because prelim- 85 inary investigation convinced us that the use of more than two methods would be dysfunctional in that the addi tional complexity involved in comparing the results of more than two analyses usually tended to obscure rather than clarify. During the past several years factor analysis has become a widely used technique in the field of inter national relations. As a result, several excellent explanations of the technique and its uses in the analysis of international relations have appeared, obviating the necessity for a long discourse here.^ Instead, we will briefly describe the basic factor analysis model with particular reference to the variations to be used in this analysis. The specific type of factor analysis employed here is based on the product-moment intercorrelation of each variable with every other variable.^ The method See R. J. Rummel, 1967, and Denton, 1965, for two excellent discussions aimed at the nonmathematically trained researcher. For the mathematically trained, see Harman, 1960. 1 7 This type is one of the many that could be used and is called "Principal Component" factor analysis. We will use squared multiple correlation coefficients as ini tial communality estimates and orthogonal rotation through out this chapter. The reader interested in the more 86 then attempts to fit a "factor" to each set of variables that have a high correlation with one another and a low correlation with the other variables in the data matrix. The factor "loadings" then can be interpreted loosely as being the correlation of each variable to each of the artificial factors found in the analysis. After rotation of the axes, each factor resulting from the analysis thus represents a single independent "dimension" that exists in the data."^ In most factor analyses calculated on social science data, the variables are arrayed across the top of the data matrix, i.e., they are the columns, and the cases along the side form the rows. A typical example using WEIS interaction data as the variables and the nations as the cases is illustrated in Figure 1. A factor anal ysis of data organized in this fashion is usually referred to as an "R-analysis," to distinguish it from other possible arrangements of the data. In this example, the factor analysis will attempt to find the underlying 18 The factor analysis also tells us which factors account for how much of the variance in the data, an important item of information when interpreting the results. 87 VARIABLES -----------* (Types of Behavior) I C A S E S (NATIONS) YIELD C O M M E N T C O N S U LT • • • • F O R C E l iisa C A N U S R F R N • n FIG U RE 1 — A F A C T O R ANALYSIS 88 dimensions of behavior and group the variables according to their relationship to each of those dimensions or factors. In some instances, as here, we are interested not only in how the different behaviors cluster together, but also in how nations might group. Factor analysis can be used to accomplish this by transposing the matrix, or turning it 90 degrees. Using the same example, the nations, after transposition, would be in the columns and the variables in the rows, as in Figure 2.^ Factor analyzing such an inverted matrix is usually referred to as a "Q-analysis." It will be this type of factor anal ysis which will be our principal tool for clustering nations according to their observed international behav ior. Hierarchical grouping analysis is a clustering technique originally derived by Ward and Hook (1963). It is much less familiar to those in the field than is factor analysis, and comparatively few readily understand- Transposing the matrix in this fashion sometimes creates other problems such as too many "variables" com pared to the number of cases available. This situation is discussed in detail below in the preliminary analysis section. 89 C A S E S (NATIONS) VARIABLES (Types of Behavior) U S A C A N U S R FRN. . . n YIELD C O M M E N T C O N S U L T F O R C E FIG U RE 2 — A Q -FACTO R ANALYSIS M ATRIX 90 able explanations exist.2® Consequently, the method will be explained in a more detailed fashion than was the case for factor analysis. As mentioned above, the ideal clustering program would maximize the between-group distance while at the same time minimizing the within-group distance. Since such an optimal solution is impossible to obtain given the restraints within which we operate, some compromise ? 1 is necessary. Hierarchical grouping concentrates on minimizing the within-group variation at each step of the analysis, at the expense of the between-group criterion. The implication of this choice will be discussed below after the method is explained with the aid of an example. The data shown below are for four nations which we wish to group on the basis of their scores on each of 20 See Veldman (1965) for a reasonable cogent and understandable discussion of H-Group analysis. We have used elements of both his and Russett‘s (1967) explanation, plus the original Ward and Hook article (1963) to piece together the explanation presented here. 2Although algorithms apparently exist which are capable of such a solution, they require several hours of computer time for even a simple problem. 91 three WEIS behavior variables, as follows: AGREE ACCUSE FORCE USA 5 10 9 ISRAEL 2 15 55 SOVIET UNION 7 17 0 EGYPT 1 12 40 The H-group computation process begins by defining each original case (nation here) as a "group,1 ' so that in this example we have four "groups" of one nation each. The error within each group, therefore, is zero. The first step in the process is to compute a matrix of potential 22 error terms for all possible pairs of nations. This matrix is constructed by summing the squared differences between all pairs of nations in the matrix for each variable, dividing by the number of objects in the poten tial group which at this stage would always be two, and taking the square root of the quotient. Computing all The data are normally transformed into standard scores before this computation is made unless one has a theoretical reason for permitting one variable to count more heavily than another. See further explanation below. 92 possible new groups, i.e., all pairs of nations for the three variables, produces the "potential error matrix" below: SOVIET USA ISRAEL UNION EGYPT USA 0 1075 45 491 ISRAEL 0 1527 235 SOVIET UNION 0 1661 EGYPT 0 It should be obvious from the matrix that the combination of the United States and the Soviet Union would be the first real group formed since the resulting error would be the smallest (we will label it “Group 1"). After this group is formed, a new error matrix would be calculated, with no values at all for the merged nations since they no longer "exist" as separate entities. Instead, the total error resulting from their grouping (45) would be inserted in the diagonal to represent the error present in Group 1 which resulted from the merger of the US and USSR which formed the group. Note also that completely new coefficients for all possible new 93 groups are calculated with reference to the new Group 1 rather than the former groups "USA1 * and "Soviet Union." The new error matrix thus would consider the possibilities at this stage to be a merger of either Israel or Egypt into Group 1 or the creation of a new group composed of the two. In an analysis which was grouping many nations, such as that reported below, the method would continue either to add a country to an existing group or to merge two existing groups, always on the basis of the combina tion that would result in the smallest new error sum. Several important factors relating to the use and interpretation of an H-group analysis should be con sidered before we proceed. First, one should be aware that each variable included in the analysis is given the same "weight"— is considered just as important— as every other variable. It is necessary, therefore, that one carefully consider the variables included in his analysis in order to avoid overemphasizing a particular aspect or dimension of the data and thereby unknowingly biasing the results. To illustrate the point, if one were attempting to group nations on the basis of two types of behavior, conflict and cooperation, he must have an equal number 94 of categories measuring each of the two dimensions or a rationale for emphasizing one over the other, i.e., using more categories for one, if the results are to be properly interpreted. Similarly, the variance or range of each variable must be taken into consideration to prevent over weighting. In an analysis in which the goal is to group all nations in the international system on the basis of variables such as "GNP per capita," "size," and "students per capita" measured in dollars, square miles, and number of students, it is easy to see that the range in the variable "size" is far greater than that for the other two. Should these data be grouped in an unnormalized form using H-grouping, the results would be almost totally due to the one variable with size, the influence of the others all but washed out since the sum of the squared distances of the cases on all variables is the criterion used to form the groups. For this reason, all H-group analyses computed here were done on standardized rather than raw data. Problems resulting from both these over- or under weighting difficulties can be resolved fairly easily when one is formulating his research design. Another problem 95 not so readily dismissed is that a group, once formed, cannot at a later stage be broken apart and its members assigned to other clusters that have been composed in the interim. As a result, it is quite possible in an H-group analysis that a nation could be assigned to one group at an early stage of the analysis and, were it free to change on the same basis of minimized least-square distance, later would be reassigned to a different group. There is no apparent way to circumvent this difficulty, so that the possibility of obtaining in some instances misleading results must be recognized. A last disadvantage of H- group analysis is that it provides no statistical measure upon which one might infer his results to a larger sample or universe. The only way in which one can assess the "goodness" of the final results is by examination of the total error resulting from the formation of all groups at a given stage of analysis, i.e., by summing the error coefficients of all groups at that stage. Whether or not the results at some point are acceptable can be decided only with reference to a specific data set. Veldman suggests, however, that one may use the summary error measure to decide which stage of the analysis yields the 96 "best1 1 set of groupings. He defines the best or "natural" grouping of the data as being at that stage just prior to the largest increase in the error measure (Veldman, 1967: 308-12). In practice this guideline is a useful one but must be interpreted in the context of one's analytic problem. In almost all cases, the largest error increase will occur when going from three to two groups. If one1s theoretical interest is in a larger number of clusters (which sometimes is the case here), there is no reason not to use the largest increase above the three- to-two step as the "natural" grouping point. Preliminary Results The reader will recall that we concluded the second section with a discussion and presentation of the 22 WEIS combevent categories which will be the variable in this analysis. The data on these variables for the 83 actors included in the study were displayed in Table 4. Before proceeding to the actual derivation of the prelim inary taxonomy, however, it is necessary that we become as familiar as possible with these data in order to facilitate our interpretation of the results. To achieve this end, a factor analysis of the data in this 83 by 22 97 matrix was computed which almost perfectly replicates the findings of McClelland and Young (1970) when only 60 actors were included. The results of our analysis are shown in Table 5. The first dimension of behavior was labeled Diplomatic Exchange. It accounted for 75 percent of the total variance and consists of fourteen of the 22 comb- event categories, including YIELD, COMMENT, CONSULT, APPROVE, PROMISE, GRANT, REWARD, AGREE, REQUEST, PROPOSE, REJECT, DENY, WARN, and REDUCE RELATIONSHIP.23 Ten of the fourteen (categories YIELD through PROPOSE, or 1 through 10) can be characterized generally as being cooperative, the next two, REJECT and DENY (categories 11 and 14) contain some elements of conflict but it is defensive in nature. WARN is the most puzzling in the matrix since our intuition tells that it logically belongs on the second factor. It is also the only one of the 22 cate gories above .60 on more than one factor, adding to the difficulty of placing it on one or the other factor. We 23 See Table 3 for a complete list of the 22 combevent categories. 98 TABLE 5 F A C T O R ANALYSIS O F 22 C O M B E V E N T C A TE G O R IES F O R 83 S ELEC TED A C T O R S 1966 - 1969* O R T H O G O N A LL Y VARIABLE U N R O T A T E D F A C T O R S R O T A T E D F A C T O R S VARIABLES COM M UNALITY** I II I I I I II I I I — "" ■ 1 1-YIELD .96 .89 .41 -.01 (.96) .18 .06 2-C O M M E N T .99 .94 .31 -.15 (.94) .25 .19 3-CO NSULT .96 .97 .14 .07 (.87) .46 .03 4-APPRO VE .98 .97 .17 .06 (.89) .44 .04 5-PROMISE .98. .96 .23 .12 (.90) .41 -.04 6-G RANT .95 .96 .17 -.05 (.88) .40 .14 7-R E W A R D .97 .91 .36 .10 (.95) .27 -.06 8-AGREE .92 .95 .06 .11 (.91) .38 .05 9-REQ UEST .96 .94 .22 -.18 (.90) .30 .24 10-PR0P0SE .97 .96 .23 .02 (.91) .04 .05 11-REJECT .85 .90 -.15 -.13 (.65) .58 .29 12-ACCUSE .96 .79 -.58 .00 .31 (.89) .27 13-PROTEST .94 .90 -.28 .25 .55 (.80) -.03 14-DENY .97 .91 .28 -.24 (.92) .22 .28 15-D EM AND .78 .73 -.49 .09 .31 (.81) .16 16-W A R N .95 .97 -.08 .01 (.74) (.61) .15 17-THREAT .84 .87 -.27 -.06 .55 (.68) .26 18-D E M O N S T R A T E .74 .71 -.40 .26 .34 (.80) -.04 19-REDUCE RELATIONSHIP .86 .90 .21 .05 (.85) .36 .02 20-EXPEL .76 .67 -.48 .25 .25 (.83) -.01 21-SEIZE .74 .61 -.58 -.15 .17 (.74) .40 22-FORCE .91 .33 -.37 -.82 .09 .21 (.93) EIGENVALUES: 16.5 2.4 1.1 \% O F TO TAL VARIANCE: 75.0 10.2 5.0 |*Squared m ultiple correlation coefficients were used as communality ! estimates in the diagonal o f the correlation m atrix. Loadings on | the rotated factors equal to or greater than .60 are indicated by | parentheses. j**The communality of each variable is equal to the sum o f the squared factor loadings fo r that variable. 99 decided finally to put it in Factor I for several reasons. First, it has a higher factor loading on Factor X than on Factor II. While this fact alone would not be decisive, the Hierarchical grouping analysis computed as a cross check on the factor analysis assigned it to the group corresponding to Factor I, as one can see by referring to Table 6. Finally, other factor analyses on similar data 9 & sets have all consistently placed WARN on Factor I. In sum, we permitted the empirical evidence in this case to take precedence over our intuitive judgment. Finally, the fourteenth category, REDUCE RELATIONSHIP, is of such a formal diplomatic nature (e.g., break diplomatic rela tions) that upon reflection one is not very surprised that it co-varies with the other categories on this dimension of behavior. The Second factor accounted for 10.7 percent of the total variance and includes seven of the 22 combevents. 24 See Skrein (1970) for 69 nations, Moore (1970) for 86 nations on 40 and 48 months of data, and McClelland and Ancoli (1970) on 152 nations for 48 months of data for further support. 100 TABLE 6 HIERARCHICAL G R O U P IN G ANALYSIS O F 22 C O M B E V E N T C ATEG O R IES F O R 83 SELEC TED A C T O R S 1966 - 1969 With ten groups, error s 11.77 With nine groups, error = 14.11 With eight groups , error = 15.76 With seven groups , error = 19.70 With six groups, error = 19.98 With fiv e groups, error = 21.82 With four groups, error = 30.09 With three groups , error = 40.40 With two groups, error = 79.81 Group Membership: G R O U P 1 G R O U P 2 YIELD C O M M E N T C O N S U LT A P P R O V E P R O M IS E G R A N T R E W A R D A G R E E R E Q U E S T P R O P O S E D E N Y W A R N R E D U C E RELATIONSHIP R E JE C T A C C U S E D E M A N D T H R E A T E N D E M O N S T R A T E E X P E L SEIZE G R O U P 3 F O R C E 101 They are: ACCUSE, PROTEST, DEMAND, THREATEN, DEMONSTRATE, EXPEL, and SEIZE. We have labeled it Nonmilitary Conflict behavior, since all of the behavior types loading on this factor are of a conflictive nature in either a verbal (e.g., ACCUSE, PROTEST) or physical (e.g., EXPEL, SEIZE) sense, although armed fighting is not included. The third and final factor accounted for almost 5 percent of the total variance had only one category loaded highly on it which was FORCE. It has been labeled Military Conflict. It is significant to note that the dimensions delineated in the factor analysis reported here account for over 90 percent of the total variance in the data, an unusually high percentage whenever one is dealing with behavioral science data. This fact, and the continued strong evidence supporting the existence of these three dimensions over varying subsets of WEIS data and two different techniques lends substantial credibility to the contention that these three dimensions, and probably no others, do indeed exist and will play a major role in 25 the preliminary nation groupings presented below. J 25 See footnote 24 for references to other similar results. 102 The next natural step would be to transpose our matrix and compute a Q-factor analysis, as described above. We are prevented from doing so, however, by having only 22 variables. If, as required for a Q-analysis, the matrix were transposed, it would have only 22 cases and 83 variables. Clearly, too few cases would be present to 2 6 compute a legitimate factor analysis. In order to derive our first cut typology, therefore, it was necessary to expand the number of cases in the transposed matrix. This was done by taking four one-year data sets of 22 variables each rather than aggregating all four years. The four one-year matrices were then "stacked" succes sively, one on top of the other, resulting in an 83 column (nations) by 88 row (22 variables times four one-year subsets) matrix, the form of which is illustrated in Figure 3. 26 A Q-Analysis, along with a Hierarchical Grouping analysis, was calculated for these data despite the major violations of the model. The results confirmed our strong doubts: Both the factor analysis and the H-Group were unable to distinguish between the nations (variables). Almost all nations (80 out of 83) loaded on one factor in the factor analysis while below 8 groups, 74 out of 83 clustered together in the H-Group. 103 NATIO NS U S A C A N .N E W YIELD C O M M E N T ! 22 C O M B E V E N T S ' 1966 122. F O R C E 22 C O M B E V E N TS / 1967 22 C O M B E V E N T S / 1968 YIELD C O M M E N T F O R C E YIELD C O M M E N T F O R C E ( 6 7 . YIELD 68. C O M M E N T 22 C O M B E V E N T S 1969 1 |88. F O R C E i FIG URE 3 - M ATRIX CO N FIG UR ATIO N F O R PRELIMINARY Q -FAC TO R ANALYSIS ! 104 Both a factor analysis and a hierarchical grouping analysis were computed for the expanded matrix, with the results of each shown in Tables 7 and 8, respectively. The most "natural" set of groupings, using the criteria described in the third section of this chapter, was used in the H-group analysis to decide the optimum set of groups. For the factor analysis, the arbitrary criterion of using only factors accounting for at least 2 percent of total variance was used to limit the number of factors in order to make the results of the two analyses com parable . That is, in order to combine the output of the two different techniques and thereby maximize the use of both, it was necessary to have the same number of group ings. In this case, seven groups resulted from the application of the above criteria. For the H-group, no difficulty was encountered in assigning the nations to particular groups since this is the specific purpose of the program. Although ambi- buity undoubtedly exists regarding which cluster some nations ought to be assigned, the technique makes the decision for us. Matters are not quite so easy, however, t in the case of the factor analysis. Each nation has a 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 TABLE 7 Q -FAC TO R ANALYSIS O F 83 N ATIO N S F O R F O U R S E TS O F 22 C O M B E V E N T S Y E A R - B Y - Y E A R 1966 - 1969* U N R O T A T E D F A C T O R M ATRIX R O T A T E D F A C T O R M ATRIX UABLE )MM.** I I I I I I IV V VI VII I I I I I I IV V VI VII 98 83 -26 17 19 -30 -23 -05 13 -26 -13 -00 (-89) 13 -27 86 84 -10 17 26 -21 -03 -11 23 -08 -11 -12 (-76) 27 -37 65 51 14 -52 08 -19 13 22 13 02 (-72) -10 -15 12 -26 58 63 15 17 2 1 10 27 06 38 15 -10 -32 (-39) 36 -12 72 2 1 -59 -18 19 23 35 -27 -09 -44 -04 01 -03 (72) 07 74 66 15 19 -34 16 29 -13 20 -11 01 (-74) -25 15 -24 53 60 28 -18 01 -02 10 -22 25 -05 -23 -28 -17 16 (-53) 52 57 -10 13 05 -04 40 1 1 08 06 -18 -39 (-41) 39 -02 74 70 26 1 1 -38 08 12 -11 24 -13 -08 (-68) -28 -03 -36 45 59 -08 25 -11 10 -07 -02 25 -23 06 -33 (-47) 04 -08 62 72 07 06 -13 -21 08 16 05 -11 -15 -39 (-49) 10 -42 35 42 05 -10 -13 -22 18 -26 -12 -09 -14 -27 -20 15 (-42) 93 90 19 12 17 00 -21 -09 5 1 -13 -12 -19 (-63) 06 -45 38 52 03 14 17 23 00 10 (45) -05 -05 -17 -33 18 -02 72 75 16 -02 25 09 -23 08 (57) -10 -23 -06 -49 05 -28 88 88 15 03 1 1 -15 -22 09 36 -15 -21 -16 (-65) 01 -49 56 68 -07 28 -03 -06 -01 -05 18 -14 02 -32 (-60) 12 -18 6 1 72 08 -02 25 -07 13 04 3 1 04 -29 -17 (-48) 30 -28 43 35 28 -36 05 -29 -01 -13 02 04 -38 -01 -09 -02 (-52) 95 88 -09 19 12 -23 -25 -03 25 -21 -13 -14 (-84) 04 -32 60 6 1 -24 -23 -14 -05 12 30 07 -25 (-53) 30 -38 14 07 72 70 04 -26 -31 -02 -04 25 20 -26 (-54) -43 -32 -11 -13 A J / ■« « \ A A 24 / U 79 / J 81 25 86 68 26 81 84 27 69 43 28 85 83 29 54 63 30 81 66 3 1 70 69 32 87 87 33 89 89 34 64 7 1 35 47 56 36 5 1 62 37 41 55 38 49 42 39 30 3 1 40 78 68 41 30 30 42 17 40 43 58 61 44 78 60 45 80 6 1 46 66 58 47 65 53 48 62 67 49 60 48 50 62 . 59 5 1 69 50 52 76 52 53 84 80 54 64 66 55 69 68 56 87 63 57 83 58 58 87 40 59 80 5 1 60 9 1 40 61 76 74 62 6 1 43 63 68 49 64 23 31 65 89 40 66 47 40 67 77 7 1 68 68 33 c.«/ -14 1 vs -22 ^ W T 02 ■ k. 03 17 -53 -14 00 27 -06 09 04 -55 08 -09 42 12 -30 13 03 05 09 19 -15 1 1 17 41 -23 -01 -36 -15 20 25 -08 -13 12 -20 00 -06 14 23 03 -10 26 12 -03 -15 07 23 -11 -13 -01 01 10 -06 -16 -13 -08 -06 23 -22 04 -12 -02 12 -28 06 -06 -23 18 20 -14 -04 -03 -09 0 1 03 -19 07 31 19 -38 33 19 -47 -00 -26 -26 -20 -07 -16 -25 -49 -06 -22 -13 07 10 04 36 -01 24 43 -09 -08 -26 3 1 -02 32 3 1 -13 18 1 1 42 40 13 39 -09 07 16 18 -03 29 -08 -06 34 28 07 -28 16 26 -09 -61 2 1 17 -03 -11 18 20 25 -27 -25 43 -24 -25 02 30 -26 -12 01 32 -12 -01 17 48 -31 03 -59 20 14 08 -04 17 -09 -77 17 01 26 -05 3 1 03 36 -10 39 09 23 -28 -50 23 -12 16 -20 08 -10 12 27 34 07 -04 -11 08 -01 -12 03 45 36 -47 26 06 24 29 24 32 17 11 31 30 -44 -05 -10 26 -03 -53 06 -71 -40 -64 25 27 -17 26 -04 -22 -14 -40 2 1 33 -15 -47 (-49) -39 -03 -21 15 -07 -33 -19 (-61) -56 -14 00 12 4 1 -08 -10 -49 (-60) 10 -13 03 -04 -10 (-63) -16 -02 46 -21 14 37 -18 -22 (-63) -46 -03 -15 -05 (53) -08 -07 -21 -31 0 1 -33 -00 (78) -10 -00 -20 -28 0 1 -26 12 02 -11 -41 (-59) -36 0 1 -21 -01 17 -11 -12 -48 (-70) 18 -24 16 23 -22 (-53) -34 -53 27 -18 03 18 10 -11 -43 (-54) 26 -21 32 10 0 1 -28 -27 (-54) 05 08 15 10 -11 -12 -32 (-61) 02 -02 27 31 -03 -29 -07 (-47) -05 -05 -18 16 15 -19 -38 -01 08 (-50) -04 18 1 1 -27 -09 -05 -06 (-40) 01 04 -55 -09 -37 (-56) 12 06 -05 (43) -06 -18 04 03 03 -28 -01 10 -05 -17 -17 (-25) 04 -20 -10 20 -00 -14 (-59) -19 18 -32 -09 37 -11 05 (-75) -13 02 -21 -05 19 -18 -56 -01 -26 -16 (-57) -29 21 -17 -19 -13 -24 -15 (-67) 18 14 -11 (-67) -08 -21 -14 -32 07 27 1 1 -23 (-42) -37 37 -17 16 (72) 03 -13 -13 -11 09 -09 09 48 -16 -22 (-51) -09 -10 -19 0 1 -00 -08 05 -02 (-66) 48 09 36 (63) 05 -17 -08 -33 40 24 -01 24 -14 06 -52 (-64) 24 -09 12 (59) -05 -06 -39 -34 -02 -13 14 (56) -11 r-26 -11 -36 39 -06 -13 26 (-67) -25 03 -32 44 -06 2 1 45 -22 (-70) 01 -07 20 -19 -24 -05 -55 -05 -01 -25 (71) 09 -11 28 (-80) -17 -19 -10 08 01 -33 1 1 (-87) -03 12 -17 3 1 -05 -12 (60) -26 -17 -31 -21 08 -39 06 (75) -15 -03 -04 -09 -06 -12 -02 0 1 -43 -09 (-68) -15 -00 03 09 (37) 02 -04 -21 -12 -11 -10 21 15 -12. (-84) -00 08 36 -14 -19 05 -08 -09 -06 -22 (61) -15 -11 33 -13 -24 -08 -41 (59) -25 -15 -05 (-75) -16 -31 01 -02 -00 105 TABLE 7 (Continued) U N R O T A T E D F A C T O R M ATRIX R O T A T E D F A C T O R M ATRIX \TION VARIABLE' COMM.** I II I I I IV V VI 69 59 54 -28 -01 -33 18 -26 70 92 85 -19 27 18 -23 -11 71 88 87 26 -18 04 -08 03 72 84 74 40 -08 -01 -07 13 73 88 89 -09 10 06 -18 -18 74 85 65 -20 -43 -17 -30 -22 75 84 7 1 -52 -02 -07 04 -14 76 87 67 -45 -43 -14 -05 -10 77 93 72 -57 1 1 19 -14 -14 78 64 65 32 10 -04 -04 04 79 69 76 10 25 14 -04 04 80 74 78 -02 23 09 -19 -17 81 75 44 42 -16 16 -19 -07 82 66 7 1 -09 13 24 -14 -07 83 76 69 -24 15 -12 -37 -12 EIG ENVALUES 34.5 6.0 4.7 :3.7 :3.3 ;1 . 7 VII I I I I I I IV V VI VII -09 16 (-61) -06 -29 -32 -08 -06 -04 18 -17 -07 -13 (-86) 19 -24 -12 35 -09 -37 -32 -41 14 (-57) -32 28 -00 -16 -39 -27 15 (-69) 08 27 -21 -25 -19 (-78) 05 -24 20 -01 -39 (-67) -08 -44 -12 -18 -19 07 (-65) -10 -14 -55 25 -12 -02 -00 (-64) -51 -12 -36 18 -14 -08 04 -46 -10 04 (-77) 32 -06 -30 27 -02 0 1 -37 -32 08 (-57) -16 30 -03 03 -28 (-58) 22 -37 -08 22 -14 -04 -17 (-73) 05 -32 -54 16 0 1 -02 -01 -12 08 (-84) 22 29 -02 -25 -07 (-70) 14 -07 20 -05 -17 -27 -22 (-78) -06 -05 2.4 % TO TAL VAR IANC E 41.5 7.2 5.7 4.5 3.9 3.3 3.0 *Squared m u ltip le c o r r e la tio n c o e ffic ie n ts were used as com m unality e stim a te s in th e diag on al o f the c o r r e la tio n m a trix . Loadings on th e ro ta te d fa c to rs equal to o r g re a te r than .60 are in d ic a te d by p a re n th e se s. **The com m unality o f each v a ria b le is equal to th e sum o f th e squared fa c to r lo a d in g s f o r th a t v a r ia b le . 106 107 loading on each factor, and by reference on Table 7, one can see that in some cases (e.g., USR) the difference between a nation's loading on two separate factors could be quite insignificant (for the USR, .532 compared to .530) . Nevertheless, we arbitrarily assigned the nations to a specific group (i.e., factor) solely on the basis of their highest factor loading (these are underlined in Table 7). By doing so, ambiguity was tolerated at this interim stage, but mitigated somewhat when the results of the two separate analyses were combined to form a single set of groupings. The necessity for combining to arrive at a "final" set of groups resulted in a systematic com parison of the two sets of results which permitted much of the ambiguity to be removed. The preliminary typology was arrived at as follows. First, since a majority of the nations in each group were the same, individual groups in the two analyses were easily "matched." For example, the first several nations in Factor III (CUB, GME, POL, etc.) were identical to those in H-group's Group 2. Each of the seven factors and groups were matched in a similar fashion and those nations which were assigned to the same group by both 108 TABLE 8 HIERARCHICAL G R O U P IN G ANALYSIS O F 83 N ATIO NS F O R F O U R S E TS O F 22 C O M B E V E N T S YEAR-BY-YEAR/1966-1969 With ten groups, error = 96.6 With nine groups, error = 102.9 With eight groups , erro r = 116.8 With seven groups , error = 124.9 With six groups, erro r = 150.2 With fiv e groups, error = 174.3 With four groups, error = 193.9 With three groups, error = 208.5 With two groups, error = 325.2 LA R G E S T - E R R O R IN C REASE "NATURAL" G R O U P IN G = 7 groups Group 1 Group 2 Group 3 Group 4 Group 5 Group 6 Group 7 U S A S W D C U B E LS C O L V E N G R C NIG 1 C A N N O R G M E LB Y B R A A R G C Y P U A R iM EX D E N P O L LE B B O L P O R A LG JO R ! P E R M O R H U N M O N C H L G UI . T U N ISR ; U N K S U D ALB C H T S W Z G H A T U R K O N N T H IRQ B U L C Z E C O P S A U K O S ! BEL JAP U S R ITA K E N Y E M L A O iFRN TAI S Y R Y U G Z A M K U W V T S ! S P N SIN C H N R U M R H O ; G M W PHI C A M TAZ SAF ! A U S AU L B T N E T H IND | FIN N E W IRN P A K S Y E M A L INS 1 N=24 N=ll N =5 N=13 N=14 N=8 N=8 109 analyses were listed under their respective groups in the composite preliminary typology. A surprisingly large number of nations— 62 out of 83, or 75 percent— were put into the same group, after matching. The remaining 21, however, had to be assigned to one of the two groups into which each was placed by the two techniques. These assignments were made on the basis of the following arbitrary criteria: 1. If the highest factor loading for a nation was .60 or higher, assign it to the group indicated by the factor analysis; 2. If criterion (1) does not apply, but the second highest factor loading is more than .10 lower than the highest, assign the nation to the group indicated by the factor analysis; 3. If neither criterion (1) nor criterion (2) holds, assign the nation to the group indi cated by the H-Group analysis. These criteria, while admittedly somewhat arbitrary, were established on the basis of an explicit rationale. First, because of one of the major problems with Hierarchical Grouping Analysis discussed above, i.e., that a group once formed cannot be broken apart, we tended to rely more on the factor analysis than the H-group. Thus, if the factor analysis was able to assign a nation to a 110 factor in a fairly unambiguous manner, i.e., with a factor loading > .60 or with a reasonably wide spread between the highest and second highest loadings (> .10), we chose to accept the factor analysis findings. If, however, the factor analysis results for a case were quite ambiguous, as when the first two criteria failed, we had no choice but to rely upon the H-group results. The Preliminary Behavioral Typology, thus arrived at, is shown in Table 9. The most easily described cluster of nations is Group 5, which consisted of all the nations in the system that have been actively engaged in a military conflict during a significant portion of the 27 four-year period 1966-1969. None of the other groupings, however, were so readily describable. Group 1 was com posed of 28 nations and might be loosely described as a western community cluster (with some exceptions), while Group 4 might be labeled a generally Communist or Social ist group. It is precisely such familiar labels as these, however, that we wish to avoid here since they have 27 Biafra was excluded from the analysis by the requirement that each entity be a sovereign nation. Ill TABLE 9 NATIO N G R O U P IN G (Preliminary) G R O U P 1 G R O U P 2 G R O U P 3 G R O U P 4 G R O U P 5 G R O U P 6 G R O U P 7 FG-5* FG-7 FG-4 FG-3 FG-2 FG -1 FG-6 H G - .■]** HG-5 HG-4 HG-2 HG-7 HG-6 HG-3 U S A A U L V E N C O L C U B U A R G R C E LS C A N N E W A R G B R A G M E JO R C Y P LE B M E X TAI P O R Y U G P O L ISR A LG M O N P E R F R N G UI TAZ ALB K O N T U R C H T U N K S P N G H A E T H U S R K O S S A U G M W B E L Z A M S Y E S Y R LA O Y E M A U S M O R R H O C Z E C H N V T N K U W FIN S W Z IND B U L C A M NIG C O P S W D ITA P A K C H I S A F N T H N O R R U M M A L H U N S U D D E N B O L INS IRQ JAP LB Y K E N T U N SIN IRN PHI V T S N=28 N=12 N =9 N=10 N=8 N=12 N=4 *FG = Factor Group, as numbered in Table 7. **HG = HGroup, as numbered in Table 8. 112 nothing whatever to do with the respective nations* behavior. Therefore, we went further in our attempt to interpret the derived taxonomy and examined the factor scores for each group on each of the three dimensions of behavior (diplomatic activity, nonmilitary conflict, and military conflict) but these results also failed to provide reasonable insights or adequate descriptions of the classification system. It was at this point that we began to doubt the usefulness of the taxonomy as presently constituted. If, as appeared to be the case, it was difficult to interpret the taxonomy in a qualitative way and quantitative descriptions were equally unrewarding, then one would have to be very skeptical of usefulness for any purpose. It was decided, therefore, that the taxonomy was unacceptable in its current form and alter native approaches to the problem would have to be found. It is this task to which we now turn. 113 References for Chapter III Ball, Geoffrey H. and David S. Hall. "Background Informa tion on Clustering Techniques.1 1 Technical paper. Menlo Park: Stanford Research Institute, 1968. Banks, Arthur S. and Phillip M. Gregg. "Grouping Politi cal Systems: A Q-Factor Analysis of 'A Cross- Polity Survey,'" American Behavioral Scientist, 9:3 (1965), 3-5. Berry, Brian J. L. "Basic Patterns of Economic Develop ment." Atlas of Economic Development. Ed. by Norton Ginsburg. Chicago: University of Chicago Press, 1961. Pp. 110-119. Cattell, Raymond B. "The Dimensions of Culture Patterns of Factorization of National Characters." Journal of Abnormal and Social Psychology, 44:2 (1949), 443-469. ________; H. Breul; and H. P. Hartman. "An Attempt at More Refined Definitions of the Cultural Dimen sions of Syntality in Modern Nations." American Sociological Review, 17:4 (1951), 408-421. Coleman, James S. "The Political Systems of the Develop ing Areas." The Politics of Developing Areas. Ed. by Gabriel A. Almond and James S. Coleman. Princeton University Press, 1960. Pp. 532-576. Cutright, Phillips. "National Political Development: Measurement and Analysis." American Sociological Review, 28:2 (1963), 253-64. Denton, Frank H. "A Handbook of Factor Analysis for International Relations." School of International Relations, University of Southern California, November 1965. (Mimeographed) 114 Dixon, W. J., ed. "BMD Biomedical Computer Programs." University of California Publication in Automatic Computation, No. 2. Berkeley: University of California Press, 1968. ________ . "BMD Biomedical Computer Programs, X-Series Supplement." University of California Publication in Automatic Computation, No. 3. Berkeley: University of California Press, 1969. Feierabend, Ivo K. and Rosalind L. Feierabend. "Aggressive Behaviors within Polities, 1948-1962: A Cross- National Study." Journal of Conflict Resolution, X, No. 3 (September 1966), 249-271. Fitzgibbon, R. H. and Kenneth Johnson. "Measurement of Latin American Political Change." American Political Science Review, 55 (1961), 515-526. Fitzsimmons, Barbara; Gary Hoggard; Charles McClelland; Wayne Martin; and Robert A. Young. "World Event/ Interaction Survey Handbook and Codebook." World Event/Interaction Survey Technical Report #1. School of International Relations, University of Southern California, January 1969. (Mimeographed) Gould, Loyal N. "The ENDC and the Press." SIPRI, Stockholm Papers No. 3. Stockholm: Almqvist & Wiksell, 1969. Harman, Harry H. Modern Factor Analysis. Chicago, 111.: University of Chicago Press, 1960. Lipset, Seymour H. Political Man. Garden City, N.J.: Doubleday & Co., 1960. McClelland, Charles A. "International Interaction Anal ysis: Basic Research and Some Practical Applica tions ." World Event/Interaction Survey Technical Report #2. School of International Relations, University of Southern California, 1968. (Mimeographed) 115 McClelland, Charles A. "International Interaction Analysis in the Predictive Mode." World Event/ Interaction Survey Technical Report #3f School of International Relations, University of Southern California, 1969. (Mimeographed) ________ . "Some Effects on Theory from the International Event Analysis Movement." World Event/Inter action Survey Technical Report #5. School of International Relations, University of Southern California, 1970. (Mimeographed) ________ and Anne Ancoli. "An Interaction Survey of the Middle East." World Event/ Interaction Survey Technical Report #7. School of International Relations, University of Southern California, 1970. (Mimeographed) Merrill, John C. "U.S. Panel Names World's Ten Leading 'Quality Dailies.'" Journalism Quarterly, 41:4 (1964). Moore, James A. "Attention and Interaction in the Inter national System: A Case Study." Unpublished Master's thesis. University of Southern Califor nia, 1970. Phillips, Warren and Dennis Hall. "The Importance of Governmental Structure as a Taxonomic Scheme for Nations." The Dimensionality of Nations Project, Research Report No. 18. Department of Political Science, University of Hawaii, n.d. (Mimeo graphed) Plato. The Republic. Trans, by Benjamin Jowett. New York: The World Publishing Co., 1946. Riggs, Fred W. "The Dialectics of Development." Depart ment of Political Science, University of Hawaii, 1968. (Mimeographed) Rummel, R. J. "Understanding Factor Analysis." Journal of Conflict Resolution, 11:4 (December 1967), 444-480. 116 Rummel, R. J. Dimensions of Nations, forthcoming, 1969. Russett, Bruce M. International Regions and the Inter national System: A Study in Political Ecology. Chicago: Rand McNally and Co., 1967. et al. World Handbook of Political and Social Indicators. New Haven: Yale University Press, 1964. Skrein, S. P. Michael. "Thesis." 1970. Snyder, Richard C.; H. W. Bruck; and Burton Sapin. "The Decision-Making Approach to the Study of Inter national Politics." International Politics and Foreign Policy: A Reader in Research and Theory. By.James N. Rosenau. New York: The Free Press, 1961. Pp. 186-192. Veldman, Donald J. Fortran Programming for the Behavioral Sciences. New York: Holt, Rinehart and Winston, 1967. Ward, J. H., Jr. and Marion E. Hook. "Application of an Hierarchical Grouping Procedure to a Problem of Grouping Profiles." Educational and Psychological Measurement, 23:1 (1963). CHAPTER IV AN INTERNATIONAL BEHAVIOR TAXONOMY: II An Alternative Approach The lack of success in using the 22 combevent categories to derive an acceptable behavioral classifica tion of nations prompted an extensive re-examination of our approach to the problem. Professor Davis Bobrow's paper on "International Indicators" (Bobrow, 1969) was extremely influential at this juncture. In it, the author presents a strong argument for the development of indi cator systems as the most promising approach to research in the field.1 In order to be successful in such an effort Bobrow asserts that it is vital that indexes be developed which are capable of accurately measuring the multifaceted phenomena of international relations (Bobrow, 1969: 9-12). This view was further reinforced by the 1The reader will recall that we dealt extensively with this notion in Chapters I and II. 117 118 recollection that we originally set out to include measures of as many different aspects of behavior as possible. This suggested the introduction of several new variables, some of which were aggregations of the 22 combevents and others which were ratios of one sort or another. During the process of deriving these new data we attempted to move away from the measurement of specific behavioral acts toward indexes or indicators or indicators of different dimensions or aspects of international interaction. A large number of preliminary computer runs were made which tended not only to confirm that composite indicator variables were better able to differentiate between the nations, but that they probably would provide a better means of describing the resulting nation clusters as well. Out of those considered, six indicator variables finally were selected. They were: 1. VOLUME— the total number of acts of any kind initiated. 2. DIPLOMATIC EXCHANGE— consisting of WEIS comb- event categories YIELD, COMMENT, CONSULT, APPROVE, PROMISE, GRANT, REWARD, AGREE, REQUEST, PROPOSE, REJECT, DENY, WARN, and REDUCE RELATIONSHIP. 3. NONMILITARY CONFLICT— consisting of WEIS comb- event categories ACCUSE, PROTEST, DEMAND, THREATEN, DEMONSTRATE, EXPEL and SEIZE. 119 4. MILITARY CONFLICT— consisting of the WEIS combevent category FORCE. 5. BALANCE— a ratio of volume to the number of nations acted to, i.e., the average number of output acts per target, or volume of output number of targets* 6. SCOPE— a ratio of volume to the total number of acts received, i.e., SCOPE = volume of output volume of intake * Variables two, three and four, which were derived from the factor analysis discussed in the fourth section of Chapter III, were chosen principally because they are an empirical aggregation of the 22 WEIS combevents rather than a purely arbitrary intuitive set of category group ings. With the one exception noted in that section— WARN— they did not violate our intuitive insights into the cate gory system and appeared to offer promise as indicators of different kinds of behavior, especially since each represents a different and independent dimension of interaction. Operationally, the scores for each nation on each combevent category included in a particular index were summed to give the individual actors a composite score on that indicator. For example, the United States' scores on categories ACCUSE, PROTEST, DEMAND, THREATEN, DEMONSTRATE, EXPEL and SEIZE were added together and the 120 sum used as an index of its nonmilitary conflict behavior. The other three indicators— VOLUME, SCOPE and BALANCE— are intended to measure a quality of international inter action as distinct from a type of behavior.^ VOLUME is self-explanatory and of course attempts to measure how active each of the 83 nations was relative to the others. This index was included principally because of the earlier empirical work done on the WEIS project which suggested that a simple measure of amount of activity could be an important dimension, especially if one is interested in crisis b e h a v i o r . SCOPE is intended to measure the rela tive spread of a nation's actions across other actors in the system. Many researchers familiar with the WEIS data have noticed, for example, that of the very active nations in the system, some interact with a large number of other polities while others direct their attention to relatively few other entities. The SCOPE indicator is designed to 2 These three indicators were developed originally by Professor McClelland and reported in the WEIS Quarterly Activities Report for March-May, 1969. This is the first attempt to utilize volume and scope in an actual research effort. ■^For an example, see Hoggard, 1969. 121 tap this particular aspect of behavior in the belief that a knowledge of the relative concentration of an actor's behavior may be an important factor to consider in the characterization of individual nation's interactions. William G. Fleming suggests that another important dimension of interstate relations is their balance. This would be the political counterpart of the international balance of payments familiar in economics. It would describe the symmetry of relations between polities— to what extent inter actions were reciprocal or asymmetrical. (Fleming, 1969: 108)4 Here we are trying to distinguish, for example, between the nation that outputs a relatively high number of acts and receives an approximately equal number and the actor that may initiate many acts but to which few other nations direct much attention. Having decided upon six indicator variables, the next step was to construct the data matrix from which the This is only one of .several variables which Fleming asserts are important in the measurement of inter national interactions. Among the others is volume. For other attempts to investigate the concept of symmetry, see Martin, 1970, Moore, 1970, and the graph theoretic work of Harary, 1961. 122 typology would be derived. Because of the necessity for transposing the matrix for a Q-Analysis, the problem of too many "variables" (nations) and not enough cases (indicators) again was encountered. We chose to solve it in the same way as when the preliminary analysis using the 22 combevents as "cases" was calculated. That is, the data were aggregated by separate time periods instead of for the total four years included in the analysis. In the earlier case we used four one-year periods to obtain 88 cases (22 combevents times 4 one-year periods), however, while here we have only six "cases" to begin with instead of 22. It was necessary, therefore, to use even smaller time periods than one year in order to even approach the requisite number of cases. We decided finally to use three-month sets of data since there are sixteen in the 48 months of data. Multiplying 16 separate time cuts by the 6 indi cator variables calculated for each yields a matrix which has 83 columns, i.e., "variables" which here are actually nations, and 96 rows, i.e., "cases" which in this matrix are the 6 variables taken over 16 separate time periods. The form of this matrix is illustrated in Figure 4. 123 N ATIO N S JAN - M A R C H 1966 < APRIL - JU N E 1966 < lO CT. - i i D E C . il 969 U S A 1. V O L U M E 2. DIP. ACTION 3. NON-MIL. CONFLICT 4. MILITARY CO NFLICT 5. S C O P E 6. B A LA N C E V O L U M E 12 B A LA N C E 9 1 V O L U M E 96. B A LA N C E C A N N E W FIG U RE 4 — D A T A M ATRIX F O R SIX INDICATOR VARIABLES O N 83 N ATIO NS F O R SIXTEEN TH R E E -M O N TH TIME PERIODS, 1966 - 1969 124 A Final Taxonomy: Results and Some Interpretations As before, both a Q-factor analysis and a Hier archical grouping analysis were computed. The results of these two analyses are shown in Tables 10 and 11, respectively. The same operational assignment decision- rules used in the preliminary grouping in Chapter III were used to assign nations to a particular group when the two analyses did not agree. The final taxonomy derived from combining the two analyses is presented in Table 12. The analysis divided the 83 entities included into seven groups, containing, respectively, 17, 18, 8, 17, 10, 10, and 3 nations apiece. In comparison to the preliminary one displayed in Table 9, it is somewhat better on a number of bases. r Intuitively, it is probably much more appealing, since the nations are distributed more evenly across the groups 5 and conform more closely to our expectations. 5 The placement of many of these nations still is somewhat puzzling at this point; the reasons for some of the "odd" groupings will be discussed in detail below. TABLE 10 Q -FACTO R ANALYSIS O F 83 N ATIO NS F O R SIX INDICATOR VARIABLES B Y TH R E E -M O N TH PER IO D S 1966 - 1969* U N R O T A T E D F A C T O R M ATRIX R O T A T E D F A C T O R M ATRIX VARIABLE NATIO N COMM.** I I I I I I IV V VI VII I II I I I IV V VI VII 1 70 53 -50 34 19 06 0 1 14 (83) 03 -05 04 -03 06 05 2 89 92 01 -04 -11 -12 15 01 44 34 -29 -11 -35 (47) 38 3 79 85 -03 -23 05 -06 07 -07 37 19 -46 -07 -28 (51) 24 4 74 82 03 -00 07 09 -23 01 46 3 1 (-54) -11 -30 15 13 5 47 35 -10 3 1 -31 14 -06 -34 3 1 17 02 (-57) 09 0 1 08 b 77 70 41 -04 1 1 30 -03 01 15 49 (-62) -09 -02 09 32 7 74 b3 48 16 15 07 04 -22 13 (68) -40 -18 03 23 10 8 64 77 14 03 -05 01 03 15 35 (39) -34 -01 -27 20 36 9 74 83 -01 -07 -11 -03 -15 06 41 25 -42 -12 (-44) 23 26 10 38 59 06 05 00 03 -15 -06 3 1 28 (-34) -15 -22 12 09 1 1 86 75 18 -29 23 27 -22 -14 22 22 (-82) -13 -12 23 09 12 77 76 20 -05 19 2 1 -26 -03 32 37 (-69) -09 -16 09 09 13 9 1 86 -25 21 23 -04 -02 1 1 (79) 28 -31 06 -19 25 11 14 69 79 12 -19 02 13 -02 10 30 25 (-55) -02 -24 25 35 15 76 77 10 -27 -03 -20 2 1 -06 19 26 -33 -02 -30 (61) 3 1 16 94 88 -37 06 12 01 04 05 (78) 10 -33 -03 -21 36 20 17 77 68 03 -14 -15 -41 -30 10 23 24 -26 03 (-72) 23 04 18 75 84 -04 -00 -02 -07 01 20 (40) 29 -33 05 -36 27 34 19 70 7 1 39 05 17 -07 -08 01 20 (60) -45 03 -20 22 1 1 20 63 72 -02 30 13 -06 00 04 (58) 44 -20 -02 -16 19 11 21 87 90 -09 08 -10 -08 -09 -16 (53) 32 -25 -27 -29 45 25 22 85 78 15 -29 05 28 ^ a 18 AA -11 2 1 A A 2 1 a n (-62) or -16 1 A -00 a n 41 nn 42 f /IO\ 22 85 78 15 -29 05 28 18 23 81 83 16 07 -23 04 -09 24 77 83 09 -23 -02 07 -10 25 80 81 -23 22 -05 10 -06 26 80 81 -00 -25 02 -08 13 27 42 58 -04 00 -03 03 -10 28 94 92 -13 01 -08 -10 24 29 75 83 04 -00 01 -03 16 30 58 64 06 -13 30 -10 -18 31 77 73 23 -28 07 24 -15 32 84 86 -07 -11 -20 -08 -19 33 88 74 -39 37 13 08 04 34 76 66 -10 -06 22 -18 16 35 69 7 1 1 1 -15 -03 -32 19 36 66 72 -07 -01 -02 -30 18 37 63 73 1 1 17 -12 -03 19 38 62 63 14 27 -12 -01 -02 39 66 59 16 -30 36 07 2 1 40 69 75 -19 08 -21 16 -11 41 63 64 30 19 2 1 -17 07 42 64 50 50 20 -13 -13 -06 43 66 66 23 01 -29 -23 -20 44 80 65 5 1 3 1 -06 -10 07 45 73 72 3 1 -30 05 09 07 46 87 66 48 34 27 -13 -04 47 74 78 18 -17 13 13 -19 48 69 69 -05 -34 -17 09 2 1 49 59 65 11 -20 -19 08 28 50 64 65 17 -07 -21 26 24 51 79 5 1 27 5 1 -36 16 04 52 50 42 03 -16 -48 2 1 03 53 9 1 85 -22 -23 03 -04 -27 54 73 79 16 -02 06 -22 -17 55 70 79 14 -14 02 13 -13 56 86 81 -29 33 04 09 -03 57 71 79 07 03 15 -16 12 58 87 73 05 04 -42 36 18 59 87 81 -34 26 00 13 -01 60 89 71 -44 40 08 09 -09 61 74 76 3 1 15 10 -04 09 62 81 47 59 46 1 1 -10 05 63 78 42 57 37 04 -22 -03 64 81 74 1 1 27 -17 09 28 65 88 85 -20 17 15 03 06 -11 2 1 2 1 (-62) -16 -00 41 42 16 36 42 -35 -14 -42 09 (42) -05 29 23 (-59) -15 -31 32 26 -15 (67) 23 -27 -33 -20 21 16 -21 29 20 -43 -17 -23 (60) 2 1 -26 30 18 -31 (-32) -18 25 03 -10 54 28 -23 -18 -24 (56) 35 18 (45) 34 -31 07 -24 34 41 15 33 25 (-50) 22 -26 22 00 13 18 24 (-72) 01 -22 13 33 -05 41 19 -41 -24 (-53) 29 22 1 1 (88) 19 -14 -04 -08 14 16 -45 36 23 -26 -22 -04 (67) -10 -06 20 33 -20 01 -34 (58) 22 15 41 27 -10 13 -35 (45) 27 01 38 (44) -15 -13 -18 29 37 33 40 (45) -14 07 -30 -05 37 13 17 2 1 (-52) 28 05 41 25 -08 (53) 13 -31 -35 -29 13 27 -14 25 (62) -24 -03 -08 34 01 26 04 (64) -15 09 -32 -04 29 01 14 41 -22 -17 (-57) 17 19 -02 15 (79) -17 -11 -17 15 25 -10 06 32 (-59) -09 -14 40 30 -01 26 (83) -29 06 -08 15 00 02 28 3 1 (-67) -03 -24 19 17 14 24 00 -37 -02 -24 39 (54) 02 15 18 -28 -11 -15 39 (52) 14 2 1 25 -32 -11 -08 18 (62) -21 26 (59) 02 -53 -07 -07 30 -12 04 -00 -21 -43 -24 1 1 (44) -07 50 04 (-56) -14 -45 34 06 -03 31 (44) -40 -03 -42 31 05 00 30 29 (-59) -11 -26 21 25 -06 (79) 26 -21 -22 -14 17 14 08 (42) 42 -29 10 -22 41 2 1 03 33 21 -28 -38 -14 10 (68) 05 (81) 17 -24 -16 -17 15 23 -02 (89) 15 -14 -19 -13 08 04 14 32 (61) -32 08 -16 22 30 -08 10 (88) -10 -07 00 05 06 -31 -01 (82) -07 -23 -08 17 -11 26 46 46 -09 -02 -13 13 (58) 25 (74) 27 -30 12 -19 2 1 29 TABLE 10 (Continued) U N R O T A T E D F A C T O R M ATRIX R O T A T E D F A C T O R M ATRIX \TION VARIABLE COMM.** I I I I I I IV V VI VII I I I I I I IV V VI VII 66 59 59 34 06 -01 28 -18 07 18 43 (-51) -13 -12 -09 26 67 86 66 -01 -17 -34 -20 -44 19 23 12 -34 -09 (-79) 03 18 68 81 77 -24 29 03 1 1 -11 -22 (70) 25 -27 -35 -13 18 03 69 84 85 -25 18 06 04 -06 -07 (72) 23 -31 -20 -21 26 13 70 81 74 -24 -09 -28 -23 -24 03 44 05 -23 -18 (-66) 25 16 71 80 83 -24 01 05 01 10 20 (65) 14 -30 07 -24 3 1 34 72 58 68 -14 -10 10 01 10 25 (46) 10 -33 18 -20 28 32 73 75 77 -21 -12 -18 14 16 -20 (44) 03 -36 -37 -21 38 3 1 74 76 77 -17 -28 08 -21 -03 -08 38 07 -41 -01 -38 (54) 08 75 78 76 -19 -23 -14 01 19 -24 36 02 -32 -30 -21 (57) 30 76 83 74 -46 24 14 04 03 05 (84) 08 -18 -05 -12 23 11 77 79 72 -48 19 04 03 -03 01 (80) 02 -17 -14 -20 21 12 78 66 76 -21 -05 -03 00 05 -17 (49) 1 1 -31 -24 -21 42 18 79 83 68 -18 -28 43 07 -23 -12 44 0 1 (-69) 04 -13 34 -15 80 82 85 -16 -23 07 -01 10 -02 46 10 -46 -04 -24 (50) 27 81 68 62 -11 -19 23 -31 28 -08 32 16 -19 14 -14 (68) 08 82 86 79 02 -40 -05 24 03 08 25 04 (-65) -06 -22 32 47 83 76 56 -17 -13 -38 -50 -07 -02 22 06 05 -13 (-71) 41 14 EIG ENVALUES 44.5 5.1 3.9 2.6 2.3 1.9 1.8 EIG ENVALUES 44.5 5.1 3.9 2.6 2.3 1.9 I.a % TO TAL VARIANCE 53.6 6.4 4.7 3.1 2.7 2.3 2.2 ♦Squared m u ltip le c o r r e la tio n c o e ffic ie n ts were used as com m unality e stim a te s in the d iag o n a l o f the c o r r e la tio n m a trix . Loadings on the ro ta te d fa c to rs equal to o r g re a te r than .60 are in d ic a te d by p a re n th e se s. **The com m unality o f each v a ria b le is equal to th e sum o f the squared fa c to r lo a d in g s f o r th a t v a ria b le . 127 TABLE 11 HIERARCHICAL G R O U P IN G ANALYSIS O F SIX INDICATOR VARIABLES B Y TH R E E -M O N TH PERIODS/1966-1969 With ten groups, error ~ 88.9 With nine groups, erro r = 89.5 With eight groups, error = 92.9 With seven groups, erro r = 113.0 With six groups, error = 144.2 With fiv e groups, error = 150.4 With four groups, error = 155.8 With three groups, erro r = 186.8 With two groups, error = 226.4 LA R G E S T -ERROR IN C R E A S E "NATURAL" G R O U P IN G = 7 groups Group 1 Group 2 Group 3 Group 4 Group 5 Group 6 Group 7 U S A C A N M O R M E X B U L E LS P E R C O P B R A E T H C Z E C U B S Y R C O L G H A ALB B O L K E N S W Z R H O U S R U N K C H N V E N Z A M LB Y G M W A LG R U M Y E M NIG BEL IND C H L S A F S U D A U S T U N TAZ S Y E U A R F R N P A K A R G T U R G R C LE B IR N JO R S P N TAI N T H IRQ C Y P S A U C H T ISR G M E C A M P O R M O N G UI K U W JAP K O N ITA LA O P O L AU L N E W K O S Y U G M A L H U N V T N FIN SIN V T S S W D PHI N O R INS D E N N=l 1 N=25 N=17 N=4 N=14 N =8 N =4 128 TABLE 12 C O M B IN E D G R O U P IN G O F 83 N ATIO NS - 7 G R O U P S Group 1 Group 2 Group 3 Group 4 Group 5 Group 6 Group 7 FG-8* FG-6 FG-5 FG-3 FG-7 FG-2 FG-4 HG-1** HG-2 • HG-6 HG-3 HG-5 HG-7 HG-4 C H N B E L B R A A R G A LG C O P A LB C Z E C A M C H T A U L A U S E T H E LS F R N C A N IRN B U L B O L G U I LB Y G M W C U B JAP C H L G R C K E N IND D E N N E W C O L K U W P O R ISR FIN R U M C Y P LEB R H O JO R G M E S W Z G H A M O R S A U K O N INS TAZ H U N P E R S Y E K O S ITA IR Q S U D V E N NIG L A O M E X T U N Y E M P A K M A L M O N U A R N O R N T H U N K PHI P O L U S A S P N S A F U S R S W D SIN V T N S Y R T U R V T S TAI Z A M Y U G N=17 N=18 N=8 N=17 N=10 N=10 N =3 *FG=Factor Group, as numbered in Table 10. **HG=H Group, as numbered in Table 11. 129 Group 1 is a good example. In Table 9, the group that included most of these 17 nations also included 11 or 12 others. Somehow we feel that a group that includes the United States, the Soviet Union and most of the other "major powers" in the world should not include nations such as Bolivia, Libya and Iran, as it did in the Prelim inary Taxonomy. Their behavior could be similar, but somehow it seems highly unlikely. Besides these intuitive feelings, there also are quantitative reasons for assert ing that this grouping is better than that previously derived. As one can see in Tables 10 and 11, as compared to Tables 7 and 8, both the Factor Analysis and the Hierarchical Grouping Analysis results were improved in the second effort. In the Preliminary Grouping, the Factor Analysis was able to account for only 69 percent of the total variance while here it accounted for almost 75 percent. In the case of the H-Group Analysis, the error for seven groups was 124.9 earlier and only 113.0 here. Only the number of nations for which the two analyses agreed on group assignment went down (from 75 to 69 percent) when the two taxonomies are compared. This was caused almost solely, however, by the H-Group 130 assignment of seven nations to Group 2 which eventually were put in Group 1 since each had high factor loadings on that factor. By tracing through the H-Group Analysis, it immediately becomes evident that this was caused by the inability of the H-Group to separate these nations from the others they had been clustered with early in the process even though they later probably were more similar g (i.e., would have reduced the error) to those in Group 1. Given that the taxonomy displayed in Table 12 is acceptable according to the criteria used, the next subject to be taken up is why did the 83 nations cluster in this manner? In order to answer this question, we must turn to an examination of the data used to derive the taxonomy. This also will permit us to characterize each of the seven groups, something we have avoided doing up to this point since the basis upon which it could have been done earlier would have necessitated that we resort to labels such as "western nations," "small nations," "underdeveloped nations," and so on— all characterizations See discussion of this problem in the third section, Chapter III. 131 A that nothing whatever to do with a nation's or group's 7 international behavior. In order to simplify the task at hand, the six indicator variables were aggregated for the total 48-month time period (it will be remembered that the taxonomy was derived using the six variables measured by 3-month time periods). Since means and standard deviations will be used as the major descriptors, such a simplification could not in any way alter our conclusions. Further, it was necessary to make the six variables comparable so that direct comparisons could be made, since four were composed of frequency scGres and two were ratios. This was 7 We wish to remind the reader at this point that the interpretations below are based upon the assumption that the data used in the analysis somehow reflect "reality." We recognize the biases in the data, as pointed out in detail in the second section of Chapter III. Since it would be dysfunctional to constantly modify our inter pretations with these reminders, we ask that the points made in that section be kept firmly in mind by the reader when this section is read. We will come back to this point again in the last part of this chapter because we do not feel it can be overemphasized. 8 The raw data for the 83 nations on the six indi cator variables is provided in Table 41 of Appendix B. 132 achieved by ranking the nations on each of the indicators g and replacing the raw data in the matrix with rank scores. This was accomplished in the following manner. For the first five index variables (VOLUME, DIPLOMATIC EXCHANGE, NONMILITARY CONFLICT, MILITARY CONFLICT, and SCOPE), the nation with the highest value was assigned a rank of one, the one with the second highest a rank two, and so on. In the case of BALANCE, however, logic dictated that this procedure be reversed. It seemed appropriate to assign the nation with the ratio closest to one a rank value of one, the actor with the ratio next closest to one a rank value of two, etc., since a ratio on exactly one would indicate that the nation's international interactions were perfectly symmetrical— it initiated the same number of acts as it received.^ These rank data for all 83 9 The data were first converted to standard (z) scores, but because of the large number of nations in the middle of the distributions, the effect was to remove almost all of the differences between them except for the large countries such as the United States and the Soviet Union. ■^Initially, the ranks on the SCOPE variable also were assigned in this latter manner, since it seemed logical that here too the value closest to one would indi cate that the nation had dispersed its acts in the widest 133 11 actors are displayed in Table 42 of Appendix B. The next step was to construct a matrix for each group which contained the rank data and compute mean rank values for each group on each of the six indicator variables. These summary descriptions are presented in Table 13 and illus trated in Figure 5. Before proceeding to the interpretation of these summary group descriptions, some insights can be gained from the raw volume figures summed for each group which are as follows: possible fashion. Upon closer examination of the data, however, it became apparent that although this might be the case theoretically, in reality it was obvious that the nations with the highest ratios, i.e., the largest number of nations in the system. Therefore, we assigned the ranks on the highest to lowest basis instead of reversing it as was done with BALANCE. It is clear that considerably more effort should be expended in the con struction of a scope index. 11 Tables 43 and 44 in Appendix B contain both the raw and ranked data for each nation displayed by group rather than for all 83 actors. It is felt that the pro vision of these data will permit the reader to draw his own— possibly different— conclusions about the groups than those presented by the author below. TABLE 13 M E A N G R O U P R A N K S B Y INDICATOR VARIABLE 134 G R O U P V O L U M E DIPLOMATIC E X C H A N G E NON-MILIT. CONFLICT MILITARY CONFLICT S C O P E B A LA N C E 1 10.6 11.4 12.4 15.0 10.4 28.9 2 35.4 34.5 40.4 43.9 42.5 55.4 3. 49.0 46.8 57.6 54.1 57.5 42.0 4 53.1 53.0 49.7 47.9 50.2 53.1 5 56.1 54.9 54.3 57.1 58.4 41.4 6 60.8 62.6 53.9 47.5 61.4 62,6 7 68.2 73.5 53.2 49.7 47.3 31.0 I I 135 Group 1 ■ 0 Group 2 Group 3 Group 4 . Group 5T Group 6 ^ Group V O L U M E DIPLOMATIC NON-MIL. MILITARY S C O P E B A LA N C E E X C H A N G E CONFLICT CONFLICT FIG URE 5*—M E A N G R O U P R A N K S B Y INDICATOR VARIABLE *Note that we have departed from the general rules covering the use o f graphs such as th is by (a) reversing the ordinate so that the zero point on the scale is at the top instead of the bottom, and (b) connecting the points in the graph with lines even though the scale along the absaisa is discrete and not continuous. This was done fo r the purpose of p ic to ria lly displaying the contents of Table 13 fo r the convenience of the reader. Since neither case caused the data to be distorted, w e f e lt ju s tifie d in taking these lib e rtie s . 136 Total Volume of Interaction Group 1966-1969 % of Total 1 15,856 71% 2 2,782 12 3 880 4 4 1,403 6 5 753 3 6 617 3 7 144 1 TOTAL 22,435 100% The high percentage of the total acts accounted for by Group 1 (71 percent) and by Groups 1 and 2 com bined (83 percent), will be seen below to be one of the distinguishing characteristics of these nation clusters. At the same time, one can observe that the highly skewed nature of the distribution may be another reason to exercise caution in the interpretation of the results. According to the figures in Table 13, Group 1 nations appear to have characteristics quite different from all of the others. The group as a whole has an * extremely high mean rank on all but one of the variables, 137 and even on this (Variable Six— BALANCE), it is still the highest although by only a very small margin. We can describe this group, then, as being one which is extremely active, engaging in a great deal of diplomatic exchange, nonmilitary conflict and, relative to the other groups, military conflict. Group 1 countries also typically spread their attention across many nations and have a tendency to initiate about as many acts as they receive. The group includes almost all of the nations generally referred to as "super" or "major" powers (e.g., United States, Soviet Union, West Germany, France, the United Kingdom) but contains also a large number of nations that usually are not thought of in this context. These are the nations (almost all of them are included in the group) that have been, or still are, involved in a serious con flict or flare-up during the four-year time period including Communist China, Czechoslovakia, India, Israel, Jordan, North Korea, South Korea, Nigeria, Pakistan, United Arab Republic, North Vietnam and South Vietnam. Upon reflection, it should not surprise us particularly that these subsets of nations were clustered together since the two largest members of the "major power" sub- 138 group— the United States and the Soviet Union— both were involved actively in serious conflicts during the four- year period. The most accurate way of describing this group of nations in terms of their behavior would be to label it as a "highly active, conflictful, and wide- lO ranging" group. A good Tcey word might be "embroiled,l,x* since, relative to the rest of the nations, all of those in Group 1 are deeply involved in international affairs on a global level, deeply involved in a conflict, or goth. Group 2 nations have roughly the same profile (see Figure 5) as Group 1, except on a different order of magnitude. Relative to the remaining five groups, how ever, it is quite low on the balance variable (only Group 7 is lower) indicating that these nations typically initiate many more acts than are directed toward them. The group consists of nations from all over the world, many of which are well known for their efforts in the realm of formal international diplomacy, e.g., Belgium, 12 This term first was suggested by Professor McClelland as a descriptor of nations such as Jordan or South Vietnam that are very active in the system only as a result of their involvement in a major conflict. We obviously have expanded the meaning substantially. 139 Canada, Denmark, Norway, and Sweden. Most of the others in the group have been engaged in intensive efforts to persuade others in the system along some policy line. Thus, Cambodia, Laos, and Thailand all have been upset not only about the Vietnam War but about the fighting within their own borders, their future security vis-a-vis China, and so on. The other Southeast Asian countries in the group also have been active in attempting to justify their policy on the war and on the future of the region. Belgium, Cuba, Italy, Spain, and Syria all are at least moderately active, especially on issues affecting their particular regional subsystem. East Germany and Yugo slavia, too, both have been quite vociferous (usually on opposite sides) on the Czech question and future relations with the West ar._l the Soviet Union. We may conclude that the nations in Group 2 are a moderately high activity group which tend to pay more attention to others than is accorded them. As a whole, nonmilitary conflict behavior is often engaged in as well as some physical conflict, and they attempt to address themselves to many nations rather than focusing on only a few. For convenience, we might label this group "diplomatic activists.1 1 140 Groups 3 and 5 have very similar profiles with the only significant difference being the volume of behavior (3 is much higher than 5). Group 5 nations also have a higher mean rank on nonmilitary conflict behav ior, though the two groups rank lowest on this variable. Otherwise, . they participate in international affairs at only a moderate (Group 2) or moderately low (Group 5) level and restrict most of their behavior to diplomatic exchanges of one sort or another. They engage in less international conflict of any kind than any of the other groups of nations, restrict it to few nations, and ini tiate only a few more acts than they receive. Although there are only eight nations in Group 3, they represent almost every geographic region of the world, while six out of the ten in Group 5 are from the Middle East/North African area. In order to make future reference easier, we will label Group 3 a "nonconflict moderately active" group and Group 5 a "nonconflict low activity" group. The least-variable profile in the entire taxonomy belongs to Group 4. In contrast to those in Groups 3 and 5, these nations rank quite high on both Nonmilitary (Variable Three) and Military (Variable Four) Conflict. 141 It is necessary, however, that we be careful in our interpretation of these results since ranks this far down the list easily might be affected by only a few acts. Reference to Tables 45 and 46 in Appendix B demonstrates that it probably would not be accurate to say that these nations engaged in a lot of military conflict since most had between zero and four such acts recorded during the four-year period (e.g., Argentina had three, Chile one, and Turkey three), with the high for the group being Iraq with eight. On the other hand, many did have significant numbers of Nonmilitary Conflict interactions (Poland had a high of 59), so that characterizing this group as initiators of a relatively large amount of this type of behavior would not be misleading. Group 4 tended not to direct their behavior to a large number of nations and initiated almost twice as many acts as were directed toward them. A short-hand descriptor of this group for ease of reference might be to label it as being a "narrow ly focused, moderately active, nonmilitary conflict" group. Where Group 4 was characterized as having the least-variable mean rank profile in the taxonomy. Group 6 142 has by far the most variable profile. As a group, the nations initiated very few acts, with a small proportion of them classified as Diplomatic Activity. Most were either nonmilitary or military conflict directed to relatively few actors. Curiously, their behavior was quite balanced, indicating that they initiated about the same number of acts as they received. The conclusion is that this particular group acts very little and then only in approximately the amount of acts they have received. When one of these nations does initiate an act, it is quite likely to be of a conflictful nature and directed TO toward one of a very small number of actors. We will label this cluster of nations, therefore, as a "low activity, conflictful, narrowly focused, and balanced" group. The last group, containing only three nations, has as its main distinguishing characteristic its extreme ly low level of activity and high proportion of nonmilitary 13 It would appear that at least in theory, it should be quite easy to predict the future behavior of any of the nations in this group. This conclusion will be quite important to us in our prediction chapter below. 143 and even higher proportion of military conflict. Relative to the., (small.),-number, of acts initiated, their scope of activity was fairly wide and they initiated many more acts than were directed toward them. We will label this small final cluster a "low activity, conflict" group. There are numerous other possible ways to view these seven groups. It is obvious that not all can be pursued here. One way that is probably quite interesting, however, is to examine them from the point of view of their variability. The research question then becomes: How widely do the nations in each of the groups range in their rank scores on each of the six indicator variables? Some insight into the matter can be gained by examination of the standard deviation of each group's distribution of ranks on each variable.^ We also have computed the 14 We have reported the standard deviations of the ranked data here in order to maintain comparability with the data in Table 13. Standard deviations also.were cal culated for the raw data, but no significant changes in the relationships resulted. The formula used in the cal culations was N (x - x)2 (Nie, Bent and Hull, S = / 1968) i=l N 144 "mean standard deviation" of each group as a simple measure of the average variability exhibited by that group over the six indicator variables.^ All of these values are provided in Table 14. Examination of these data indicates that the best ("tightest/ 1 "most similar") group is Group 1, with a mean variability score of 0.6. This means that, on the average, nations in Group 1 varied in rank across the variables little more than one-half rank per nation in the group. Groups 2 and 4 were next, both with values of 1.1; while Group 6 had 1.8; Group 5, 2.0; Group 3, 2.6; and Group 7, a very poor 6.8. These results indicate that most of the nations assigned the groups derived in the analysis (Group 7 being the notable excep tion) were reasonably similar. This would seem to lend even more credence to our contention above that the groups 15 We felt it necessary to control for the number of nations of each group in order to make these averages more comparable. The formula used was where s = staple standard deviation ±=l n = number of variables n N = number of cases, i.e., ---------- nations in the group N 145 TABLE 14 G R O U P R A N K S T A N D A R D DEVIATIONS B Y INDICATOR VARIABLE -- -- -- -- -- -- -- -- -- -- --n P U O W r C T T O N '- M IL . ' M I L I T A R Y " " - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - G R O U P V O L U M E E X C H A N G E . C O N E LLC T CONFLICT S C O P E BALAN C E STD.DEV. N 1 7.6 9.4 9.5 15.3 7.4 12.9 0.6 2 19.2 18.0 23.0 20.6 19.1 19.6 1.1 3 21.4 22.7 18.7 21.0 20.3 21.2 2.6 4 16.9 16.6 19.1 20.3 14.3 24.5 1.1 5 17.7 16.1 23.4 19.2 19.8 25.7 2.0 6 14.3 15.4 13.9 21.2 18.0 27.2 1.8 7 14.5 9.8 22.5 27.2 20.0 28.6 6.8 146 making up the taxonomy are a fairly accurate reflection of the data, especially when conceived as a "first-cut" attempt at building such a classification system. Other Considerations We mentioned above that numerous other aspects of the taxonomy, its implications and possible uses, could be examined and discussed. Some, such as its actual role in a prediction model will be taken up briefly in the last section below and in later chapters. Two other topics are of particular interest, however, and will be considered here. First, some of the dynamic aspects of taxonomy will be discussed, following which we will com pare the groups resulting from our analysis with those derived by others from quite different data. The reader will recall that an earlier chapter contained a lengthy and somewhat detailed discussion of the relationship between prediction, explanation, and, of particular interest here, projection. The question that concerns us is the matter of the relative stability of the taxonomy. Put another way, is one to consider the groups derived in the analysis to be a static picture applicable only to analyses of the international system during the 147 1966-1969 period, or is it possible that the behavior of most nations is stable enough to permit us to assume that our taxonomy will not change significantly for some specified future time period? What we need to know is how project able is the taxonomy? Is it possible to project it three years, five years, ten years, or twenty years? And if it cannot be projected, upon what basis do we arrive at this conclusion? The original intent was to examine the matter in an empirical fashion by dividing the four-year period into two-year periods (1966-1967 and 1968-1969), replicat ing the four-year analysis, and then testing for signifi cant differences between the three resulting taxonomies (66-67, 68-69, and 66-69). We had hoped also to accumu late the data year-by-year and follow the same procedure, thereby computing a taxonomy for 1966, one for 1966 and 1967, one for 1966, 1967, and 1968, and one for 1966- 1969.16 Actually, this would have involved deriving taxonomies only for 1966 alone and for 1966-1968, since the two and four-year cuts already would be on hand. 148 Unfortunately, this turned out not to be feasible because we quite literally ran out of data. With only four years available and the necessity of building a matrix amenable to Q-factor analysis, it was not possible, even using two-month data cuts, to do analyses of data 17 for less than two and one-half years. What probably would be the best way to gain insight into the stability of the taxonomy— an empirical investigation and partial replication— is not, therefore, possible. Thus, we are forced to rely on less direct evidence to arrive at some tentative conclusions. One factor which affects the dynamic quality of the taxonomy is the way in which our matrix was constructed. The necessity for using three- month data cuts prevented an empirical test of the proposi tion (since we did not have enough three-month periods) but at the same time injected a dynamic quality into the taxonomy. Thus, the indicator variables used to cluster the nations were not aggregated for the total four-year 17 Using one month cuts would have given us a sufficient number of rows in the matrix but would have caused too large a number of zero cells. 149 period— which would have yielded a truly static set of relationships— but were permitted to vary as widely as they would every three months. The result is that the taxonomy to some degree reflects the individual nation*s behavioral profile over the four-year period as opposed to for the four years.A reasonably intimate knowledge of the data leads one also to feel that in general nations do not change their behavior radically in a three or four- year period of time. The single major exception to this statement might be some of the Group 1 nations, such as Czechoslovakia, that within a matter of several months become involved in a situation which causes a distinctive change in their behavioral pattern. This case probably is the exception, however, rather than the rule. On the basis of the above discussion, and lacking empirical evidence to the contrary, we feel it reasonable to proceed on the assumption that the taxonomy i£ project- The implication is that while the actual analysis has some dynamic elements, the interpretations of the results and descriptions of the groups are static in nature. This would not seem to invalidate them in any way, but the fact should be recognized nonetheless. 150 19 able for a few (three or four) years out. In operation al terms, this will permit us later in the analysis to assume that if the United States is a "Type One" nation in our taxonomy based on 1966-1969 data, it will be— for predictive purposes— a "Type One" nation in 1972 or 1973. The last subject to be discussed before a final summation is the relationship of the classification of nations derived here to a few of the other taxonomic schemes that exist in the literature which are based on quite different data. The reader will recall that in Chapter II we discussed in considerable detail the differences between field theory and system theory, pointing out that this study was being carried out within the framework of the latter. At that time we also noted that nonbehavioral variables such as size, wealth, or GNP would not enter into our analysis, since we were convinced that they measured factors different from those in which we were interested. The comparisons reported below, therefore, will enable us to obtain some notion about ■ L9But probably not any more. 151 whether or not our taxonomy, derived from behavioral data, is significantly different from classification systems based on nonbehavioral data, as we have assumed to this point. Three sets of nation groupings were selected on the basis of (1) number of groups (as close to seven as possible, (2) time period for Which data were collected (as recent as possible), (3) number of nations included in the taxonomy (as many as possible), and (4) general acceptance of the taxonomy (as well-known and thought of as possible). All of these criteria were chosen to maxi mize the comparability and usefulness of the comparison. The three taxonomies selected were: 1. Russett's Sociocultural groups (Russett, 1967). 2. Banks and Gregg's Political System groups (Banks and Gregg, 1965). 3. Russett's UN Roll Call Voting groups (Russett, 1967) . Making the comparisons proved to be somewhat more difficult than at first was anticipated. It was neces sary, for example, to collapse the seven categories in our taxonomy into five to make direct comparison with the two of the others (Russett Sociocultural and Banks and 152 Gregg) more feasible. This was accomplished by deleting Group 7 altogether and merging Group 5 with Group 3, since they are the two most similar in our taxonomy. For the third comparison (with the UN Roll Call grouping) it was possible to leave our classification system in its original form. We then "matched-up" the groups in the respective classification systems in the following manner. First, nations which did not appear in both taxonomies were deleted so that we had a common N. The largest group in the other classification then was compared to each of our groups and "matched" with the one in which the largest number of nations appeared in both. We then proceeded to the second largest group and compared it with each of our remaining groups, finding the most comparable one using the same criterion, and so on through all of the groups. Thus, our Group 2 was matched with Russett's Western Community group from his Sociocultural Taxonomy because seven of the nations were the same, the largest number of "common" nations in that comparison. This procedure, of course, probably causes the taxonomies to appear more similar than they actually are since no attempt was made to consider the conceptual comparability 153 of the matched groups; the sole basis was the maximization of "match-ups." As one can see in Tables 15, 16, and 17, however, our nation-groups were so different from any of the other three that even this procedure could not force them to appear similar. In the comparison with Russett's Socio cultural classification (see Table 15), one can see that of the 57 nations included in both analyses, only 17, or 30 percent of the total were in the same group. The percentage figure was exactly the same (30 percent) in Table 16, where our Behavioral Taxonomy is arrayed against Banks and Gregg's Political System Groupings, only there 22 out of 73 nations were matched. We would expect that of the three, Russett's grouping based on UN Voting behavior would be the most similar, since it is the only one using behavioral data. Quite the opposite is the case, however. Of the 67 nations in both taxonomies, only 19 of them, or 28 percent, are in corresponding groups, the lowest figure of all. The lack of similarity between the groupings, especially considering the lengths gone to make them as comparable as possible, does not surprise us greatly. 154 TABLE 15 C O M P A R IS O N O F Y O U N G BEH AVIO R AL A N D R U S S E TT S O C IO C U LTU R AL TA X O N O M IE S R U SSETT Y O U N G R U S S E T T Y O U N G R U S S E T T Y O U N G R U S S E TT Y O U N G R U S S E T T Y O U N G Gp.2 Gp. 2 Gp.l Gp.3 Gp.5 Gp.4 Gp.3 Gp.5 Gp.4 Gp. 1 A R G ALG----- A LG A R G B O L P O R C H L C Z E AU L A U S A U L B R A C U B F R N A U S B R A BU L----- B U L C O L P O R G M W B E L----- B E L B O L C Z E C H L M E X S P N IND C A N ------C A N C H T — C H T G M E C O L P E R ISR C Y P C U B IND G R C C Y P PHI JO R D E N---- -D E N INS H U N -----H U N VEN----- V E N K O S FIN------FIN IRN----- IRN IRQ P A K F R N G M E IR Q JAP M E X U A R G M W INS JO R N T H U N K G R C K O S P O L----- P O L U S A ISR LEB----- LEB U S R T U R U S R ITA-----•ITA M A L Y U G JAP M A L M O R -----M O R N E W P A K N E W N O R---- N O R S Y R R U M N T H PHI TAI S W Z S P N TUN----- T U N S W D ----- S W D T U R S W Z S Y R U A R U N K TAI U S A Y U G N =21 N=16 N=17 N=l5 N=8 N=l2 N =7' '"T = 2 " N=4 N=12 Number Number Number Number Number Corres- Corres- Corres- Corres Corres ponding=7 ponding=6 ponding=3 ponding^ ponding^ Key to Russett Groups: Group 1 = Afro-Asia Group 2 = Western Community Group 3 = Semi-developed Latins Group 4 = Latin America Group 5 = Eastern Europe Total N =57 Total Number Corres ponding = 17 Per Cent Corres ponding = 30% 155 TABLE 16 C O M P A R IS O N O F Y O U N G BEHAVIO RAL A N D B A N K S A N D G R E G G POLITICAL S Y S T E M TA X O N O M IE S B & G Y O U N G . B & G Y O U N G B & G Y O U N G *B & G PER- Y O U N G B & G TRA- Y O U N G PO LYAR C . Gb. 2 CENTRIST Gb. 1 ELITIST Gp. 3 'SONALIST Gp.4 DITIONAL Gp.5 A U L A LG C A M A LG ARG--ARG C O P A U S B U L A U S A U L E T H BEL— - B E L C H N — C H N C O P B O L B U L G U I B O L C A M C U B B R A C H L IRN P O R B R A CZE----- C Z E G H A G R C C O L LA O S A U C A N — -CAN E T H F R N IRN C Y P NIG V E N C H L G M E G M W GUI JAP G H A Y E M ----- Y E M C O L C U B H U N IND LEB H U N C Y P ISR INS M O R IRQ—-IR Q D E N — -DEN JOR-----■ J O R P E R M E X FIN —-FIN K O N ----- K O N P A K R U M K O S M O N F R N G M E M O N K O S N T H G M W INS P O L NIG S U D —-SUD LEB P O L G R C P O R PA K S W Z S A F IND R U M TAZ— -TAZ P E R T U R ISR S A U ITA— -IT A S P N T U N —-TUN S Y R JAP L A O U AR — - U A R M A L — -MAL U N K V T S TAI M E X U S A M O R N E W U S R ----- U S R NOR--NOR VTN---- V T N N T H PHI— -PHI Y U G V T S S A F S P N Total N =73 S W D —-SWD Total Number Corres S W Z ponding =22 TU R S Y R Per cent Corresponding=30 U N K U S A TAI Y E M Y U G N=37“ N=18 N=zir N=17 N=l0 N=16 N=7 N=15 N=4 N=7 #Corres- #Corres- #Corres- #Corres- #Corres- ponding=9 pondim r7 ponding=3 ponding=2 ponding=l Young Groups Three and ■iv e combined to form th is group. TABLE 17 C O M P A R IS O N O F Y O U N G BEHAVIO RAL A N D R U S S E T T U N VO TING BEH AVIO R TA X O N O M IE S R U S S . Y O U N G R U S S . Y O U N G R U S S . Y O U N G R U SS. Y O U N G R U SS. Y O U N G R U S S . Y O U N G R U S S . Y O U N G Gp.l Gp.2 Gp.4 Gp. 6 Gp.3 Gp.4 Gp.5 Gp.l Gp.6 Gp. 5 Gp.2 Gp. 3 Gp.7 Gp.7 AU L ALG ARG----- A R G ALB JO R A LG B O L B R A P O R ALB A U S C A M C O P A U L B U L A U S C O P C H T S P N E LS BEL —— BEL E TH ------E T H B R A B U L C U B B O L IRN LB Y C A M G H A CHL-— C H L CZE-— C Z E G R C JAP C A N — — C A N GUI----- G UI COL----- C O L H U N F R N KUW-— K U W N E W C H T C U B IND P O R CYP“ ---- C Y P M O N IND LB Y R U M D E N —— D E N INS V E N C H A P O L ISR LEB----- LE B TAZ E LS IRQ H U N R U M JO R N O R FIN— — FIN M O R IRQ NIG P E R F R N NIG MEX----- M E X P A K S U D G R C INS S U D P A K M O N U A R T U N IRN S Y R P E R N T H U N K ITA - — ITA TAZ PHI P O L U S A JAP M A L T U N TAI T U R U SR — ---- U S R N T H U A R V E N N E W Y E M — Y E M N O R —— N O R PHI S P N Y U G SWD-— S W D 156 TABLE 17 (Continued) R U S S . Gp.l Y O U N G Gp.2 R U S S . Gp.4 Y O U N G Gp. 6 R U S S . Gp. 3 Y O U N G Gp.4 R U S S . Gp.5 Y O U N G Gp. 1 R U S S . G p .6 Y O U N G Gp.5 R U S S . Gp.2 Y O U N G Gp.3 R U S S . Gp.7 Y O U N G Gp.7 T U R U N K U S A S Y R TAI Y U G N=20 N=16 N=17 N =6 N=13 N=14 N=9 N=ll N=4 N=10 N =2 N =7 N=2 N=3 Number Corres-Number Corr- Number Corr- Number Corr- Number Corr- Number Corr- ponding=3 esponding=5 esponding=5 esponding=2 esponding=0 esponding=0 Number Corres- ponding=7 Key to Russett Groups: Total N =67 Total Number Corresponding = 19 Group 1 = Western Community Per cent Corresponding = 28 Group 2 = Brazzaville Africans Group 3 = Latin America Group 4 = Afro-Asians Group 5 = Communists Group 6 = Conservative Arabs Group 7 = Iberia 158 Instead, it appears to offer some support for our conten tion that we are dealing with data quite different from that from which existing taxonomies have been derived. Thus, the groupings formed on the basis of these data may provide a new theoretical perspective on the classifi- 20 cation of nations in the international system. The implications of this conclusion for our analysis will be taken up in the summary to this part of the study. Summary and Conclusions Chapters III and IV have been concerned with the development of a taxonomy of nations on the basis of their international behavior during the four-year period 1966- 1969. It was asserted that such a classification system might be useful not only as an element in the prediction paradigm being developed in this study, but also could 20 It is possible that the lack of correspondence is due to the noncomparability of the time periods from which dates were drawn for the respective analyses. Though the only way to prove or disprove the possibility is to replicate the Russett and Banks and Gregg analyses with recent data, it is doubtful that the results would change enough to significantly alter our conclusions. 159 have a number of ancillary uses in other investigations of international phenomena. The data used in the analysis were a large subset of those collected from the New York Times by the World Event/Interaction Survey at the University of Southern California. Of the 159 actors in the WEIS data, 83, accounting for 87 percent of the 25,760 discrete acts in the collection, were included in our analysis. An attempt was made and repeated several times through the two chapters to point out some of the possible biases and the special nature of the data collection so that the reader will be able to interpret properly our results. After deriving the data for the 83 selected actors across the 22 combevent categories selected as our initial set of variables, the two principal techniques used in the analysis, Factor Analysis and Hierarchical Grouping Analysis, were discussed at some length. The underlying dimensions of the data then were delineated and a prelim inary taxonomy derived. Upon examination it was decided that for a number of reasons the preliminary grouping was unacceptable. After some experimentation, six new variables were 160 selected which were indicators of different dimensions or aspects of behavior rather than measurements of specific types of acts. The final taxonomy then was presented and each of the seven groups characterized both by the six indicator variables and by the internal variability of each. The final section considered two related factors. First, the temporal aspects of the taxonomy were examined, after which it was compared to three other nation-state classification systems which had been derived by other authors from different kinds of data. A number of conclusions were reached during the analysis which, subject to the limitations set forth throughout the discussion, are as follows: 1. It may be a mistake to assume that all nations in the international system are basically the same in terms of their underlying motives, methods of statecraft, etc. 2. There is a great need for many more well-defined nation-state classification systems than now exist that are capable of discriminating between types of nations on a wide variety of bases. 161 3. Although a number of taxonomies of nations exist in the literature, only one other analysis even has considered the possibility of classifying nations on the basis of their international behavior. 4. The data upon which the behavioral taxonomy was derived in our analysis are considered a "special purpose" set of international interaction informa tion which are intended only to record the "control efforts" of nations. They also contain some bias. The taxonomy derived in the analysis must be interpreted, therefore, within the context of these factors. 5. Hierarchical grouping analysis, while it has some difficulties associated with its use, is consid ered a good supplementary technique, especially if one is interested in avoiding the possibility of arriving at results that could be technique- dependent . 6. There are three underlying dimensions of nation state behavior in the four years of WEIS data: Diplomatic Exchange, Nonmilitary Conflict, and 162 Military Conflict. 7. The use of indexes or indicators of different aspects of international interaction permit one to discriminate better between nations on the basis of their external behavior than the use of the 22 WEIS combevent categories. 8. When the 83 nations included in the analysis are clustered on the basis of six aspects of their behavior during the four years 1966-1969, seven groups emerge: a) Group 1 is composed of 17 nations which together account for 71 percent of the total interaction recorded in the four-year period. It is the least-variable of all with a mean rank deviation of only 0.6 and is described as a "highly active, conflictful, and wide- ranging" group. b) The 18 nations in Group 2 account for 12 per cent of the interactions. They varied on the average only 1.1 ranks. The group is described as being made up of "diplomatic activists." c) Group 3 contains only eight nations which 163 together account for 4 percent of the events in our data. It is one of the least similar groups in the taxonomy with an average rank deviation of 2.6, and is described as a "non conflict, moderately active" group. d) Seventeen nations are included in Group 4, accounting for a total of 1,403 events or 6 percent of the total. The countries vary little in their ranks across the variables, averaging only 1.1 deviations. Taken together, they are described as a "narrowly focused, moderately active, nonmilitary conflict" group. e) Group 5 contains 10 nations and accounts for 3 percent of the recorded interactions. It is moderately variable (mean deviation equals 2.0) and is described as a "nonconflict, low activ ity" group. f) Group 6 also includes 10 nations accounting for 3 percent of the recorded interaction. It is slightly less variable than Group 5 with an average variability score of 1.8, and is described as a "low activity, conflictful, 164 narrowly focused, and balanced" group, g) Only three nations were assigned to Group 7, and together they account for 144 acts, or slightly more than one-half of 1 percent of the 22,435 events in the data file. The com bination of the small amount of data for these nations and their extreme lack of similarity (the group had a mean rank deviation score of 6.8) causes one to seriously question whether the cluster should be considered a true group. It is described loosely as a "low activity, conflict" group. 9. It probably is reasonable to "project" the nation classification a maximum of four years into the future. 10. On the basis of the lack of similarity between our taxonomy and three others derived from different types of data, we conclude that event/ interaction data of the kind used here probably provide a valid alternative means of classifying nations in the international system. 165 One matter may be of particular interest to those wishing to attempt applications of our findings in this portion of the study to other efforts. It relates to item number six above which asserts that there are three underlying dimensions of nation-state behavior in the four years of World Event/Interaction Survey data, namely, DIPLOMATIC EXCHANGE, NONMILITARY CONFLICT, and MILITARY CONFLICT. The question which may arise is how applicable is this finding to subsets of WEIS data? Would one expect, for example, to find these same three dimensions if data for only European or Southeast Asian nations were analyzed in a similar manner? We think not. Our belief is that these underlying dimensions of international interaction are strong and relatively clear-cut ones, but they were derived in a system— not subsystem— level analysis. Although the question is an empirical one, we doubt that the basic general characteristics of nation-state behavior necessarily are repeated in specific subsystems. Further, one should understand that these dimensions were derived in a static analysis of international behavior, i.e., the data were aggregated for the four-year period 1966-1969 and the variations in behavior as it varied across nations. 16.6 A dynamic analysis which would focus upon variations across time, e.g., months, might be expected to yield 21 quite different results. Before placing these results in the context of the overall study, we feel it necessary to again emphasize their tentative nature. Not only are the data of a very specific type, but we realize that there are as many different ways to proceed— as many different techniques that might be used and as many different aspects of the data that might be included— as there are researchers in the field. The analysis was conceived originally as a "first-cut," pilot study of a relatively unexplored area of research. We hope, therefore, that others will con sider the taxonomy as a tentative one, subject to further revision and improvement. The major goal of this analysis, as set down in detail in Chapter II and mentioned at the beginning of Chapter III, was to derive a taxonomy of nations which would be based upon different aspects of the international behavior of nations, rather than on nonbehavioral 21 See Chapter VI for a full discussion of this point. 167 attributes such as socioeconomic status, level of economic development or political system. This goal was set in the belief that, if successful, such a classification system could function as a "control1 * or "decision box" in a forecasting system, permitting one to postulate predictive relationships that would be applicable to specific types of nations rather than only to a single country or uniformly to all nations in the international system. It was asserted that such a capability would increase our ability to discover useful predictive rela tionships which, in turn, would make our projections and subsequent forecasts more accurate. Whether or not we are successful in attaining the overall goal of the study— an improved ability to test predictive relationships by making forecasts of international behavior— remains to be seen. It is felt, however, that some measure of success has been gained in deriving a behavioral taxonomy which is capable of discriminating among nations in the inter national system on the basis of their observed foreign policy outputs. The second factor we feel may be important to consider, which was discussed in detail in Chapter II, 168 is the situational context within which nations interact. Where the behavioral classification just derived focuses upon the actors in the system, a situational taxonomy would consider a more system-oriented factor: the varied situations in which nations find themselves as they go about the business of coping with their external environment. 169 References for Chapter IV Bobrow, Davis B. "International Indicators." Paper prepared for delivery at the 1969 Annual Meeting of the American Political Science Association, New York City, September 2-6, 1969. Banks, Arthur S. and Phillip M. Gregg. "Grouping Politi cal Systems: Q-Factor Analysis of 'A Cross- Polity Survey.'1 1 American Behavioral Scientist, 9: 3 (1965), 3-5. f* Fleming, William G. "Sub-Saharan Africa: Case Studies of International Attitudes and Transactions of Ghana and Uganda." Linkage Politics. Ed. by James N. Rosenau. New York: The Free Press, 1969. Pp. 94-125 Harary, Frank. "A Structural Analysis of the Situation in the Middle East." Journal of Conflict Resolu tion, V, No. 2 (June 1961) . Hoggard, Gary D. "Indicators of International Inter action." University of Southern California, July 1969. Martin, Wayne R. "A Study of Limited War: Vietnam and Korea." Unpublished Ph.D. dissertation, University of Southern California, 1970. McClelland, Charles A. WEIS Quarterly Activities Report for March-May 1969. University of Southern California. Moore, James. "Attention and Interaction in the Inter-*"- national System: A Case Study." Unpublished Master's thesis. University of Southern California, 1970. (Mimeographed) 110 Nie, Norman H.; Dale H. Bent; and C. Hadlai Hull. Statistical Package for the Social Sciences. Provisional Users Manual, November 1968. Russett, Bruce M. International Regions and the Inter national System: A Study in Political Ecology. Chicago: Rand McNally and Co., 1967. (Yale University) CHAPTER V AN INTERNATIONAL SITUATION TAXONOMY Theoretical Considerations One of the very early organizing concepts proposed in the field of sociology was put forward and developed by William I. Thomas. As far back as the 1920's Thomas suggested, in a series of books and articles, that one of the keys to the understanding of human behavior is the consideration of the context within which behavior takes place. In his words, "... the behavior reaction can be studied only in connection with the whole context, i.e., the situation as it exists in verifiable, objective terms . . ."■*• Thomas' situational approach offers, according to Edmund H. Volkart, the editor of a volume on his contributions to the field, at least three specific advantages to the social scientist. First, it helps in ■^Thomas and Thomas, 1932, pp. 553-554, as quoted in Volkart, p. 6. (Emphasis added) 171 172 the achievement of the overall goal of science which is "to determine that under certain conditions [i.e., certain situations], certain results will follow in certain pro portions. The second advantage of the situational approach, according to Volkart, is that it provides an opportunity to bring out the variables involved in behavior. That is, by study ing behavioral expressions comparatively in differ ent situations, the various types of behavior determ inants (biological, physiological, and cultural) and their relations can be revealed. (Volkart, 1951: 18) The last benefit perhaps is the most applicable to this study: the situational approach emphasizes the need for controlling for various factors when doing research. This is so particularly in the social sciences where the neces sity for controlling for different sources of variance is not quite as obvious as in the physical sciences. Psychologists, too, have recognized the need to consider the situation in attempts to account for human behavior. In a chapter on "Behavior and Development as 2 Thomas and Thomas, 1932, p. 565, as quoted in Volkart, p. 18. 173 a Function of the Total Situation," Kurt Lewin asserted that behavior and development depend upon the state of the person and his environment, B=F (P,E). In this equation the person (P) and his environment (E) have to be viewed as variables which are mutually dependent upon each other. In other words, to understand or to predict behavior, the person and his environment have to be considered as one constellation of interdependent factors.3 During the 1940*s and 1950's many social scientists began to advocate the "situational approach." Among them was Gordon Allport, who, in an article discussing the study of prejudice, stated that "sometimes one and sometimes another disposition or habit is evoked— according to the situation . . . it grows more and more apparent that it is the varying situational contexts that set off varying action tendencies" (Allport, 1962:381— emphasis in origi nal) . ^ 3 Lewin, in Cartwright, 1951, pp. 239-240 (emphasis in original). ^See, for example, Boguslaw, 1960-1961, pp. 212- 19? Cartwright, 1951, pp. 30-60, 238-304; Carr, 1948, pp. 1-38, 45-61, 90-100? Cottrell, 1950, pp. 705-12? Meadows, 1944-45, pp. 354-64 (includes extensive biblio graphical information)? Thomas, 1931? Thomas and Znaniecki, 1927; Thomas, 1927? Thomas, 1937? Volkart, 1951, pp. 1-34. 174 Almost as soon as situational analysis was introduced in sociology, a very few political scientists realized the possible utility of such an approach in their field and attempted to integrate it into their work. In 1930, for example, G. E. G. Catlin's excellent book, A Study of the Principles of Politics (Catlin, 1930), contained a rather amazingly modern-sounding discourse on the science of politics which included a substantial section discussing the concept of a "situation" with regard to politics and history (Catlin, 1930: 31-39). Harold Lasswell briefly touched upon the subject in the introduction to Politics and Personal Insecurity (Lasswell, 1935: 5, 22-24), but the first extensive treatment of the subject by a political scientist probably was by David Easton (Easton, 1953: 149-218). He felt that political scientists actually had been concerned, at least implic itly, with the effects of the situation on political behavior for many years: It can be said that students in traditional research have quite unwittingly been trying to establish the influence of the situation upon authoritative policy. In effect, they have been suggesting that what we must look at to understand the variables influencing policy is the impact of the situation created by individuals out of their structured relations. Although few in political science can use the notion 175 of situation deliberately and clearly, the ideas underlying it constantly appear. Traditional political science has been preoccupied with the role of institutions in the political process. It has been so because instinctively it has felt that the political situation or pattern of power created by the existence of various institutions has been a factor shaping policy. As a determi nant of policy, the situation by implication has been conceived to be quite distinct from and more significant than the kind of people who acted in the situation. (Easton, 1953: 152) Easton, along with Robert Dahl and others, was among the early proponents of the study of "political behavior."® Among the major tenets of these inquiries was the notion that the behavior of individuals engaged in political activity would vary according to the situational context and the related factor of the specific issue being con tested.® At about this same time (the early- to mid- Fifties), Richard Snyder and his associates produced the first of a series of articles and books expounding the 5 See Easton, 1953 and Dahl, 1961. For more recent discussions of the role of the situation in studies of political behavior, see Davies, 1963, and Greenstein, 1969. 6See the discussion of Rosenau's use of "issue- areas” as organizing concepts in the study of foreign policy behavior below. 176 "decision-making approach" to the study of international politics (Snyder, Bruck and Sapin, 1954). Their theoret ical framework included specific reference to a set of situational variables that we ought to recognize that a systematic frame of reference for the study of international politics will require several typologies, one of which will be concerned with situations as defined by decision makers . . . : 1. Structured versus unstructured situations— 2. Situations having different degrees of requiredness— . . . 3. The cruciality of situations— . . . 4. Kinds of affect with which the situation is evidenced by decision-makers— . . . 5. How the problem is interpreted and how its major fxanctional characteristic is assigned— . . . 6. The time dimension— . . . 7. The degree to which objective factors impose themselves on decision-makers— . . . (Snyder, Bruck and Sapin, 1962: 81. Emphases in the original) One is able to see clearly from the phrase "as defined by o decision-makers," by the frequent reference to decision makers in the listing of different types of situations— indeed from the name "the decision-making approach," that this school of political analysis uses the various per ceptions of the individual, i.e., the decision-maker, to define the situation. 177 These are not necessarily the only criteria by which one might derive a taxonomy of situation, however. Charles Hermann correctly identifies two of the possible approaches, with reference to studies of crisis: Definitions of crisis which identify a specific class of situations can be constructed with reference to either of two approaches which are among those prevalent in the contemporary study of international politics. These two are the systemic and decision making approaches. A set of crisis situations does not automatically emerge once the analyst selects either the decision-making or systemic framework for organizing his research. But the approach helps structure the kind of hypotheses in which crisis can prove to be a significant explanatory variable. . . . (Hermann, 1969: 411) Professor Hermann then proceeds to develop a "situational cube" as a possible framework for the analysis of situa tions but, unfortunately for our purposes here, he chose to do so within the decision-making framework, using as his three dimensions threat, decision time and awareness. Indeed, it probably is reasonable to assert that most of the systematic empirical research in international rela tions on situations has been conducted according to the decision-making to this approach.^ 7 See, for example, Pruitt, 1965, and Snyder et al., 1962. 178 As pointed out above, the decision-making approach classifies situations on the basis of the decision-maker's perceptions. Where Professor Hermann identified the two major approaches to the situational analysis in Inter national Relations but pursued the decision-making option, Professor McClelland has suggested recently that a very promising approach for the international system analyst is 11. . . event analysis of international situations" (McClelland, 1970: 22. Emphasis in original). The idea is that one should be able to construct a taxonomy of international situations which is based upon the recurrent situations in which governments typically become involved as they undertake to carry out their foreign policy objectives. Thus, rather than being decision-maker centered, the analysis becomes interaction centered. Situations are defined not by perceptions but by the subject of the interaction, identified in concrete terms. McClelland feels that these (situations) might be characterized as recur ring interaction modes. A few of these modes are well known and are recognized readily by name: there are prominent negotiation situations and there are crisis situations. Chronic border war- 179 fare may be another distinctive type of inter national situation.8 A third approach, which is neither specifically interaction nor decision-making oriented, has been pro posed recently by James N. Rosenau. Within the framework of his study of national-international linkages, Rosenau has suggested that specific "issue-areas" be controlled for: The conclusion that national societies can be organized as penetrated political systems with respect to some types of issues— or issue-areas— and as national political systems with respect to others is consistent with mounting evidence that the functioning of any type of political system can vary significantly from one issue-area to another. Data descriptive of local, party, legis lative, national, and international systems are converging around the finding that different types of issue-areas elicit different sets of motives on the part of different actors in a political sys tem, that different system members are thus activated in different issue-areas, and that the different interaction patterns which result from these variations produce different degrees of stability and coherence for each of the issue- areas in which systemic processes are operative. (Rosenau, 1966: 71) Although promising in many respects, researchers have encountered some difficulty in operationalizing Rosenau's 8McClelland, 1970, p. 23. See Hermann, 1969, pp. 410-411 for a discussion of crises as one type of international situation. 180 conceptualization so that, at least at this time, one is unable to test empirically its usefulness. A similar problem is encountered in the use of one of the most widely used sets of interaction types of issues. This set has consisted of the abstract "politi cal," "economic," "social," and "cultural" factors with others such as "ideological" sometimes added for good measure. From an operational point of view, the biggest problem with these categories is their almost total lack of both rigorous definition and mutual exclusivity. It is very difficult, for example, to conceive of any impor tant economic interchange between two governments which also does not have political implications. In actuality, all of these categories could and probably should be "politico-economic," "politico-social," and so on. Also familiar to us is the "socioeconomic" or the "socio cultural" category. Besides being so difficult to use in an operational sense, such abstract conceptualizations suffer from what Marion Levy calls the "fallacy of reifi cation," which he defines "as the confusion of what one ordinarily distinguishes as aspects of an object with parts of the object" (Levy, 1969: 98). He goes on to 181 say that we would commit it with regard to the physical universe if we referred to the shape and mass of a table as though those aspects were in no way different from the top or legs of the table. This confusion exists . . . (in the use) of the terms economic, political, social, economy, polity, society, and/or culture. (Levy, 1969: 98-99) Thus, despite their rather widespread use in the field, the classification of international behavior according to such categories probably is not very useful and often Q may turn out to be dysfunctional. Probably the most extensive and systematic development of these categories was done by Walter Corson (Corson, 1969). In his analysis of US-Soviet interaction during the post-World War II period, he coded his data according to "Functional Interest-Are'as in International Relations," which was comprised of the following "Interest-Areas": Geographic, Economic-Technological, Military, Diplomatic, Political-Legal, Cultural, and Ideological. When these codes are considered in tandem with his 33 "action-categories" which themselves "are classi fied as either verbal ("words") or nonverbal ("deeds") and the "relevant capabilities or national attributes" the author asserts that "the coding then indicates what type of capability is being used as a basis of influence or control by the actor" (Corson, 1969: 5). As the author admits, however, it is possible that given actions might involve more than one interest-area (he uses an example of a political-military act). After several attempts to code data using this system we concluded that, depending upon one's willingness to interpret the actions, almost all could be coded in at least two of his categories and often three or four. 182 We have attempted so far in our discussion to make several related points. First, the development of the "situational approach" in the social sciences and its subsequent wide acceptance as an important factor in the study of human behavior was outlined briefly. Next, it was demonstrated that a few political scientists also had recognized at an early date the potential usefulness of such an approach and its subsequent utilization both by those studying political behavior and by Snyder and his associates in their decision-making approach to the study of international relations. We then pointed out that besides Rosenau's exploratory use of "issue-areas" in the study of national-international linkages, there are at least two possible theoretical frameworks within which one might conceptualize international situations: Snyder's decision-making approach and the international system approach advocated by McClelland. Finally, it was asserted that the categories most often used by analysts in classifying international events probably are not adequate. In sum, our aim has been to establish that many social scientists, including a number of political scientists, have asserted that the situational context is 183 an important consideration in the analysis of human behavior at all levels, but the concept has not been as well-developed in the field of international relations as perhaps is possible. This, then, is the general rationale for the derivation of a situational taxonomy in this section of the study. Two other reasons of a more specific nature also should be presented. The first is our conviction, dis cussed at some length in previous chapters, that the more factors we are able to control for— to hold constant— the greater our chances of generating predictive hypotheses and using them to forecast. An ability to analyze the international behavior of actors, or types of actors, in a variety of situations which the analyst is able to specify should significantly increase our capability to manipulate experimentally our data. Thus, it would be possible to begin attempting to answer questions such as "If the Communist Chinese become involved in another crisis with the United States, what pattern or patterns of behavior should we expect?" "Is violence likely?" "Should United States decision-makers pay close attention to Chinese demands and warnings in such a situation?" and 184 so on. One naturally would have to have access to the appropriate data, use the proper techniques, etc., but the point is that only by identifying "crisis" as a par ticular type of international situation and collecting and coding data on that basis can one begin to systemati cally gain the ability to answer such questions on other than an intuitive basis. We have used as an example the situation (crisis) which has received, by a substantial margin, the greatest amount of attention from scholars in the field, but as will be seen below, there are many other situations of a less spectacular nature which, if properly identified and analyzed, could provide new insights into the international behavior of nations. The second specific reason for undertaking a study of international situations is only indirectly related to this research effort. Nevertheless, it is so important in the analysis of international event data that it exercised considerable influence on our choice. This is the question of the "importance" or "weight" of given types of international acts. The criticism has been made often, particularly regarding the coding of the WEIS data, that, for example, every warning is treated 185 exactly like every other warning, i.e., each has a "weight" of "one," when in fact we know that some warnings are much more important than others. Thus, a Communist Chinese warning to the United States not to approach any closer to its border in 1951 somehow must be more signifi cant than the "317th serious warning" to the United States about an intrusion into their coastal waters. Yet, critics of the WEIS Project properly have pointed out that under that Project's coding rules, the two warnings would be exactly equivalent to one another. This matter has been discussed at some length by Professor McClelland both in the past (McClelland, 1968: 33-36), and more recently (McClelland, 1970: 20-22). The essence of the argument is stated succinctly by Dr. McClelland: We object strenuously, however, to several attempts to correct the WEIS approach where "in tensity" scores have been attached directly to types of acts. [For examples, see Richman, 1967; Corson, 1969; and Beattie et al., 1970.] There are two complaints we have. The first is that this overlayering of a scaling on a system for categorizing types of action results in an absurd ity of evaluating relationships between types of international acts so that, in effect, the occur rence of an act called WARNING is worth three occurrences of an act called PROTEST, and so on. The second complaint is that weighting by "intensity" usually has had a vague reference— most often to an evaluation of something like the 186 historical impact of a particular event. We think measures need to be taken to incorporate the fac tor of ”importance" or impact of certain events in historical context but by means other than rather ad hoc judgments of intensity applied across all the codings. Our position is that event data as reported normally will not support these evalua tions. Importance or saliency of events varies according to the situation both with respect to the orientations of the involved actors and to the time and place of occurrence of an act in the sequence of events of a particular international situation. . . . (McClelland, 1970: 21. Emphasis added) Thus, the availability of a situational taxonomy would provide, in addition to the other uses cited above, a means to begin approaching the event scaling problem in what appears to be a much more promising manner than those attempted so far. The object of this section of our study will be, therefore, to derive a preliminary set of international situation categories which may prove more useful in international system research than those currently in use. This will be done in a relatively inductive manner .through the collection of information on a large number of international events from which it will be possible to gain enough insight to permit the derivation of general situational categories. During this process, we will attempt to construct the taxonomy in such a way that 187 insofar as possible, the categories will be (1) a generalized reflection of the actual situations in which nations become involved; (2) mutually exclusive, i.e., minimize the necessity for the "double-coding" of indi vidual items; and (3) exhaustive, at least with regard to the data used in its derivation. After a discussion of the data collection, coding and derivation procedures in the next section, the first-cut situational taxonomy finally arrived at will be presented. Data Collection, Coding and Derivation Procedures In order to derive the taxonomy from actual international events, data were collected for two years, 1955 and 1967 from the front page of the New York Times, in the form Who did What (type of act) to Whom, about What (situational context). The New York Times was chosen as the source for the same reasons discussed in the second section of Chapter III. Briefly, we asserted there that there is substantial agreement among scholars 10Thanks are due Vivian Schwartz for her willing help in this data collection effort. 18 8 and practitioners that the New York Times is the single most comprehensive source of international news available. Further, the possible biases of the newspaper which also were discussed in that section here probably are less important. This is because in the previous study we were interested in the types of acts output by nations and, therefore, we were largely dependent upon the perceptions of the news reporter who described the various specific acts as they took place. In this case, the main objec tive is to record, however, the context of the act— not the act itself— which is a more general kind of coding. We are less likely, therefore, to have a reporter's perceptual bias introduced into the taxonomy. One might expect greater problems to result from our decision to use only situation data from the front page of the newspaper. While on the one hand there is no doubt more agreement across different data sources on which international events are important enough to be put on the front page (thus lessening the possible bias of our source), it is possible that we will not find as wide a variety of situations on only the front page as might be found by culling through the entire newspaper. 189 Since in this pilot study we are interested only in obtaining a reasonable sample of these situations, how ever, and it is felt that an acceptable span of time is covered by the collection, we do not expect major problems to occur. Further, one might contend that in analyzing international behavior, it is necessary to look only at the front page, and to go deeper may be for some purposes a waste of time and resources. Thus, Professor Michael Driver contends that one need collect only front page headlines on the "radical premise . . . that only dramatic events, meriting front page headlines, are of real value in developing a model of international behavior" (Driver, 1969: 2). Dr. Driver cites the work of Deutsch and Merritt (1965) in support of the contention that only the most spectacular events have the capacity to influence public opinion of other nations. One might also add that other studies of international interaction, e.g., Gamson and Modigliani (1966) also have demonstrated that for some purposes, at least, headline data appear to be a sufficient source of information. In sum, it is felt that enough evidence exists to permit us to proceed with the development of a taxonomy on the basis of these 190 front-page data. A rationale also should be provided for our choice of the two years, 1955 and 1967. A pilot study indicated that an average of between three and four international events per day appeared on the front page of the newspaper. On that basis, it was projected that one year would provide in excess of 1,000 data items. We also wished to draw our sample from more than one time frame (e.g., the late sixties) while at the same time obtaining a reasonably long continuous run of data on a daily basis. Further, it was important that years be selected that were as active, at least internationally, as possible, i.e., have a fairly good mix of crises, major conferences, and so on. It also was considered desirable that one year from the early part of the post- World War II period and one from the later part be included. Finally, a less important factor for our pur poses, but one that nevertheless also was considered was the comparability of the years selected with the years included in other data collection efforts.^ When all of This was for the purpose of future comparisons with (1) Rummel data for 1955 (Rummel, 1963); (2) the 191 these factors were considered together, 1955 and 1967 were chosen. With the 1,262 data items from 1955 and 1,345 from 1967 in hand, the same basic procedure was followed to derive the taxonomy as that successfully used in the 12 development of the WEIS category system. That is, initially each datum was considered individually and the best general descriptor of the situation represented by that item was listed, always keeping in mind the require ments set down for the final taxonomy, at this stage principally that the categories be general but at the same time as concrete as possible. After this process was completed the resulting list of descriptors was tabu lated and those with apparently similar meanings or connotations were grouped. This process yielded a rough, first-cut classification system. Chronology of International Events for 1955; (3) WEIS data for 1967 (see Chapter III, section 2 for descrip tion) ; and (4) the Index of Le Monde for 1967, in order to permit conclusions regarding the possible use of other data sources for the collection of interaction data. 12 See McClelland (Access to Berlin) in Singer, McClelland et al., 1965; Martin and Young, 1>966, for detailed descriptions of that procedure. 192 The next step was to begin the process of repeated revision. This was done initially by going through the data in somewhat random fashion in an attempt to code individual items according to the current version of the taxonomy. Subsequently, successive day-by-day runs of items were tested instead of the more haphazard method used previously. During this process more atten tion was paid to both the mutual exclusiveness and exhaustiveness of the preliminary taxonomy. It was felt also that the classical situational contexts of students of international relations, especially WAR, CRISIS, and NEGOTIATION should be included in the taxonomy even though doing so caused some difficulties, particularly with regard to minimizing the necessity for double-coding. Several months of data drawn from both 1955 and 1967 then were coded on a trial basis. During this step we focused on the labels used to describe, and the level of abstrac tion represented by, each of the categories. Thus, the relationships between and definitions of, crises and disputes or wars and negotiations were considered and at least tentative solutions to the problems encountered found. The possibility of deriving subcategories where 193 they appeared appropriate also was investigated at this time. Again, we felt that at least some of the more specific types of situations suggested by experts in the field (e.g., land, sea and air incidents) ought to be retrievable from sets of data coded according to our final taxonomy if at all possible. At this point a trial coding session was held. Three students were given a preliminary version of the taxonomy and asked to code one month of data from each 13 of the two years in our sample. All were advanced graduate students who were familiar enough with the con cept of the situational taxonomy as to make extensive instructions or explanations unnecessary. After their criticisms and suggestions were considered and, wherever possible, implemented, the final version of our prelim inary international situational taxonomy was arrived at, as presented and discussed in the next section. 13 Thanks are due James Moore, Vivian Schwartz and Wayne Martin for their many helpful suggestions at this juncture. 194 A Preliminary Situational Taxonomy The Preliminary International Situation Taxonomy finally derived at is presented in Table 18 and consists of twelve general types of situations into which nations become involved as they attempt to cope with their exter nal environment. Before discussing this classification system in detail, however, two points should be emphasized. First, one should note that the taxonomy is exhaustive only with regard to the data collected and used in its derivation. That is, every international act on the front page of the New York Times during the years, 1955 and 1967, can be assigned to one of the situational categories in the classification system. This is not to say, however, that all situations that have occurred, or for that matter that could occur, can be classified in this system. On the contrary, we feel sure that a large data collection even for a period during the 1960's might well contain situations not included in the taxonomy. A set of international interaction data for the immediate post-World War II period (i.e., late 1940's) almost certainly would have situational contexts for which categories would have to be created. This 195 TABLE 18 A PRELIMINARY SITUATIONAL T A X O N O M Y 1. W A R 2. CRISIS 3. F O R M A L INTERNATIONAL NEGOTIATIONS 3.1 About War 3.2 About C risis 3.3 About International Incidents/Disputes 3.4 About Trade or Monetary Policy 3.5 About Future Diplomatic Relations/Recognition 3.6 About Disarmament (Including Nuclear W eapons and Space) 3.7 About External Dominance 4. INTERNATIONAL INCIDENTS/DISPUTES 4.1 Land 4.2 Sea 4.3 A ir 4.4 Other or Combination 5. INTERNAL INCIDENTS 5.1 Espionage 5.2 Involving Embassies, Consulates, O ffic ia l Representatives 5.3 Private Citizens (Non-Espionage) 6. T R A D E O R M O N E T A R Y POLICY (ECO NO M IC) 7. F O R M A L DIPLOMATIC RELATIO NS 8. D IS A R M A M E N T (INCLUDING N U C LE A R A N D S P A C E ) 9. INTERNATIONAL E X C H A N G E S 9.1 Educational/Cultural 9.2 Technological (Including Nuclear and Space) 10. AID 10.1 Non-Military 10.2 M ilita ry 11. E X TE R N A L D O M IN A N C E 12. G E N E R A L DIPLOMATIC A S S E S S M E N T S 196 shortcoming is not considered excessively serious, however, since our aims were more modest. It will be recalled that we set out only to create a taxonomy that might serve on one level as an aid in the delineation of predic tive hypotheses and forecasts and on the other as a pilot study illustrating that this approach to the classifica tion of international interaction is both useful and feasible. A second factor to be noted is that the taxonomy consists of nominal categories. Thus, we have not attempted to range the situational categories along any dimension and thereby make judgments regarding the relative "importance" of each, the relative level of conflict inherent in each, etc. We are not asserting that any is "more than" or "less than" any other on any basis whatever. This is not to say, however, that one (or any number of) such dimensions or continua could not be constructed. On the contrary, as pointed out above in some detail, we feel strongly that this kind of ordering probably is the most promising approach to the scaling and weighting of international interaction. The point we wish to emphasize is not that situations cannot be 197 arranged along various continua, but simply that we have not done so here as it was beyond the scope of this 4 . j 14 study. With these two points in mind, it is possible now to discuss the situational taxonomy shown in Table 18. As mentioned before, the category system consists of twelve general kinds of situations. The first three categories, WAR, CRISIS, and FORMAL INTERNATIONAL NEGO TIATIONS, all represent different (usually subsequent) phases of situations identified in the remainder of the taxonomy. A crisis, for example, usually is an outgrowth of an INTERNATIONAL INCIDENT but could result from a disagreement over TRADE AND MONETARY POLICY, EXTERNAL DOMINANCE, or a combination of several issues. Similarly, a WAR usually is preceded by a crisis of some kind. INTERNATIONAL NEGOTIATIONS often take place in an attempt to end a major system disturbance, e.g., a war, a crisis, or an arms race, but have of course been used as a means of resolving many different kinds of problems. As one can see in the taxonomy, this fact has been recognized 14 Such an effort currently is underway in the WEIS Project. 198 through the provision of subcategories that permit one to code the specific subject of the negotiation under con sideration. Were it not for the great interest in the international negotiation process, we might have deleted this category altogether and included these items under their appropriate general headings, putting, for example, negotiations about war under category 1, about crisis under category 2, and so on. Our decision not to do so, moreover, means that if one were interested in analyzing all data, including negotiations, relating to wars, it would be necessary to combine the items assigned both categories 1 and 3.1. This drawback is not considered serious, however, since on balance it was felt that an ability to retrieve easily all data relating to negotia tions was more desirable than rigidly including all items relating to a given situation under that heading. Following the successful procedure used on the WEIS Project to code types of international interaction, it was decided to use the source, i.e., the newspaper, as an aid in assigning data to our situational cate 199 gories.^ Thus, the operational rule on the WEIS Project is that insofar as possible, the actual wording of the datum should be used in coding. Thus, if the news items state that the United States "accused" the Soviet Union of accelerating the arms race, it would be coded an "accuse." In a like manner, when an item refers to the Vietnam WAR it is coded under category 1, "WAR." Typical examples from the 1967 data for this category would be: DeGaulle urges US to quit Vietnam War, and US bombers attacked targets in demilitarized zone between North and South Vietnam. Somewhat more difficult to operationally identify, because of the usually shorter duration, is a "CRISIS" situation. It was found, however, that the rule adopted— that the situation be defined by the data source— also worked 15 See WEIS Technical Report #1 for details of this procedure. 16 Unless it refers specifically to negotiations about the war, in which case it would be coded a 3.1. Note also that once the conflict was labeled a "WAR" by the news media, we did not require that the word be in every item. 200 reasonably well in this case. In the crisis that developed previous to the June, 1967 Arab-Israeli War, for example, we labeled the situation a "CRISIS" on May 27 as a result of the following item: The USSR said today it was considering joining other nations in an attempt to end the crisis in the Middle East. As pointed out above, category 3, FORMAL INTER NATIONAL NEGOTIATIONS, is a somewhat special category in that it includes items about most of the other categories. All interactions pertaining to negotiations should be coded here including those that occur before the start of actual formal negotiations, during what might be termed the pre-negotiation stage. Particularly included would be those "procedural" discussions that almost always take place, formally or informally, during which time agendas, condition for the start of formal negotiations, etc., are agreed upon. The SALT talks at Helsinki are a good example as are the following items which preceded the Paris Peace Talks. President Johnson said today he would be glad to meet Hanoi more than halfway in any possible cease-fire or peace conference negotiations. The US asked North Vietnam to clarify through diplomatic or other channels whether it has 201 softened its position on Peace Talks. INTERNATIONAL INCIDENTS/DISPUTES probably con stitute the broadest category in the taxonomy, and include all those international disagreements and incidents that are not codable under any of the other categories in the system. It was decided also to provide the subcategories of "LAND," "SEA," and "AIR," to permit the analysis of these more specific kinds of interactions, since evidence exists which suggests that the behavior of at least some nations in these different situations may differ enough to warrant the separation.^ It should, be clear that a dispute over trade policies would not represent an INTERNATIONAL DISPUTE as we have conceived it here, but the following events are good examples of items which be assigned this category. The US said that Soviet planes had made an unwarranted attack on a US patrol aircraft over international waters in the Bering Straits. Congo asked Algeria to extradite Tshombe to Kinshasa where he faces the death sentence. 17 For example, it has been suggested that the Communist Chinese react quite differently to LAND, SEA, and AIR intrusions. See McClelland etal., 1965. 202 Subcategories also were provided for the next general situation, INTERNAL INCIDENTS. These were intended mainly to permit one to distinguish between incidents related to official buildings and representa tives, quasi-official representatives (i.e., espionage), and private citizens who, one way or another, have managed to create an internal incident while living or traveling in a foreign country. The situations labeled INTERNAL INCIDENTS are distinguished from other contexts princi pally by virtue of the fact that they take place physi cally within the borders of a single nation. A typical example is: The Soviet Union strongly protested the breaking of windows in its embassy in Peiping by a mob of Chinese students. Although economic relations usually are among the most routine of all international interaction, on occasion they can become major international issues. In such cases the context generally is a question either of TRADE or MONETARY POLICY, and thus would be coded under category 6. Recent examples include the controversy which con tinued for several months over international gold flows, and the attempted embargo or Rhodesia. In the earlier 203 part of the post-World War II period, FORMAL DIPLOMATIC RELATIONS often were the subject of sometimes bitter debate between nations, particularly in Europe. The formal diplomatic relations of many nations with the two Germanies, the two Koreas, the two Chinas, and the two Vietnams also have been the subject of sometimes intense international debate. An example of an act which should be coded as occurring within this situation context is: Spain and Rumania signed an agreement today establishing full consular relations. Category 8, DISARMAMENT, which includes dialogues about nuclear weapons and space as well as conventional arms probably will not be a widely used category since in most instances in which the context is disarmament nations are either involved in setting up formal negotia tions or are in the midst of them. The appropriate category thus usually is 3.6, FORMAL NEGOTIATIONS ABOUT DISARMAMENT. It was included as a separate category because from time to time statements are made that are not related in any way to formal negotiations. Note that this is another instance in which it would be necessary to combine two or more categories (in this case, 8.0 and 3.4) if one wished to analyze all data concerned with 204 disarmament. One of the few items which fell under this heading in our two years of data was: India and the Soviet Union jointly called for a ban on nuclear weapons and a substantial reduction in conventional arms. INTERNATIONAL EXCHANGES (category 9) and AID (10) are similar in some waysr but item-by-item examination of the data strongly suggested that the situational context is in fact quite different in most instances, making it relatively easy to differentiate between them. The two- way flow inherent in EXCHANGES was quite apparent and contrasted, sometimes sharply, with the generally one-way flow of acts in the AID category. It was decided also that both of these categories should be subdivided in the former case into EDUCATIONAL/CULTURAL and TECHNOLOGICAL (9.2) EXCHANGES, and in the latter into Nonmilitary (10.1) and Military Aid (10.2) . General examples of these-two categories are: The US said today it would give military aid to Costa Rica if that nation's request was approved by the OAS. A 12-man agricultural delegation from the Soviet Union arrived in the US today to study corn and hog production. Visit was part of an agricul tural exchange agreement between the two nations. 205 EXTERNAL DOMINANCE, category 11, focuses mainly on situations in which one nation's internal processes are controlled by another nation.The common example of this situation was the colonialism of many of the major powers in the early post-war period, although other series of events (e.g., Hungary in 1956 and the Dominican Republic in 1965) also would be coded under this heading, at least until they reached the crisis or war stage. The French difficulties in North Africa in 1955 was the set ting for the following example of a category 11 item: The US expressed concern to France today about reports that US military equipment was being used to fight Moroccan rebels. The last category, GENERAL DIPLOMATIC ASSESSMENTS, was included to permit the coding of those nonspecific statements made by national leaders about the state of the world or a region or sometimes the relations between just two nations. In such cases, the situational context was felt to be simply the international diplomatic rela tions of nations. A typical example is: French President DeGaulle hailed the easing of tensions in Europe. In a final assessment of this preliminary situa tional taxonomy, one can always see ways in which it 206 might be improved, but it was felt that the greatest source of confusion probably would arise center on cate gory 4, INTERNATIONAL INCIDENTS/DISPUTES. One might well assert that such a category could subsume many— perhaps all— of the others. In a purely literal sense, of course, that would be the case. The rule used to code the two years of data gathered for this analysis, however, worked 19 quite well to clarify this potential difficulty. It was simply that if an international dispute or incident was not accounted for by one of the other categories, it would be assigned category 4. While such a procedure in some ways causes this category to become a "catch-all," it nevertheless has the advantage of permitting the easy classification of most international events while at the same time allowing those which for some reason we feel take place within an identifiably different context to be so coded. Experience with larger data sets may well dictate that several new categories be set up and category 4 be eliminated, but lacking that broader knowledge, we feel that the taxonomy, as presented, is capable of ■^Intercoder reliability using this rule-of-thumb was 88 percent, thus supporting our contention. 207 providing significant new insights along the lines out- 20 lined in the first section above. The Situation Data Before proceeding to an examination of this and the nation-state taxonomies from the point of view of prediction and forecasting, a brief quantitative descrip tion of the situation data used to derive the situation classification system will be presented to permit at least some familiarity with it before it is analyzed in a more extensive manner in the chapter which follows. Table 19 contains a listing of the twelve general situa tional categories together with the number of acts that were assigned to each in 1955, 1967, and for the two years combined. Of the 2,607 acts that were recorded, 1,262 or 48 percent occurred in 1955, and 1,345, or 52 percent, occurred in 1967. This nearly even split of the data between the two years sampled probably reflects one of 20 In order to permit as complete as possible familiarity with the taxonomy, approximately 100 coded data items have been provided in Appendix C of this chapter. 208 TABLE 19 TO TAL A C T S O U T P U T B Y M A JO R SITUATION C A T E G O R Y 1955 1967 1955&1967 SITUATION F R E Q U E N C Y R A N K F R E Q U E N C Y R A N K TO TAL FR EQ . W A R 101 4 452 1 553 CRISIS 0 12 33 7 33 F O R M A L INTERNATIONAL NEGOTIATIONS 329 1.5 309 2 638 INTERNATIONAL INCIDENTS/DISPUTES 329 1.5 291 3 620 TR A D E /M O N E TA R Y POLICY 24 9 28 8 52 F O R M A L DIPL. RELATIONS 2 1 10 24 9 45 D IS A R M A M E N T 72 6 22 10 94 INT'L E X C H A N G E S 15 1 1 3 11.5 . 18 AID 50 7 53 6 103 E X T E R N A L D O M IN A N C E 3 1 8 3 11.5 34 GEN.DIPL. A S S E S S M E N T 213 _3 54 J5 267 TO TAL F R E Q U E N C Y 1262 1345 2607 209 the major constraints placed upon the data collection— that only items on front page be recorded— as well as the occurrence during both of these years of several major 2 1 international events or incidents. Thus, in 1955 the conflict between Nationalist and Communist China over the Tachen Islands occupied a significant amount of attention during the early part of the year while the latter months were largely concerned with the Geneva Conference in October, the general lessening of tensions in Europe, and the beginning of the build-up to the 1956 war in the Middle East. The June War in the Middle East was the most spectacular international event of 1967, although the Vietnam War probably was the major issue of continuing concern throughout the year. This assertion appears to be confirmed in Table 19 when one notes the extremely high number of acts in 1967 in which the situational context was WAR (452), especially in comparison to 1955 when there were only 101 sets in the WAR category. 21 It will be recalled that one of the criteria used to choose these two years was that both were rela tively active in terms of international affairs. 210 Other interesting comparisons also can be made regarding the distributions for the two years. Besides the radical difference in the WAR category, other major shifts were recorded. CRISIS went from zero in 1955 to 33 in 1967?22 DISARMAMENT, from 72 in 1955 down to 22 in 1967? EXTERNAL DOMINANCE from 31 in 1955 to 3 in 1967? and GENERAL DIPLOMATIC ASSESSMENTS from 213 in 1955 to only 54 in 1967. These shifts also can be noted in the relative ranks of these particular situational contexts between the two years. The 33 crisis acts recorded in 1967 of course are those related to the crisis which existed prior to the outbreak of the June War in the Middle East. This conflict and the continuation of the Vietnam War throughout the year (which continually got front page treatment) together explain the extremely large number of acts in that category (over 33 percent of the 1345 for the entire year). It is interesting to note that the rest of the significant shifts pointed out 92 Although often referred to as the "Tachens Cri sis," the events recorded during 1955 did not meet the criterion set up to assign acts to that category. In retrospect our feeling is that more attention needs to be paid to the criteria for assignment to this category. 211 above— in DISARMAMENT, EXTERNAL DOMINANCE, and GENERAL DIPLOMATIC ASSESSMENTS— all were downturns, which conform to our intuition that attentiveness to these situations decreased from 1955 to 1967. Since all of the other categories contained very near the same number of acts, we are forced to conclude that at least on the front page in 1967, WAR and to a lesser extent CRISIS, replaced these other categories as major foci of attention. With the exception of the situations already noted, the distribution of acts was reasonably stable across the two years.^ In both cases, FORMAL INTER NATIONAL NEGOTIATIONS and INTERNATIONAL INCIDENTS/DISPUTES each accounted for approximately 300 acts, while INTERNAL incidents had about 75, AID about 50, both TRADE or MONETARY POLICY and FORMAL DIPLOMATIC RELATIONS had 20 to 25, and INTERNATIONAL EXCHANGES had 15 and 3 acts respec tively in 1955 and 1967. These same data can be examined from a somewhat different viewpoint if one ranges the total acts across the different nations in the system as is done in Table 20. 23 The rank order correlation between the two years is rg = .70. 212 TABLE 20 TO TAL N U M B E R O F R E C O R D E D A C TS IN SITUATION D A T A CO LLECTIO N 1955&1967 JATIO N 1955 1967 C O M B IN E D NATIO N 1955 1967 U S A 298 411 709 C O P 0 7 U S R 218 121 339 A LG 3 3 ISR 52 97 149 C A M 0 6 U N K 99 . 44 143 E E C 0 6 C H N 68 64 132 ITA 5 1 U A R 46 67 113 K O N 6 0 U N O 50 63 113 S U D 2 4 F R N 79 33 112 A U S 5 0 V T N 0 86 86 S P N 3 2 G M W 53 18 7 1 S W D 4 1 M L G 26 40 66 IRN 2 2 V T S 8 41 49 IRQ 1 3 V C G 0 43 43 N A T 3 1 IND 28 8 36 P O L 1 3 C H T 26 0 26 S A U 3 1 S Y R 9 13 22 A F G 2 1 Y U G 15 5 2 1 C U B 0 3 JO R 2 16 18 C Y P 2 1 V A T 2 15 17 E C U 1 2 C A N 6 10 16 H U N 1 2 G M E 1 1 4 15 K U W 0 3 R U M 2 12 14 LEB 3 0 G R C 4 9 13 M E X 0 3 JAP 10 3 13 N E W 3 0 K O S 9 3 12 SAF 3 0 AU L 6 4 10 Y E M 1 2 C Z E 4 6 10 BEL 0 2 M O R 9 1 10 E T H 0 2 NIC 10 0 10 FIN 2 0 T U R 3 7 10 INS 2 0 C O S 9 0 9 IRE 2 0 NIG 0 9 9 LA O 2 0 B U L 8 0 8 LAP 2 0 B U R 6 2 8 LBY 0 2 O A S 7 1 8 P A N 2 0 P A K 6 2 8 P A R 2 0 BIA 0 7 7 PHI 1 1 19558.1967 C O M B IN E D rororororororororororocococooocooocotoGocA>co-P».p»-p».p»4itncjiC7i<y>c7>cy»c7>c7>CT>''j TABLE 20 (Continued) NATIO N 1955 1967 P L O 0 2 TAI 0 2 A R G 1 0 B O L 0 1 B R A 0 1 E FT 0 1 G U A ICE 1 0 1 H O K 0 1 K E N 0 1 LB R 0 1 P O R 1 0 S E A 0 1 S W Z 1 0 TAZ 0 1 V E N 0 1 W B E 0 1 TO TA L 1262 1345 1955&1967 C O M B IN E D 2 2 2607 214 Doing so permits us to see which nations most often were involved in major or important international interactions during each of the two-year periods. As expected, the United States is the most active nation according to these data, accounting for a total of 709 acts when both years' totals are combined. Also as anticipated, the Soviet Union occupied second place with 339 acts in both years. Thereafter, Israel (149), the United Kingdom (143) , Communist China (139), Egypt (113), the United Nations and other international organizations (also 113) , France (112), North Vietnam (86), and West Germany (71) round out the ten most active entities during these two years with regard only, of course, to front-page data. In order to see the different situations into which at least a few of the nations became involved during 1955 and 1967, we have provided in Table 21 the distributions for all of the nation-state actors (the United Nations and other international organizations was not included) that had at least 100 acts when both years are combined. There one can see that some of these countries were involved pre dominantly in very different situations in 1955 as com pared with 1967 and that there are some striking 215 TABLE 21 TO TAL A C T S O U T O F S E V E N MOST-ACTIVE NATIO NS B Y SITUATION U S/T O S R ISR CNlC C H N “ TM FRIT SITUATION 55 67 55 67 55 67 55 67 55 67 55 67 55 67 W A R 31 220 7 2 1 0 18 5 3 24 3 0 10 1 3 CRISIS 0 2 0 3 0 6 0 3 0 0 0 7 0 1 F O R M A L INT'l NEGOTIATIONS 58 78 84 33 4 24 35 9 8 5 5 14 36 14 INT'L INCIDENT/ DISPUTES 46 48 33 33 33 39 2 1 8 10 27 27 31 14 3 INTERNAL INCIDENT 25 10 13 12 0 2 0 6 14 25 0 0 2 1 TR A D E O R M O N E T A R Y POLICY 8 5 6 3 1 0 1 8 0 1 0 0 0 4 F O R M A L DIPLO MATIC RELAT. 4 1 4 0 0 0 1 3 1 3 1 2 0 0 D IS A R M A M E N T 28 13 22 4 0 0 4 0 0 0 0 0 4 0 |INT'L EX C H AN G ES 5 1 5 0 0 1 0 0 2 0 0 0 0 0 | AID 16 22 2 6 1 1 5 4 0 1 0 9 1 0 1 O !E X TE R N A L D O M INANCE 2 0 0 0 0 0 5 1 0 0 1 0 1 3 0 i jGEN'L DIPLO MATIC ASSESS.75 1 1 42 6 3 2 23 3 8 0 3 2 9 6 TO TAL F R E Q U E N C Y 298 411 218 121 52 97 99 44 68 64 46 67 79 33 216 differences between the nations in the distribution of acts across the various situational contexts. Thus, the United States acted a large proportion of the time in WAR situations in 1967 (over 50 percent), but only about 10 percent of the time in 1955, reflecting the major pre occupation of United States foreign policy with its Southeast Asian involvement. Most of the nations had significant proportions of their acts in the FORMAL INTERNATIONAL NEGOTIATIONS and INTERNATIONAL INCIDENTS/ DISPUTES categories, and all of them showed a decrease in the number of times they issued GENERAL DIPLOMATIC ASSESS MENTS from 1955 to 1967. Beyond these similarities, each of the nations had fairly distinctive situational profiles which in each case reflected their major involvements during the years for which the data were collected. The United States, as already mentioned, was intensely pre occupied with the Vietnam War in 1967. The Soviet Union was directly involved in Geneva as well as other negotia tions in 1955 (84 acts), but was concerned only periph erally with discussion of negotiations on the Vietnam War in 1967 (33 acts). Israel was involved in a major dispute which escalated through a crisis into a short war in 1967 and subsequently negotiated a ceasefire with the Arab 217 nations, all situations reflected by increases in acts in the appropriate categories (WAR and CRISIS) from 1955 to 1967. The United Kingdom also was directly and vitally concerned with the Geneva Talks in 1955, as was France, but neither was involved in a similar situation in 1967, a fact which is reflected in a steep drop in the number of FORMAL DIPLOMATIC ACTS recorded for each. The United Arab Republic's profile is very similar to Israel's, as one would expect, while Communist China's much larger number of WAR acts in 1955 in comparison to 1967 was caused by its ongoing fight with Nationalist China over the Tachens. This apparent ability of the situational taxonomy we have developed to reproduce in a general way the various contexts in which nations become involved as they interact in the international system is, in itself, an encouraging finding and suggests that the classification may be useful in at least some of the ways suggested earlier in this chapter. Our task in this study, however, is to attempt to demonstrate its utility in testing pre dictive hypotheses and subsequently making forecasts of future international interaction. This will be one of the subjects taken up in the chapter which follows. 218 References for Chapter V Allport, G. W. "Prejudice: A Problem in Psychological and Social Causation." Toward a General Theory of Action. Ed. by T. Parsons and E. A. Shils. New York: Harper and Row, Publishers, 1962. Torchbook ed. Pp. 365-87. Boguslaw, R. "Situational Analysis and the Problem of Action." Social Problems, VIII, No. 3 (1960-61), 212-19. Carr, L. J. Situational Analysis. New York: Harper Brothers, 1948. Cartwright, P., ed. Kurt Lewin's Field Theory in Social Science. New York: Harper and Brothers, 1951. Catlin, George E. G. A Study of the Principles of Poli tics, Being an Essay Towards Political Rationaliza tion. London: George Allen and Unwin, Ltd., 1930. Re-issued in 1967 by Russell and Russell. Cottrell, L. S.* Jr. "Analysis of Situational Fields in Social Psychology." American Sociological Review, 7 (1942), 370-382. ________. "Some Neglected Problems in Social Psychology." American Sociological Review, 15 (1950), 705-12. Deutsch, Karl W. and R. L. Merritt. "Effects of Events on National and International Images." Inter national Behavior. Ed. by H. Kelman. New York: Holt, Rinehart, Winston, 1965. Pp. 130-187. Driver, M. J. "Crisis and Reciprocity in Soviet-American Interaction: A Newspaper Headline Analysis." University of Southern California, August 1969. (Mimeographed) 219 Easton, David. The Political System. New York: Alfred A. Knopf, 1953. Gamson, W. and A. Modigliani. "Knowledge and Foreign Policy Opinions." Public Opinion Quarterly, 30 (1966), 187-199. Hermann, Charles F. "International Crisis as a Situa tional Variable." International Politics and Foreign Policy. Ed. by James N. Rosenau. New York: The Free Press, 1969. Lasswell, Harold D. World Politics and Personal Insecur ity. New York: McGraw-Hill Book Co., 1935. Levy, Marion J. Jr. "'Does it Matter if He's Naked?' Bawled the Child." Contending Approaches to International Politics. Ed. by Klaus Knorr and James N. Rosenau. Princeton: Princeton■Univer sity Press, 1969. Martin, Wayne R. and Robert A. Young. "World Event- Interaction Study: Pilot Study Report." Center for Research on Conflict Resolution. University of Michigan, 1966. (Mimeographed) McClelland, Charles A., et al. "The Communist Chinese Performance in Crisis and Non-Crisis: Quantita tive- Studies of the Taiwan Straits Confrontation, 1950-1964." Prepared for the Behavioral Sciences Group, Naval Ordnance Test Station, China Lake, California, 1965. ________. "The Access to Berlin: The Quantity and Variety of Events, 1948-1963." Quantitative International Politics: Insights and Evidence. Ed. by J. David Singer. New York: The Free Press, 1968. "Some Effects on Theory from the International Event Analysis Movement." University of Southern California, School of International Relations, 1970. (Mimeographed) 220 Meadows, P. "The Dialectic of the Situation: Some Notes on Situational Psychology." Philosophy and Phenomenological Research, 5 (1944-45), 354-64. Pruitt, Dean G. "Definitions of the Situation as a Determinant of International Action." Inter national Behavior; A Social-Psychological Analysis. Ed. by Herbert C. Kelman. Holt, Rinehart and Winston, 1965. New York: Rosenau, James N. "Pre-theories and Theories of Foreign Policy." Approaches to Comparative and Inter national Politics. Ed. by R. Barry Farrell. Evanston, 111. 1966. Northwestern University Press, Royal Institute of International Affairs. Chronology of International Events. London: Royal Institute of International Affairs, 1955. m Rummel, R. J. "Dimensions of Conflict Behavior within and between Nations." General Systems Yearbook, 8 (1963). Snyder, Richard C.; H. W. Bruck; and Burton Sapin. Foreign Policy Decision-Making: An Approach to the Study of International Politics. New York: The Free Press, 1962. and Decision-Making as an Approach to the Study of International Politics Princeton: Foreign Policy Analysis Series, No. 3, 1954. Thomas, W. I. Primitive Behavior. New York: McGraw- Hill Book Co., 1937. and Dorothy S. Thomas. The Child in America. New York: Alfred A. Knopf, 1932. "The Relation of Research to the Social Proc ess." Essays in the Social Sciences. Ed. by W. F. G. Swann et al. Washington, D.C.: The Brookings Institution, 1931. 221 Thomas, W. I. and F. Znaniecki. The Polish Peasant in America. 2d. ed. New York: Alfred A. Knopf, 1927. 2 vols. ________ . "The Behavior Pattern and the Situation." Publications of the American Sociological Society Papers and Proceedings. 22nd Annual Meeting, Vol. 22, 1927. U. S. Department of Commerce. Le Monde Index. Washing ton, D.C.: U.S. Government Printing Office, 1967. Volkart, Edmund H. Social Behavior and Personality: Contributions of W. I. Thomas to Theory and Social Research. New York: Social Science Research Council, 1951. CHAPTER VI FINAL ANALYSES, SUMMARY, AND CONCLUSIONS As we have seen in the preceding chapters, the central thesis of this study, stated very simply, is that one of the most promising ways to improve our ability both to analyze the relationships among international phenomena and make better forecasts is to develop better and more relevant taxonomies. We undertook to derive two such classification systems in an effort to support this thesis. One was intended to group nations on the basis of their international behavior, and the other to categorize international situations. In this final analytic chapter we shall attempt to demonstrate that the two taxonomies can be useful in actual empirical analyses, while at the same time illustrating in a more specific way than was possible in Chapter I the operational differences between prediction, projection, and forecasting. We shall focus first on the nation-state classifi cation system and then on the situational taxonomy, in 222 223 each case demonstrating the utility of controlling for these particular aspects of international interaction. The next part of the chapter will contain some suggestions on ways to control for both of these factors simultaneously and a general discussion of the problems encountered in attempts to forecast international interaction. The last section is devoted to a final summary of research under taken and an enumeration of the conclusions reached during the course of the study. The Nation Behavior Taxonomy In order to test the utility of the nation- behavior taxonomy, a simple hypothesis requiring only a bivariate analysis was chosen. The notion is one familiar to those working in international event analysisthe frequency of physical conflict interactions varies with the frequency of verbal conflict interactions. On the basis of this proposition one would expect that as the number of verbal conflict acts increases, the number of ^See Fitzsimmons, 1969, and Corson, 1969. 224 physical conflict acts also will increase in approximately the same, proportion.^ The same would be expected for decreases in numbers of acts. The proposition was opera tionalized first by defining WEIS combevent categories 11 through 19 as VERBAL CONFLICT, and combevent categories 20, 21, and 22 as PHYSICAL CONFLICT. Thus, VERBAL CON FLICT was composed of REJECT, ACCUSE, PROTEST, DENY, DEMAND, WARN, THREATEN, DEMONSTRATE, and REDUCE RELATION SHIPS, while PHYSICAL CONFLICT included EXPEL, SEIZE, and FORCE. These data were retrieved from the WEIS data bank for the 83 nations included in the analysis for which the nation behavioral taxonomy originally was derived. Subse quently, the three "type seven" nations, Albania, El Salvador, and Libya were excluded since that group did not have a large enough N to make analysis possible. As before, data for the four years, 1966-1969, were included. These data are shown in Table 22 where one also may note 2 No implication that the change m verbal acts causes the corresponding change in physical acts is being made; only their co-occurrence or association is being tested. 225 TABLE 22 V E R B A L A N D PHYSICAL CONFLICT B Y NATIO N 1966 - 1969 V E R B A L PHYSICAL V E R B A L PHYSICAL N ATIO N CONFLICT CONFLICT NATIO N CONFLICT CONFLICT U S A 787 85 C O P 24 6 C A N 52 4 K E N 22 4 C U B 64 10 TAZ 16 2 M E X 6 1 E T H 6 1 C O L 4 6 Z A M 38 7 V E N 9 3 R H O 34 14 P E R 17 9 S A F 40 2 B R A 17 8 M O R 8 0 B O L 4 1 A LG 34 18 C H L 13 4 T U N 1 1 2 A R G 18 5 S U D 15 4 U N K 196 18 IRN 7 2 N T H 17 8 T U R 3 1 4 B E L 28 0 IRQ 42 12 F R N 142 4 U A R 278 179 S W Z 7 1 S Y R 125 40 S P N 40 7 LE B 46 30 P O R 37 2 JO R 134 128 G M W 103 1 1 ISR 370 487 G M E 117 14 S A U 2 1 5 P O L 74 7 Y E M 10 10 A U S 5 1 S Y E 16 4 H U N 2 1 6 K U W 1 1 0 C Z E 101 12 C H N 537 82 ITA 32 8 M O N 14 3 Y U G 45 2 C H T 13 4 G R C 39 10 K O N 69 70 C Y P 4 3 K O S 32 57 B U L 16 5 JAP 47 3 R U M 38 0 IND 97 17 U S R 741 62 P A K 29 6 FIN 1 0 TAI 23 4 S W D 16 2 C A M 97 1 1 N O R 7 2 L A O 25 23 D E N 10 2 V T N 394 78 G UI 1 1 2 V T S 105 40 G H A 14 7 M A L 14 2 NIG 48 47 SIN 4 1 226 TABLE 22 (Continued) V E R B A L PHYSICAL NATIO N CONFLICT ~ 'CONFLICT PHI 24 1 INS 20 12 AU L 17 4 N E W 8 0 227 that the form of the matrix was a simple two-by-eighty matrix with the nations comprising the eighty rows and the two variables, VERBAL and PHYSICAL CONFLICT, the two columns. The next step was to reorder the data by behavioral type so that six two-by-N (the number of nations in each behavioral group) matrices were created. These data sets are shown in Table 23. With the data organized in this manner, it was possible to calculate seven different correlation coeffi- 3 cients. The first coefficient reflects the degree of association between verbal and physical conflict when all eighty nations are included in the analysis, and the remaining six the correlation between the two variables when each of the six groups of nations is analyzed sepa rately. By controlling for nation-type in this way we were able to test our contention that it may be a mistake to assume that all nations are the same, at least in a behavioral sense, so that the results of an analysis which includes all nations do not apply necessarily to each of These were Pearson Product-Moment correlations. For the formula, see Hays, 1963, p. 505. 228 TABLE 23 V E R B A L A N D PHYSICAL CONFLICT B Y INTERNATIONAL BEHAVIO R G R O U P 1966 - 1969 V E R B A L PHYSICAL V E R B A L P H Y S IC G R O U P O N E CONFLICT CONFLICT G R O U P T H R E E CONFLICT C 0N FL1 C H N 537 82 B R A 17 8 C Z E 101 12 C H T 13 4 F R N 142 4 IRN 7 2 G M W 103 1 1 JAP 47 3 IND 97 17 N E W 8 0 ISR 370 487 R U M 38 0 JO R 134 128 S W Z 7 1 K O N 69 70 TAZ 16 2 K O S 32 57 NIG 48 47 G R O U P F O U R P A K 29 6 U A R 278 179 A R G 18 5 U N K 196 18 AU L 17 4 U S A 787 85 B U L 16 5 U S R 741 62 C H L 13 4 V T N 394 78 C O L 4 6 V T S 105 40 C Y P 4 3 G H A 14 7 G R O U P T W O H U N 2 1 6 IRQ 42 12 B E L 28 0 M E X 6 1 C A M 97 1 1 M O N 14 3 C A N 52 4 N T H 17 8 C U B 64 10 P O L 74 7 D E N 10 2 SAF 40 2 FIN 1 0 SIN 4 1 G M E 117 14 TU R 3 1 4 INS 20 12 Z A M 38 7 ITA 32 8 LA O 25 23 G R O U P FIVE M A L 14 2 N O R 7 2 A LG 34 18 PHI 24 1 A U S 5 1 S P N 40 7 B O L 4 1 S W D 16 2 G R C 39 10 S Y R 125 40 K U W 1 1 0 TAI 23 4 LEB 46 30 Y U G ..... 45 2 M O R 8 0 229 TABLE 2 V E R B A L PHYSICAL G R O U P FIVE CONFLICT CONFLICT P E R 17 9 S U D 15 4 T U N 1 1 2 G R O U P SIX C O P 24 6 E T H 6 1 GUI 11 2 K E N 22 4 P O R 37 2 R H O 34 14 S A U 21 5 S Y E 16 4 V E N 9 3 Y E M 10 10 (Continued) 230 them when they are considered individually or in subgroups. The results of our test appear to support this assertion. As can be seen in Table 24, the correlation between verbal and physical conflict when all eighty nations are aggre gated is only .49. If this were the only evidence upon which one were going to reach a conclusion, that conclu sion would have to be that nations do not vary their behavior as hypothesized. The correlation coefficients when the relationship was tested by nation-behavior group, however, suggest that such a conclusion would be at least partially erroneous. These figures varied considerably both with regard to the overall correlation and as com pared to one another. At one extreme, there was no correlation (.00) whatever between the verbal and physical conflict behavior of Group 3 nations while the correlation for Group 5 countries was almost perfect (r=.90). The rest of the nation groups had correlations which fell somewhere between these two substantially different results. Thus, Group 1 nations had a correlation coeffi cient of .27, Group 6 nations .36, Group 4 nations .45, and Group 2 nations .68. What has happened’ , of course, is that the correla tion between verbal and physical conflict in the general 231 TABLE 24 C O R R E LA TIO N O F V E R B A L A N D PHYSICAL CO NFLICT 1966 - 1969 P E A R S O N G R O U P ___________ r ALL 80 N ATIO NS .49 G R O U P 1 .27 G R O U P 2 .68 G R O U P 3 .00 G R O U P 4 .45 G R O U P 5 .90 G R O U P 6 .36 232 case, i.e., when the data for all eighty nations were aggregated, represented an average degree of covariation. When the nations were grouped by behavioral type, however, it was possible to do a much finer-grained analysis which yielded more specific and accurate results. The over simplification inherent in the assumption that all nations behave in approximately the same way resulted here in an initially misleading conclusion— that there is little correlation between the verbal and physical conflict behavior of nations (r=.49)— when in fact this is true for only some of these nations, e.g., Group 4, where r=.45 while for others the conclusion is highly in accurate , e.g., Group 5 where r=.9 0. These findings thus support and illustrate what we consider to be a critical theoretical point: Nations can and should be differentiated not only on the basis of their size, wealth, or population but also according to the way in which they behave. In testing for predic tive relationships, the failure to do so probably will cause misleading conclusions to be drawn. Further, we demonstrated in an earlier chapter that the particular behavioral taxonomy developed on a pilot basis in this 233: study is composed of subgroups of nations each of which exhibits observably different patterns of international behavior. Here we have shown that these nation types can be used in an operational way to test more accurately for the existence of a hypothesized predictive relation ship. This simple demonstration gives an empirical reason to believe that controlling for type of nation using the taxonomy developed can be of use in testing for statistical relationships among variables, although other cases should be explored in order to test further this conclusion. We also wish, however, to examine the usefulness of the taxonomy with regard to forecasting. In order to do so it is necessary, at least in any scien tific forecast, to introduce the element of time. The analysis above was a purely static one since the four years of data were aggregated and examined with regard to actors and type of behavior. Thus, it was impossible to project any established relationship. To do so in any mathematical or statistical sense, the dynamic element of time must be introduced into one's original research design. In our example, this can be accomplished by 234 retrieving the data by month rather than by actor as was done in the static analysis above. When this approach is followed the rows of all of the matrices represent months instead of actors, while the columns are verbal and physical conflict as before. We also have added a column which represents time for the reasons explained below. These data matrices are provided in Table 25 for all eighty nations and Table 26 for each of the six types of nations. The research question to be investigated with these data also must be altered. In the first analysis we were asking whether, for the four years included, verbal and physical conflict varied together across nations, but here we wish to inquire whether the two variables vary together across time and/or with time. The Pearson correlation coefficients for these comparisons are given in Table 27 for the eighty nations and the six behavioral types of nations. The results of the comparison of verbal and physical conflict over time (column one) are basically the same as those in Table 24 except that they are somewhat lower. Given the substan tially different nature of the data arrays (compare Table 24 with Table 27), these results represent a somewhat 235 TABLE 25 ALL NATIONS-VERBAL A N D PHYSICAL CONFLICT B Y " MONTH-1966-1969 V E R B A L PHYSICAL Y E A R -M O N T H CONFLICT CONFLICT TO TAL 6601 103 16 119 6602 120 18 138 6603 112 12 124 6604 86 16 102 6605 137 17 154 6606 94 8 102 6607 104 18 122 6608 161 16 177 6609 128 1 1 139 6610 148 20 168 6611 127 25 152 6612 103 6 109 6701 103 23 126 6702 106 14 120 6703 67 9 76 6704 110 27 137 6705 278 64 342 6707 138 59 197 6708 110 33 143 6709 75 33 108 6710 84 22 106 6711 99 18 117 6712 59 8 67 6801 101 28 129 6802 106 23 129 6803 105 27 132 6804 148 3 1 179 6805 141 32 173 6806 90 33 123 6807 162 22 184 6808 203 46 249 6809 167 49 216 6810 126 46 172 6811 98 34 132 6812 161 41 202 236 TABLE 25 (Continued) V ER BAL PHYSICAL Y E A R -M O N T H CONFLICT CONFLICT TO TA L 6901 123 30 153 6902 159 33 192 6903 154 65 219 6904 130 53 183 6905 104 60 164 6906 95 56 151 6907 80 9 1 171 6908 100 57 157 6909 113 75 188 6910 159 ■ 6 8 227 6911 131 146 277 6912 109 90 199 5869 1767 7673 M O N T H 6601 6602 6603 6604 6605 6606 6607 6608 6609 6610 6611 6612 6701 6702 6703 6704 6705 6706 6707 6708 6709 6710 6711 C m71 O TABLE 26 TYPE 1-TYPE 6 NATIONS-PHYSICAL A N D V E R B A L CONFLICT B Y MONTH-1966-1969 G R O U P 1 G R O U P 2 G R O U P 3 V E R B A L PHYSICAL V E R B A L PHYSICAL V E R B A L PHYSICAL CONFLICT CONFLICT TO TAL M O N T H CONFLICT CONFLICT TO TAL M O N T H CONFLICT CONFLICT 68 7 75 6601 17 5 22 6601 4 0 80 6 86 6602 14 9 23 6602 4 0 64 2 66 6603 18 5 23 6603 5 0 53 8 6 1 6604 12 4 16 6604 1 2 79 8 87 6605 27 2 29 6605 7 1 64 1 65 6606 13 2 15 6606 6 1 76 8 84 6607 12 2 14 6607 2 0 108 8 116 6608 24 1 25 6608 4 0 84 5 89 6609 2 1 4 25 6609 1 0 111 12 123 6610 20 5 25 6610 2 1 92 14 106 6611 10 6 16 6611 5 0 67 3 70 6612 7 3 10 6612 2 0 69 13 82 6701 26 6 32 6701 1 1 83 10 93 6702 13 3 16 6702 5 0 57 6 63 6703 5 0 5 6703 1 0 76 14 90 6704 18 5 23 6704 1 1 112 27 139 6705 25 3 28 6705 0 0 182 42 224 6706 3 1 4 35 6706 0 1 107 53 160 6707 7 2 9 6707 2 0 86 23 109 6708 14 3 17 6708 2 0 56 3 1 87 6709 9 0 9 6709 1 0 54 19 73 6710 12 3 15 6710 2 0 52 OO 14 * 7 66 on 6711 c m n 2 1 1 - 7 0 n 21 1 - 7 6711 m o 3 T 0 A b/IZ 6 / L / jy b/ i z 1/ 6801 75 25 100 6801 20 6802 77 16 93 6802 6 6803 70 13 83 6803 14 6804 110 25 135 6804 17 6805 111 23 134 6805 13 6806 66 27 93 6806 16 6807 117 17 134 6807 19 6808 127 42 169 6808 37 6809 126 41 167 6809 23 6810 88 40 128 6810 25 6811 72 31 103 6811 16 6812 130 36 166 6812 8 6901 84 23 107 6901 7 6902 118 24 142 6902 2 1 6903 131 54 185 6903 9 6904 110 51 161 6904 9 6905 70 50 120 6905 10 6906 74 49 123 6906 3 6907 59 74 133 6907 7 6908 89 50 139 6908 2 6909 86 63 149 6909 9 6910 96 54 150 6910 3 1 6911 93 131 224 6911 14 6912 72 83 155 6912 1 1 4163 1383 5546 740 G R O U P 4 6601 5 1 6 6601 5 6602 10 1 1 1 6602 1 6603 6 1 7 6603 6 6604 10 0 10 6604 1 6605 9 4 13 6605 6 6606 7 2 9 6606 1 6607 13 2 15 6607 0 6608 2 1 4 25 6608 2 6609 14 1 15 6609 0 6610 7 2 9 6610 2 6611 15 1 16 6611 2 6612 12 0 12 6612 1 3r—•C'JCVlCOCMCMi— r — i — >tNr r — ir>OOOU3COOOCVlLO<\ICMCO CVJi— O O i — O O C O O O r — O 1/ b /iz I u i 2 1 6801 1 1 2 8 6802 9 0 9 16 6803 6 2 8 20 6804 3 0 3 15 6805 2 2 4 18 6806 1 0 1 20 6807 6 0 6 38 6808 8 0 8 24 6809 2 0 2 29 6810 3 0 3 18 6811 2 0 2 9 6812 3 0 3 8 6901 6 0 6 26 6902 10 0 10 17 6903 3 0 3 9 6904 5 0 5 16 6905 5 0 5 6 6906 6 0 6 15 6907 2 1 3 4 6908 2 1 3 14 6909 4 4 8 33 6910 0 1 1 16 6911 1 0 1 14 6912 1 0 1 884 153 20 173 G R O U P 6 7 6601 4 1 5 2 6602 11 1 12 6 6603 13 4 17 1 6604 9 2 1 1 7 6605 4 1 5 1 6606 1 2 3 0 6607 1 6 7 5 6608 2 0 2 0 6609 6 1 7 2 6610 6 0 6 3 6611 3 3 6 1 6612 12 0 12 LO G R O U P 4 6701 3 1 4 6701 6702 2 1 3 6702 6703 0 1 1 6703 6704 7 3 10 6704 6705 4 5 9 6705 6706 26 3 29 6706 6707 10 2 12 6707 6708 2 1 3 6708 6709 2 1 3 6709 6710 13 0 13 6710 6711 15 2 17 6711 6712 6 0 6 6712 6801 2 1 3 6801 6802 5 4 9 6802 6803 9 7 16 6803 6804 11 3 14 6804 6805 9 0 9 6805 6806 2 0 2 6806 6807 11 2 13 6807 6808 15 3 18 6808 6809 12 5 17 6809 6810 4 2 6 6810 6811 4 0 4 6811 6812 5 1 6 6812 6901 15 4 19 6901 6902 6 1 7 6902 6903 4 1 5 6903 r Aft a n - i TABLE 26 (.Continued) G R O U P 5 G R O U P 6 1 1 2 6701 3 1 4 2 0 2 6702 1 0 1 2 1 3 6703 2 1 3 5 0 5 6704 3 4 7 7 0 7 6705 4 3 7 29 1 1 40 6706 6 2 8 5 1 6 6707 7 1 8 2 4 6 6708 4 2 6 5 0 5 6709 2 1 3 2 0 2 6710 0 0 0 1 1 2 6711 7 1 8 1 0 1 6712 2 1 3 1 0 1 6801 2 0 2 1 1 2 6802 7 0 7 0 0 0 6803 6 3 9 2 0 2 6804 3 0 3 2 5 7 6805 4 0 4 2 4 6 6806 2 0 2 6 1 7 6807 0 1 1 6 0 6 6808 7 0 7 3 2 5 6809 0 0 0 3 0 3 6810 3 0 3 0 0 0 6811 4 1 5 6 2 8 6812 9 1 10 4 2 6 6901 7 0 7 4 3 7 6902 0 0 0 5 2 7 6903 2 0 2 &yiw c I W - T , 6905 9 2 11 6905 8 6906 9 4 13 6906 0 6907 1 0 1 6907 2 6908 2 1 3 6908 4 6909 3 1 4 6909 7 6910 4 1 5 6910 16 6911 5 1 6 6911 6 6912 5 373 1 85 6 458 6912 9 190 2 10 6905 0 0 0 0 0 6906 0 0 0 2 4 6907 • 0 0 0 3 7 6908 1 0 1 2 9 6909 0 0 0 10 26 6910 9 0 9 5 1 1 6911 7 5 12 1 10 6912 4 2 6 75 265 190 51" 241 ito U) CO 239 TABLE 27 C O R R E LA TIO N S O F V E R B A L CONFLICT, PHYSICAL CONFLICT, A N D TIME 1966 - 1969 V E R B A L CONFLICT/ PHYSICAL CONFLICT/ V E R B A L CONFLICT/ G R O U P PHYSICAL CONFLICT TIME TIME 80 .23 .74 .08 1 .26 .78 .20 2 .04 .13 .19 3 .09 .00 .00 4 .36 .03 .29 5 .74 .29 .20 6 .22 .32 .30 240 unexpected convergence of findings. For all of the nations aggregated, the correlation here is .23 compared with .49 in the static analysis. Group 5 remains the highest with a coefficient of .74 (compared to .90), while Group 1 went from .27 to .26 and Group 3 from .00 to .09. Group 2 dropped substantially from .68 to .04 here and Groups 4 and 6 increased their correlation slightly, going from .36 to .45, and .22 to .36, respectively. Two important conclusions can be drawn from these results. The first does not relate to the specific correlation and Group 3 a low one is not of central concern. What is to be emphasized, however, as pointed out above, is the finding, based upon two different cuts into the data (across nations and across time), that controlling for the type of nation on the basis of exhibited international behavior produced results quite different from those reached when all nations were aggregated and undifferen tiated. The second conclusion relates to a comparison of the overtime and static relationships. As mentioned earlier, an ability to project a relationship into the future is dependent upon the inclusion of time in that 241 relationship. The important point to be made here, since we will be discussing this matter more fully below, is that a comparison of the very same variables, in our case verbal and physical conflict, can produce significantly different results when the unit of analysis is changed. Thus, in the first set of comparisons when the unit of analysis was the nation Group 2, exhibited a moderately strong (.68) relationship between physical and verbal conflict while, using the same data but arraying them by month over time, the correlation approached zero. This result indicates that the relationship found in the static analysis may have been, for example, caused by events which took place in a relatively short period of time, but whatever the reason, it did not hold over time and cannot therefore be projected. The last two columns of correlation coefficients in Table 27 demonstrate a related set of simple predictive relationships that are of particular concern if one is interested in projecting the results of an analysis. These bivariate correlations indicate the degree of covariation of physical conflict (column two) and verbal conflict (column three) with time, which was measured in 242 simple incremental units. Thus, in Tables 25 and 26, the first variable, labeled "TIME" is 1 in month one, 2 in month two, and so on through month forty-eight. The result, of course, is that in order for a variable to correlate highly with TIME it must increase proportionatly at the same steady rate. Thus, if a group of nations had 5 verbal conflict acts in January, 1966 (month one), they would have had to have 10 in February, 15 in March, 20 in April, etc., ending up with 240 in December of 1969 in order to correlate perfectly with time measured in this way. This procedure has some disadvantages, but its chief virtue is, however, that it permits one easily to project a relationship into the future. The figures in Table 27 indicate that of the six different nation types, only the Group 1 nations' physical conflict behavior correlates at even a moderately high level (.78) with time measured in this manner and none of the groups of nations1 verbal conflict correlated higher than .30. This means that neither all of the nations, taken as a whole, nor any of the six groups increased its verbal conflict at the steady rate which would correspond with our time measure, 243 and only Group 1 and all nations taken together (row one in Table 27) appeared to do so when physical conflict is correlated with time. A review of the data in Table 25 confirms our feeling that because the 17 Group 1 nations have such a large number of verbal and physical conflict interactions compared to the remaining 63, they completely dominate these calculations, causing it to appear that all of the nations have a relatively high correlation in this case when in fact it is only these plus perhaps a few others. This result illustrates that the use of a taxonomy such as the nation behavior classification system not only can help to establish relationships and permit projections and forecasts for subsets of nations that do not show up when all nations are aggregated, but can prevent one from concluding that a general relationship holds for all nations when in fact it is true for only a few. In order to illustrate how the correlations of verbal and physical conflict with time might be used to make an actual forecast we selected the highest correla tion coefficient in Table 27, which was between the physical conflict acts of Group 1 and TIME. Relying upon 244 this relationship, one might forecast the amount of physical conflict in which these nations will engage in the month following the last one included in the analysis, which would be January, 1970. This could be done by first calculating the simple linear regression which is of the form: Y = a + bX where Y is the dependent variable (amount of physical conflict), X is the independent variable (time in months) , b is the slope of the regression line (how many units Y increases for every unit change in X), and a is the inter cept value or the value of Y when X equals zero. For our case, the equation was: Y (Physical Conflict) = -5.6 + 1.4X (Time). Since the last month in the analysis was December, 1969, or month 48, we can insert the number of the month we wish to forecast which is January, 1970, or month 49. Solving the equation, we multiply 1.4 times 49, subtract 5.6 and the result is a value of 63.0. Thus, on the basis of our best correlation coefficient, we would expect Group 1 nations to initiate 63 physical conflict acts in January 1970. The observed (actual) value, however, was 105, 245 making our forecast of 63 a poor one by any standard. This example thus demonstrates how one might go about making an actual-forecast using the results of an analysis which includes consideration of the nation type and at the same time illustrates the fact that many other factors must be considered or controlled for in the fore casting of international interaction. Among them are the length of time included in the analysis (three months, one year, two years, etc.), the technique used to make the projection (simple linear regression was used here but many others such as the numerous nonlinear regression equations available, lagged correlations, or time-series analysis might also be utilized), and the variables included in the research design (we might, for example, have used both time and amount of cooperative activity to predict physical conflict). Still another possible way to improve forecasts, which is related very directly to the original research design, is to control for the specific SITUATIONAL context in which one wishes to forecast. This possibility is explored in some detail in the next section after a brief summary of the conclu sions drawn in this portion of the chapter. 246 The central purpose of this section has been to demonstrate the usefulness of the behavioral taxonomy developed in an earlier part of this study in testing predictive relationships and making forecasts of inter national interaction. We began by hypothesizing a simple predictive relationship using a static research design and then introduced the time element necessary in a dynamic analysis from which one can make forecasts. By comparing the results of our analyses when all nations were aggregated and when nation type was controlled for, we were able to show in both instances that better— more accurate— findings were arrived at in the latter case than in the former. Controlling for nation type made a difference— sometimes a very substantial one— in the kinds of conclusions drawn from the results obtained. Con currently, we have attempted to illustrate in a more operational way than was possible in our discussion in Chapter I the differences between prediction, projection, and forecasting particularly with regard to the importance of the TIME element in such efforts. Finally, we illus trated with an actual forecast the necessity of consider ing as many factors as possible in forecasting attempts, 247 suggesting in that discussion that consideration of the situational context of inter-nation behavior may be another factor which, if controlled for, might increase the accuracy of one's conclusions. It is to an examina tion of that assertion that we now turn. The Situational Taxonomy In the last section of Chapter V, it was shown that even with the modest amount of data collected to derive the situational taxonomy, one is able to describe in summary quantitative fashion the shift in major situa tions which arose in the international system in two different years (1955 and 1967) and, at least for the most active nations, the variations in their involvements. Our purpose here is to go one step further and investigate the taxonomy's usefulness as a device for delineating the various behavioral patterns of nations or groups of 4 nations. An example can best illustrate our point. We The other major and related use is in attempts to scale international behavior types. See Chapter V for a discussion of this use. 248 hypothesized in the previous section that there was a relationship between the verbal and physical conflict behavior of nations and proceeded to test the proposition by summing the numbers of certain kinds of acts initiated by given entities during specified periods of time. Had we been able to control for the occurrence of these behavior types in varying situational contexts, however, new insights and possibly different conclusions might have resulted. Type I nations, for example, might have had a low over-time correlation between physical and verbal conflict when their behavior in all situations was aggre gated, but might have had a very high correlation in crisis situations. What is being proposed, then, is that the ability to control for still another important factor, the situational context in which interaction occurs, can contribute to an increase in our ability to make better predictions and forecasts. We will attempt to test this contention by com paring the distribution of acts in selected cases across the 22 WEIS combevent categories of each of the twelve general situational contexts. For example, suppose Nation X had the following distributions of behavior in three different situations: WEIS Comb. Situation Situation Situation Category A B C 1 0 3 0 2 0 11 0 3 0 13 1 4 2 17 0 5 0 9 0 6 0 5 0 7 1 12 0 8 0 14 0 9 0 25 1 10 0 31 1 11 0 19 0 12 2 8 1 13 3 3 3 14 1 1 2 15 5 1 8 16 6 1 4 17 11 1 4 18 9 0 6 19 9 0 6 20 12 0 5 21 17 0 7 22 53 0 27 One can see by inspection that both Situation A and Situation C have similar distributions since each shows a concentration of behavior in categories 12 through 22, while Nation X's behavior in Situation B is quite different with most of the acts falling into categories 1 through 12. If one were to intercorrelate these three distribu 250 tions, the correlation coefficient between A and C would be high, while those comparing B with A and B with C would be low. Using this statistical technique, we can test to see whether or not the situational taxonomy has the abil ity to delineate differing patterns of behavior. If our assertion that entities probably exhibit different pat terns of behavior in different situations is correct, the correlation coefficients calculated on the real data should be relatively low, indicating that the distribu tions of acts are dissimilar and controlling for this factor might be a useful enterprise. If they are high, our tentative conclusion must be that this may not be as important a factor as we had supposed and/or the taxonomy is not capable of delineating these differences. In order to operationally reach conclusions regarding the relative similarity of the distributions compared, we set an arbitrary minimum of .60, above which a correlation will be considered "high." Any coefficient below that level will be interpreted as indicating that the two distributions being compared are dissimilar. By calculating the percentage of all correlation coefficients in a given table which are classified as "high," it will 251 be possible to render some judgment about the relative similarity of the twelve situations in that particular case. Since the relationship between the variables is symmetrical, the number of coefficients that can vary is N(N-l) 2 . With N (the number of different major situations) equal to 12, the denominator for our coefficient of 5 similarity for each table will be 66. Thus, we divide the absolute number of high (over .60) correlations by 66 and use the resulting number, which can vary from 0 to 1.0, as a summary measure of the degree of similarity of the twelve situations. The most general test of our hypothesis that nations exhibit different patterns of behavior in differ ent situations can be made by aggregating all of the situation data for all actors in both years, in effect providing us a system view of both time periods. One can see from the correlation coefficients shown in Table 28 that only 20 out of the 66 correlations, or 30 percent When no data were recorded in a given category, e.g., crisis in Table 29, that situation automatically will correlate 1.0 with all others. That is, it will not be free to vary. Therefore, the number of comparisons is reduced accordingly. TABLE 28 SITUATION C O R R E LA TIO N COEFFICIENTS-ALL N ATIO N S 1955 and 1967 1 2 3 4 5 6 7 8 9 10 1 1 12 1 1.00 .51 .06 .83 .37 -.01 -.01 .08 -.09 .09 .12 .20 2 1.00 .43 .66 .38 .46 .31 .40 .16 .52 .46 .71 3 1.00 .54 .27 .89 .75 .80 .60 .79 .63 .62 4 1.00 .53 .36 .30 .50 .19 .45 .43 .55 5 1.00 .16 .27 .30 .01 .27 .61 .53 6 1.00 .83 .59 .82 .89 .59 .53 7 1.00 .38 .67 .70 .71 .44 8 1.00 .23 .57 .42 .75 9 1.00 .71 .32 .15 10 1.00 .48 .62 1 1 1.00 .54 12 1.00 252 253 of the total, were equal to or greater than .60, indicat ing that on the whole the distributions were dissimilar. Seven of those 20 high correlations were due to category 3*s (NEGOTIATIONS) similarity to categories 6 through 12. Reference to Table 59 in Appendix D confirms the intuitive hunch that these high correlations are due to the rela tively frequent use by nations in many of these situations of WEIS combevents one through ten, which generally are considered to be the cooperative types of behavior. Thus, nations involved in negotiating situations seem to exhibit patterns of behavior somewhat similar to those manifested in ECONOMIC (.89), FORMAL DIPLOMATIC (.75), DISARMAMENT (.80), INTERNATIONAL EXCHANGES (.60), AID (.79), EXTERNAL DOMINANCE (.63), and GENERAL DIPLOMATIC ASSESSMENT (.62) situations. It is interesting to note, however, that even though all these situational contexts are similar in some degree to negotiations, they do not all correlate highly with one another. Interactions relating to economic matters, for example, were quite similar only to FORMAL DIPLOMATIC, INTERNATIONAL EXCHANGE, and AID situations involving FORMAL DIPLOMATIC RELATIONS correlate at only a moderate level with INTERNATIONAL EXCHANGES (.67) , AID 254 (.70), and EXTERNAL DOMINANCE (.71). Also, as might be expected, the behavior patterns in both WAR and CRISIS situations correlated with those in situations involving INTERNATIONAL INCIDENTS AND DISPUTES. These are, however, almost all of the similarities between situations that can be found in the table, so that these results (i.e., only 30 percent high correlations) seem to support our contention that the situational taxonomy is capable of delineating different behavior patterns. Even stronger support is found when each year is analyzed separately. In Tables 29 and 30 one can see that in 1955 only 16 percent (9 out of 55) of the correlations were above .60 and in 1967 only 16 out of 66 or 24 per cent. Many of the high correlations are for the same situations as those pointed out above. In 1955 negotiating behavior was similar to 5 out of the 7 situations noted above (EXTERNAL DOMINANCE and GENERAL DIPLOMATIC ASSESS MENTS dropped below .60), interaction in economic contexts still correlated highly with INTERNATIONAL EXCHANGES and AID, but not FORMAL DIPLOMATIC RELATIONS, and WAR behavior increased its correlation with INTERNATIONAL INCIDENTS and DISPUTES. The only other high correlations were economic TABLE 29 SITUATION C O R R E LA TIO N COEFFICIENTS-ALL N ATIO N S 1955 1 2 3 4 5 6 7 8 9 10 1 1 12 1 1.00 1.00 .25 .91 .31 .06 .20 .37 -.14 .40 .20 .47 2 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 3 1.00 .53 .39 .81 .67 .63 0 0 • .71 .52 .55 4 1.00 .43 .33 .31 .49 .18 .52 .36 0 0 • 5 1.00 .40 .52 .37 .21 .47 .72 .58 6 1.00 .58 .26 .85 .73 .54 .45 7 1.00 .30 .33 .53 .56 .51 8 1.00 .11 .43 .33 .67 9 1.00 .54 .38 .20 10 1.00 .46 .80 1 1 1.00 .48 12 1.00 255 TABLE 30 SITUATION C O R R E LA TIO N COEFFICIENTS-ALL N ATIO N S 1967 1 2 3 4 5 6 7 8 9 10 1 1 12 1 1.00 .46 .02 .74 .38 -.03 -.07 .08 -.04 -.04 .11 .16 2 1.00 .47 .68 .23 .39 .21 .56 .20 .27 .37 .60 3 1.00 .53 .06 .80 .52 .74 .55 .63 .44 .66 4 1.00 .56 .32 .01 .43 .16 .22 .55 .61 5 1.00 -.12 -.31 .04 -.11 -.14 .36 .36 6 1.00 .68 .71 .54 .76 .25 .44 7 1.00 .49 .74 .72 -.07 .09 8 1.00 .43 .64 .21 .73 9 1.00 .68 -.09 -.01 10 1.00 .12 .23 1 1 1.00 .51 12 1.00 to U 1 0\ 257 situations with INTERNATIONAL EXCHANGES (.85) and AID (.73), INTERNAL INCIDENTS with EXTERNAL DOMINANCE (.72) and GENERAL DIPLOMATIC ASSESSMENTS with both DISARMAMENT (.67) and AID (.80). The coefficients in Table 30 for 1967 bear some resemblance to those in Table 28 when both years combined. INTERNATIONAL INCIDENTS and DISPUTES again is related to both WAR and CRISIS, though at a lower level (.74 and .68 respectively), and behavior in NEGOTIA TIONS is similar to 4 out of the 7 situations to which it correlated in the previous case, dropping down when com pared with FORMAL DIPLOMATIC RELATIONS, INTERNATIONAL EXCHANGES, and EXTERNAL DOMINANCE. Situation 10, AID, was similar to the largest number of other situations (5 in all here and 4 when both years were combined), including NEGOTIATIONS (.63), ECONOMIC (.76), FORMAL DIPLOMATIC RELATIONS (.72), DISARMAMENT (.64), and INTER NATIONAL EXCHANGES (.68). The large majority of the situations that inter correlate above .60 in all three of the cases has two common factors. First, a high proportion of the acts is in the more cooperative WEIS combevent categories (1 through 10) and second, many of the categories had rela- 258 tively few acts occurring in them during the years included in the analysis. The notable exceptions are NEGOTIATIONS and GENERAL DIPLOMATIC ASSESSMENTS. We feel also that more detailed analysis perhaps over longer periods of time probably would bring out differences, for example, in the mixes of cooperative acts initiated in the different situations which we now only suspect exist. Whatever the reasons for the high correlations, the results in which we are most interested here have to do with the degree of similarity of the behavioral patterns in the 12 situations. These figures, 30 percent high correlations for both years combined, 16 per cent for 1955 alone, and 24 percent for 1967 alone suggest that at least from a systemic point of view, controlling for the situa tional context of international interaction may be a useful method of disaggregating nation behavior into its component patterns. We need not be satisfied with testing our hypoth esis only on this very general level, however, since there appear to be enough data available to examine the behavior of at least two nations individually, the United States and the Soviet Union. The same sets.of correlations, for TABLE 31 SITUATION C O R R E LA TIO N COEFFICIENTS-UNITED S TA TE S 1955 and 1967 1 2 3 4 5 6 7 8 9 10 11 12 1 1.00 .21 .20 .15 .26 .36 .10 .17 -.17 .14 -.09 .28 2 1.00 .73 .68 .24 .32 .13 .92 -.13 .50 -.10 .76 3 1.00 .90 .52 .41 .47 00 « .29 .83 r ~ . o • 1 .80 4 1.00 .45 .40 .53 .72 o CM • .84 .01 .83 5 1.00 -.02 .11 .37 .19 .43 .15 .51 6 1.00 .31 .28 .41 .49 -.21 .21 7 1.00 .21 -.13 .34 -.11 .39 8 1.00 .04 .56 -.03 .83 9 1.00 .41 .10 o o • 10 1.00 o CM « 1 CM r - * • 11 1.00 -.05 12 1.00 259 TABLE 32 SITUATION CORRELATION COEFFICIENTS-UNITED STATES 1955 1 2 3 4 5 6 7 8 9 10 11 12 1 1.00 1.00 .53 .58 .13 .37 .19 .63 -.03 .54 -.19 .76 2 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 3 1.00 .87 .47 .63 .49 .48 .52 .82 -.01 .63 4 1.00 .54 .64 .23 .53 .39 .85 .05 .73 5 1.00 .10 .02 .19 .27 .29 .27 .32 6 1.00 .56 .37 .14 .60 i ro o .34 7 1.00 -.00 -.10 .16 -.09 .05 8 1.00 -.01 .35 -.03 .77 9 1.00 .62 -.11 .03 10 1.00 -.17 .59 11 1.00 o • I 12 1.00 260 TABLE 33 SITUATION C O R R E LA TIO N COEFFICIENTS-UNITED S TATES 1967 1 2 3 4 5 6 7 8 9 10 1 1 12 1 1.00 .17 .17 .09 .39 -.11 .28 .13 -.05 .11 1.00 .21 2 1.00 .87 .77 .42 .13 .69 .74 -.07 .64 1.00 .77 3 1.00 .79 .56 .07 .85 .89 I • o r o .79 1.00 .84 4 1.00 .46 -.02 .69 .64 .06 .69 1.00 .83 5 1.00 .06 .71 .46 -.13 .47 1.00 .52 6 . 1.00 -.07 .00 -.07 .14- 1.00 -.03 7 1.00 .72 -.05 .74 1.00 .80 8 1.00 .09 .63 1.00 .63 9 1.00 -.15 1.00 -.09 10 1.00 1.00 .77 1 1 1.00 1.00 12 1.00 262 both years combined and each year alone were calculated, as was a coefficient of similarity for each case. The correlations for the United States, as shown in Tables 31 (1955 and 1967 aggregated), 32 (1955), and 33 (1967), show some similarities to those for all actors combined, at least insofar as the percentage of high correlations is concerned. Thus, in the two years combined, 20 percent of the United States' distributions were similar compared to 30 percent for all nations. In 1955 the comparable figures were 16 percent versus 21 percent and in 1967 24 percent and 33 percent. As was seen previously, behavior in NEGOTIATING situations correlated with the largest number of other distributions, in this case with almost all at fiarly high levels. Only three of the five, however, were with those situations which we previously pointed out generally would be expected to elicit com paratively cooperative behavior patterns. Thus, in Table 31 one can see that United States NEGOTIATING behavior correlates not only with DISARMAMENT, AID, and GENERAL DIPLOMATIC ASSESSMENTS, but also with INTERNATIONAL INCIDENTS/DISPUTES (.90) and CRISIS (.73). These two unexpected high relationships probably can be attributed 263 to the fact that during those years, the United States was a direct participant in neither the single major crisis that erupted (in the Middle East) nor in many of the disputes and incidents that occurred. Thus, many of the acts initiated by this country were conciliatory in nature, e.g., suggestions, proposals, promises, etc. The distribution of acts in Situation 12, GENERAL DIPLOMATIC ASSESSMENTS, also correlated highly with those in five other situations, including CRISIS (.76), NEGOTIATIONS (.89), INTERNATIONAL INCIDENTS/DISPUTES (.83), DISARMAMENT (.83), and AID (.72), probably for the same reason noted above since, as one can see, these are exactly the same set of categories that intercorrelated with NEGOTIATIONS. Almost the same pattern emerges when the two years are analyzed separately. That is, CRISIS, NEGOTIATIONS, INTERNATIONAL INCIDENTS/DISPUTES, DISARMAMENT, AID, and GENERAL DIPLOMATIC ASSESSMENTS often have distributions which are similar to one another, but not to the other situations, nor do the behavior patterns in the remaining situational contexts resemble one another significantly. Thus, even when the percentage of similar situations increases from 21 percent in 1955 to 33 percent in 1967, 264 the increases still were attributable to the similarity of the behavior in four of the categories noted above. The most important fact to be noted, however, is that these three figures are, as were those when all nations were aggregated, relatively low. Soviet behavior in various situations differs somewhat from both the general system and the United States. Most significant is the fact that even fewer of that nation's distributions were similar to one another than was the case for all actors and for the United States. As can be seen in Table 34 with both years combined, only 13 percent of the patterns correlate higher than .60. In 1955 (Table 35) that figure dropped to 9 percent and in 1967 (Table 36) it was 22 percent. Reference to all of those tables indicates that most of the high correla tions resulted from similarities between Soviet behavior in INTERNAL INCIDENTS, DISARMAMENT, and GENERAL DIPLOMATIC ASSESSMENTS. This contrasts somewhat with our findings for the United States, which suggested that in addition to behavior in the latter two situations, the distribu tions in NEGOTIATIONS, INTERNATIONAL INCIDENTS/DISPUTES, and AID most often were similar to patterns of behavior TABLE 34 SITUATION C O R R E LA TIO N COEFFICIENTS-SOVIET U N IO N 1955 and 1957 1 2 3 4 5 6 7 8 9 10 1 1 12 1 1.00 .50 .40 .58 .60 .25 -.18 .46 .09 -.10 1.00 .59 2 1.00 .25 .46 .39 CO CM • -.15 .12 -.17 -.16 1.00 .75 3 1.00 .54 .48 .46 .32 .79 .06 .29 1.00 .32 4 1.00 .62 .48 .03 .68 -.03 .14 1.00 .63 5 1.00 .21 .11 .44 .02 .04 1.00 .46 6 1.00 .26 .58 .42 .23 1.00 .65 7 1.00 .19 .38 • 0 0 1.00 .03 8 1.00 .11 .14 1.00 .41 9 1.00 .32 1.00 .00 10 1.00 1.00 -.09 1 1 1.00 1.00 12 1.00 TABLE 35 SITUATION C O R R E LA TIO N COEFFICIENTS-SOVIET U N IO N 1955 1 2 3 4 5 6 7 8 9 10 1 1 12 1 1.00 1.00 .34 .37 .29 .13 -.19 .61 -.08 -.16 1.00 .58 2 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 3 1.00 .51 .54 .48 .42 .61 .15 -.18 1.00 .14 4 1.00 .62 .50 0.00 .68 .06 -.14 1.00 .55 5 1.00 .56 .47 .36 .25 -.26 1.00 .51 6 1.00 .44 .21 .52 .12 1.00 .60 7 1.00 0.00 0 0 CO • -.12 1.00 .06 8 1.00 .06 -.11 1.00 .32 9 1.00 .47 1.00 .01 10 1.00 1.00 -.11 11 1.00 1.00 12 1.00 TABLE 36 SITUATION C O R R ELATIO N COEFFrCIENTS-SOVIET U N IO N 1967 1 2 3 4 5 6 7 8 9 10 1 1 12 1 1.00 .34 .59 .61 .69 .12 1.00 .01 1.00 .04 1.00 .56 2 1.00 .72 .37 .33 CO CM • 1.00 .12 1.00 -.14 1.00 .68 3 1.00 .52 .62 .13 1.00 .19 1.00 .05 1.00 .62 4 1.00 .54 • o ■p* 1.00 0.00 1.00 .23 1.00 .41 5 1.00 -.09 1.00 -.22 1.00 -.02 1.00 .29 6 1.00 1.00 .39 1.00 -.14 1.00 .68 7 1.00 1.00 1.00 1.00 1.00 1.00 8 1.00 1.00 .61 1.00 .29 9 1.00 1.00 1.00 1.00 10 1.00 1.00 -.16 1 1 1.00 1.00 12 1.00 268 in other situations. One can note further that the cor relation coefficients in these three tables generally are much lower than those in previous cases. The coefficients for the United States, for example, often were above .80 and occasionally were higher than .90. None of the cor relations of Soviet behavior in different situations, whether for both years combined or each year individually, are above .79, and a large majority of those considered "high" are between .60 and .70. On the basis of these data, therefore, one may conclude that the Soviet Union, even more than the United States, tends to exhibit quite different patterns of behavior according to the situation in which it is involved. Our purpose in this section, as stated at its beginning, has been to test the ability of the situational taxonomy developed earlier in this study to delineate the varying patterns of behavior exhibited by nations in different situational contexts. The classification system's ability to discriminate along these lines was tested by comparing, for selected cases, the distribution of acts across the 22 WEIS combevent categories in each of the twelve situations. It was suggested that if the 269 correlations generally were low this result could be interpreted as meaning that the taxonomy is a useful one for the purpose stated. A summary of the percentages of "high” (>.60) correlations found in each of the nine cases tested is as follows: 1955 and 1967 Combined 1955 1967 All Actors 30% 16% 24% United States 20 21 33 Soviet Union 13 9 22 / The mean percentage for these comparisons is just under 20 percent, and the range is a low of 9 percent to a high of 33 percent. Even though constituting a limited test, we believe these results to be very convincing. They are supported further by simple inspection of the data tables in Appendix D. Nations apparently do exhibit different patterns of behavior in different situational contexts, and the results cited tend to indicate that the situational taxonomy developed in this study is a useful device for disaggregating international interaction in this 270 manner. And, as discussed in detail earlier, in addition to other possible applications, such a capability probably is an important factor in improving our ability to develop better predictive hypotheses and make better forecasts. A Preliminary Event/Interaction Forecasting Paradigm In the first two sections of this chapter, we demonstrated the usefulness first of the nation-behavior classification system and second of the situation taxonomy. i A last subject of concern before final conclusions are drawn is how the two might be used together to forecast future international interaction. The idea here is to overlay one on the other, permitting one in effect to control for both factors simultaneously. The matrix shown in Figure 6 illustrates our conception of how this might be accomplished. It was formed by ranging the twelve major situational contexts down the side as the rows, and the six major types of nations across the top as the columns. Each cell thus represents the behavior of a given type of nation in a specified situational context. For example, if one were interested in the behavior of Type I nations in CRISIS situations (context 2), the cell in M A J O R IN TE R N A TIO N A L SITU ATIO N AL C O N T E X T S NATION BEHAVIO RAL TY P E S .......... } I II III IV V VI 1 2 3 4 12 FIG U RE 6 — The Form of a Possible Relationship Between Actor Type and Situation 272 position 3-1, i.e., row 3, column 1 would be the single target of concern. In order to operationalize our measure of behavior, a third dimension must be added, as shown in Figure 7. There one can see that ranging the 22 WEIS combevent categories from the front to the back of the cube permits us to conceive of a behavior pattern that would be unique to each cell of the formerly two-dimensional matrix. In our previous example, one can see that the behavior of Type I nations in CRISIS situations now can be specified in a clear and unambiguous manner in terms of the specific mix of events that historically have occurred within the context of that situation, e.g., two YIELDS, ten COMMENTS, thirty CONSULTS, etc. Before proceeding to a demonstration of how such a pattern can be translated into sets of probabilities for forecasting purposes, it is necessary to introduce two other factors, both of which must be controlled for in some manner. First, we must ask ourselves what period of time is relevant in making this forecast? The answer will be dependent, as always, upon available theory and experience. Thus, it has been tentatively suggested that 273 22 WEIS COMBEVENTS NATION BEHA 'IOR.TYPES II. *o o , CO FIGURE 7 — THE RELATIONSHIP BETWEEN ACTOR TYPE, SITUATION AND BEHAVIOR 274 if one is interested in forecasting one month into the future, a short run— perhaps the most recent three months— is likely to yield a more accurate forecast than if one year of data are included (Moore and Young, 1969) . A forecast intended for a longer period of time, perhaps two or three years out,' might require that three or four years of historical data be included in the calculations. The second major factor that should be mentioned is the target of the behavior we are interested in fore casting. Just as one might wish to specify the actor to actors to be considered, it may be important also to con trol for the target. In the example above, we specified that only the behavior of a certain group— Type I nations— was to be included in the analysis. For a wide variety of reasons, we might be interested in these nations' acts toward Type III nations only, toward each other only, toward nations in Southeast Asia only, or toward Communist countries only. Conversely, of course, one might wish also to vary the actor in the same manner. We now have suggested that in a paradigm which is designed to forecast future behavior patterns, five factors must be considered: ACTOR, TARGETS, TIME, SITUA- 275 TION, and BEHAVIOR. It is not feasible to represent these five considerations in a diagram, but it should be clear that the paradigm we have in mind is five-dimensional and therefore extremely complex. We shall first demon strate with the aid of a simple example how one might employ this very preliminary model to make actual fore casts, after which several important aspects and implica tions of our proposal will be discussed. In the example forecast, we controlled for each of the five factors in the following manner: ACTORS: All Type I Nations. TARGETS: All Nations. \ TIME: (a) Four years of historical data, 1966-1969. (b) Forecast to be applicable one year out, i.e., for the calendar year 1970. SITUATION: For comparative purposes, three different situations were chosen: (a) WAR; (b) NEGO TIATIONS; and (c) INTERNATIONAL INCIDENTS/ DISPUTES. BEHAVIOR: The full range of possible acts (WEIS combevents 1-22) was included. The raw data used to forecast the behavior pattern of Type I nations in the four different situational contexts is shown in Table 37. In order to make the actual fore cast, these relative frequencies can be converted to GROUP 1 NATIONS' OUTPUT TABLE 37 BY COMBEVENT FOR SELECTED 276 SITUATIONS 1 1966 - 1969 3 4 WAR NEGOTIATIONS INT'L DISPUTES YIELD 3 0 6 COMMENT 54 43 21 CONSULT 19 19 16 APPROVE 7 4 6 PROMISE 5 3 1 GRANT n 0 2 REWARD 3 0 8 AGREE 7 26 6 REQUEST 8 12 4 PROPOSE 8 30 9 REJECT 7 10 5 ACCUSE 15 30 14 PROTEST 8 1 11 DENY 1 0 2 DEMAND 0 9 6 WARN 4 2 8 THREATEN 3 2 3 DEMONSTRATE 1 1 8 REDUCE RELATIONSHIP 2 5 3 EXPEL 0 0 0 SEIZE 3 0 6 FORCE 192 0 23 277 percentages and treated as a priori forecasts of likely future behavior patterns. These values, shown in Table 38 again demonstrate the significant differences between behavior in different situations. In the first case, WAR, one can see that our forecast suggests that we ought to expect Type I nations to allocate over half of their acts to category 22— Force— and 15 percent to Comment. Agreements are rare, occurring only 2 percent of the time, as are Approvals (2 percent) and Rewards (1 percent). This contrasts with Negotiations, where the large majority of the acts can be expected to be Comments (22 percent), Consultations (10 percent), Agreements (13 percent), proposals (15 percent), and Accusations (15 percent). The behavior forecast for these nations in INTERNATIONAL DISPUTES suggests a quite different pattern than for either of the other two situations. In this case one would expect a much greater variety of behavior, with 14 percent in Force, 13 percent in Comment,'10 percent in Consult, 8 percent in Accuse, 7 percent in Protest, 5 per cent in Reward, War, Propose, and Demonstrate, and 4 per cent in Approve, Agree, Demand, and Seize. These examples are provided only as a demonstra- 278 TABLE 38 GROUP 1 NATIONS OUTPUT BY COMBEVENT FOR SELECTED SITUATIONS IN PERCENTAGES 1966 - 1969 1 3 4 WAR NEGOTIATIONS INT L DISPUTES YIELD .83 0.0 3.57 COMMENT 14.96 21.83 12.50 CONSULT 5.26 9.64 9.52 APPROVE 1.94 2.03 3.57 PROMISE 1.39 1.52 .60 GRANT 3.05 0.0 1.19 REWARD .83 0.0 4.76 AGREE 1.94 13.20 3.57 REQUEST 2.22 6.09 2.38 PROPOSE 2.22 15.23 5.36 REJECT 1.94 5.08 2.98 ACCUSE 4.16 15.23 8.33 PROTEST 2.22 .51 6.55 DENY .28 0.0 1.19 DEMAND 0.0 4.57 3.57 WARN 1.11 1.02 4.76 THREATEN .83 1.02 1.79 DEMONSTRATE .28 .51 4.76 REDUCE RELATIONSHIP .55 2.54 1.79 EXPEL 0.0 0.0 0.0 SEIZE .83 0.0 3.57 FORCE 53.19 0.0 13.69 279 tion of the possible use of this kind of a paradigm for purposes of forecasting. Relying as it does upon the theory suggested originally by Professor Charles McClelland that the best clue to what nations are likely to do in the future may be what they have done in the past (McClelland, 1969: 19-23), we believe it is suggestive of a method for better anticipating future patterns of inter national behavior. We feel that the ability to provide a table of probabilities, such as those shown in Table 38, may contribute to a move away from the more deterministic kinds of forecasts toward those that are more probabilis tically oriented. This ability to state, in the early part of 1970, that the probability of United States, Type I, or European nations* exhibiting a given pattern of behavior during the year in specified situations should increase our interest and perhaps our accuracy in inter national interaction forecasts. One may note that our preliminary paradigm bears some resemblance to one reportedly being developed at the University of Michigan which is called a "Computer-Aided System for Conflict and Cooperation Intensity Measure ment" (Beattie etal., 1970). Both obviously have their 280 roots deeply buried in the work of Charles McClelland and his associates on the World Event/Interaction Survey. Our conception is somewhat simpler but more explicit than CACIM, however. Where we have demonstrated how it is possible to derive a probability table which can be used to forecast future behavior, the Michigan system is in tended to compute intensity separately for conflictive and cooperative actions, for deeds, and for three types of words. The three types are (1) the actor's evaluation of action by the target (e.g., disapproval, approval); (2) the actor's desire for action by the target (e.g., demands, proposals); and (3) the actor's intent=regarding his own future actions vis-£-vis the target (e.g., threats, promises). CACIM thus provides eight types of intensity measurement for monitoring events; non verbal intensity and three types of verbal inten sity for both conflict and cooperation. . . . CACIM provides the following indicators applicable to crisis anticipation and management; (1) Intensity for each discrete action prior to and during a crisis. (2) Aggregate intensity for each of the eight types of action (conflictive and cooperative, verbal and nonverbal) over time (e.g., by day). Using these indicators, the analyst can monitor the following aspects of conflicts: (1) The diversity of conflictive or cooperative action types a party uses as an index of the seriousness with which a party engages in con flict or cooperation. 281 (2) The extent to which verbal statements of desire and intent are accompanied or followed by corresponding nonverbal deeds. (3) The intensity of verbal expressions of intent relative to expressions of desire and evalua tion as an index of a party's commitment to carry out conflictive or cooperative deeds. (4) The intensity of conflictive or cooperative deeds supporting verbal expressions of desire and intent as an index of the credibility of verbal actions. (5) The interest-area involved in a party's actions as an index of the seriousness of its intentions (e.g., military actions may indi cate a greater commitment to achieve objec tives than diplomatic or political actions). (Beattie et al., 1970: 1-2) We must admit to being somewhat skeptical of such an enterprise at this time. Lacking further information, it seems to us that several major problems remain unsolved. To mention only one, the immense difficulties encountered in the scaling of international interaction have not been overcome to our knowledge, although as discussed in Chapter V we have some hopes that the situational taxonomy developed in this study may make a modest contribution to such an effort. The essential difference between the two approaches, however, probably is one of approach. Again following the lead of Charles McClelland, we are convinced that the most promising procedure is to begin building a forecasting model by keeping it as simple as 282 possible, taking into account only the most essential factors at the beginning and using as the major criterion for adding factors each one's ability to produce a better forecast, to complicate it in a gradual manner. The Michigan approach, on the other hand, attempts to take into account all of the factors which appear to be rele vant at the outset, thereby running the risk of immediately becoming overwhelmed with immense complexity and seemingly g insoluable problems. Which of the two approaches will eventually be successful can only be conjectured about at this time, although we feel that, given the difficulties encountered previously by those attempting to solve the problem in one step, a new approach such as that orig inally suggested by Professor Charles McClelland and dis cussed above appears more promising. In this section we have attempted to demonstrate how the two taxonomies developed earlier in this study can be "overlaid" upon one another and used along with certain g The Michigan group certainly is not the only one to favor such an approach. Another well-known example was the TEMPER project completed several years ago under the sponsorship of the Joint War Games Agency. 283 other basic factors in a first-cut forecasting paradigm, thus, we were able to provide an example, calculated from real data, of how a table of probabilities might be derived which could be used as a forecast for one type of nations' behavior in several different situations. Relying upon the theory that the best information about what nations will do in the future is what they have done in the past, we suggested that the distribution of past acts in a given period of time, when converted to proportions, could be used as a forecast of expected future behavior. Finally, we noted the apparent similarity of our scheme to one proposed by scholars at the University of Michigan and discussed some basic differences in approach. We will turn now to a final summary of the entire research effort. Final Summary and Conclusions As suggested by the title, this study has been exploration into the extremely broad and complex subject of prediction and forecasting in international relations. We began with a discussion of the relationship between explanation and prediction and then attempted to define as clearly as possible the concepts of prediction, 284 projection and forecasting. The basic nature of inter national relations phenomena was taken up in the next section which included a discussion of "patterned1 1 and "unique" events, after which we concluded that on both theoretical and practical grounds, prediction and fore casting in international relations are possible. Finally, the major normative arguments for and against such efforts were presented and it was asserted that the real issue is not whether to predict or not to predict, but rather how to go about it (Sprout and Sprout, 1961: 66). Before beginning the actual analysis, four guide lines or limitations were established. First, the study was confined to an examination of short-term prediction and forecasting. Next, we restricted the investigation to questions of more academic than practical significance, although one can see that some of our results probably are relevant to the latter. We also opted for the system approach over the field theoretic approach as a framework within which to do our analysis and elected to proceed in a generally inductive manner, although not in any rigid sense. It then was suggested that international relations probably is in one of the classic early stages of a science 285 during which the development of better taxonomies for sorting and classifying data is an essential task. On that basis, it was decided that the development of two different classification systems would constitute the main research effort. The first was designed to categorize nations on the basis of their observed international behavior, thus challenging the assumption that all nations have basically the same motivations, aspirations, methods of statecraft and so on. On this basis, we hypothesized that more accurate results might be obtained if one tested for predictive relationships and made forecasts that were applicable to given types of nations rather than to all entities in the system. Four years of data (1966-1969) were analyzed with two different statistical techniques, Factor Analysis and Hierarchical Grouping Analysis, and after concluding that composite indicators of behavior were better able to discriminate between nations than event types, a first-cut nation-state behavioral taxonomy was derived which con sisted of seven types of nations. We then asserted that although the classification system essentially was static in nature, enough dynamic aspects of the data were 286 included (i.e., the indicators were allowed to vary by quarters) to permit the projection of the results three or four years into the future. It also was seen that this taxonomy is so different from other well-known classification systems based upon different types of data that it provides a reasonable alternative means of describing and grouping the nations in the international system. The second taxonomy developed was based upon the theory, well-developed in other social sciences but some what neglected in international relations, that the situational context of social interaction is an important determinant of behavior. Two more specific reasons for the choice of this area of investigation also were pre sented. The first was our conviction that the more factors we are able to control for, the greater our chances of generating accurate predictive hypotheses which subse quently could be used to make better forecasts. Secondly, we felt that the taxonomy might be used in future research efforts to derive scale or intensity values for individual types of behavior which would be based on the demonstrated importance of a given act in a specified situational 287 context. In order to develop a classification system that would be capable of categorizing international situations, data for two years, 1955 and 1967, were collected from the front page of the New York Times. Following the method successfully employed in the deriva tion of the WEIS Event category system, varying subsets of the collection were analyzed item by item and general descriptors of the situational context of each listed. The resulting list of descriptors then was tabulated, a rough classification system derived and, after several revisions, a preliminary situational taxonomy was arrived at which was composed of twelve major types of inter national situations. The final portion of the study, presented in the earlier sections of this chapter, was intended to accom plish several ends. First and perhaps most important, each of the taxonomies was tested to ascertain its useful ness for the purposes for which it was intended. Thus, a specific predictive hypothesis was tested for all nations aggregated and for each of the groups of countries the nation behavior taxonomy in order to determine whether or not better, more accurate results were obtained when 288 nation type was controlled for than when all entities were assumed to be equal. In order to examine the useful ness of the taxonomy for forecasting purposes, the same hypothesis also was tested over time in addition to the static analysis. We next considered the situational taxonomy. The basic question there was whether or not the classifi cation system was capable of categorizing international interaction in such a way that one would be able to recognize the different patterns of behavior we hypoth esized nations exhibited in different situational con texts. This test was performed by comparing the distribu tions of acts for several cases in varying situational contexts in an attempt to see if these patterns emerged as we suggested they would if the taxonomy had the capability ascribed to it. Another goal of this section was to discuss and, where possible, demonstrate in a more operational way than was possible earlier in the analysis the relation ships among the key terms prediction, projection and forecasting. Here we concentrated mainly on the role of time, pointing out that many, perhaps most, predictive 289 relationships are hypothesized and tested in a static way so that no conclusions regarding the future can be drawn. We also attempted to show how quite different results might be obtained when the same relationship is tested using both a static and dynamic model. The final portion of this concluding analysis was devoted to a presentation of some preliminary ideas on how the two taxonomies developed in this study might be used, together with other basic factors, to begin building an international interaction forecasting model. The paradigm was based upon the theory, initially proposed by Charles McClelland, that the best predictor of a nation's future behavior may be its past record. On that basis, we suggested that the distribution of past behavior in a given situation for a specified actor or group of actors could be converted to proportions or percentages and treated as expected values for purposes of short-range forecasting. With this short summary of our study completed, the final matter to be considered is the question of what, after all, has been accomplished as a result of this research effort? What conclusions can be drawn on the 290 basis of this study? Our intention, it will be recalled, was to conduct an exploratory analysis of prediction and forecasting in international relations. In pursuit of this end, it was necessary to delve into a variety of areas ranging from the philosophy of science and episte- mology on the one hand to statistics on the other. As a result, the conclusions reached vary widely both in their import and originality, i.e., many were derived from the work of others. We consider each of them to be important in some degree, however, to the structure of the study and to an understanding of this difficult area of investi gation. Subject to the limitations set forth throughout the analysis, they are as follows: From an operational and possibly a theoretical point of view, there are essential differences between explanation and prediction. Popular usage to the contrary, a prediction does not necessarily refer to the future, but involves only a statement of the form of the relationship between two or more variables. In science, a projection provides an estimate of the value of a variable at some future time, and a forecast is a result of the 291 projection of at least one element of a predictive hypothesis. Nations exhibit discernible patterns of behavior in their interaction with other members of the system. Probabilistic, class forecasts in international relations are both theoretically and technically possible. From a normative point of view, the weight of opinion is that forecasting in international relations is not only desirable but necessary. There are quite significant differences between short-, mid-, and long-range forecasts both with regard to the variables considered and the projec tion techniques employed. System theory and field theory, while they have much in common, are quite different— perhaps comple mentary— when viewed from the perspective of the nature of the predictive relationships investigated. 292 The field of international relations probably is in one of the classic early stages in the evolution of a science during which the development of rele vant taxonomies is of particular importance. Classifying actors according to "state descriptions" of their actions and classifying international' interaction on the basis of the situational context of that behavior are important requirements for the development of international system theory. Attempts to derive "global relationships" or relationships which purport to be equally applicable to all nations may be a dysfunctional research approach. There is a great need for many more well-defined nation-state classification systems than now exist that are capable of discriminating between types of nations on a wide variety of bases.' Although a number of taxonomies of nations exist in the literature, only one other analysis even has considered the possibility of classifying nations 293 on the basis of their international behavior. No source of data on international interaction is completely free of bias. The data from which the nation behavior taxonomy was derived are considered a special purpose set of international interaction information which are intended only to record the "control efforts" of nations and probably contain an American bias or view of the world. Hierarchical grouping analysis, while it has some difficulties associated with its use, is considered a good supplementary technique, especially if one is interested in avoiding the possibility of arriving at results that could be technique- dependent . There are three relatively clear underlying dimen sions of nation-state behavior in the four years of WEIS data: DIPLOMATIC EXCHANGE, NONMILITARY CON FLICT , and MILITARY CONFLICT. 294 The use of indexes or indicators of different aspects of international interaction permit one to discriminate better between nations on the basis of their external behavior than the use of the 22 WEIS combevent categories. A taxonomy consisting of seven types of nations was derived. Each of the nation types can be characterized and differentiated from the others on the basis of summary measures of its character istics. The nation behavior classification system can be assumed to be relatively stable in the short run so that it probably can be "projected," i.e., the assignment of nations to specific categories can be held constant, a maximum of four years into the future. On the basis of the lack of similarity between our nation behavior taxonomy and three others derived from different types of data, we conclude that event/interaction data of the kind used here probably provide a valid alternative means of 295 classifying nations in the international system. Controlling for the type of nation on the basis of nation behavior taxonomy permits one to more accurately test for predictive relationships in both dynamic and static cases. Results also indicated that more accurate forecasts can be obtained if this factor is considered. Many social scientists, including some political scientists, have demonstrated that the situational context is an important consideration in the under standing of human interaction, but the concept has not been well developed in the field of international relations. Research in the field which has considered the situational context of international behavior generally has been carried out according to the decision-making or international system approaches or with reference to "issue-areas." Despite their rather widespread use in the field, the classification of international behavior 296 according to abstract categories such as "political," "social," "economic," etc., probably is not very useful and often may turn out to be dysfunctional. The availability of a situational taxonomy may provide a means for approaching the event scaling problem in a more promising manner than those attempted to date. The data used to derive the situational taxonomy contain the same biases as those used for the nation behavior taxonomy, but are further limited since only information on the front page was recorded. A situational taxonomy was derived which consisted of twelve categories arrayed on a nominal scale of measurement. The situational taxonomy developed here has the ability to reproduce in a general way the various situational contexts in which nations become involved as they interact in the international system. 297 The situational taxonomy developed here probably is a useful device for disaggregating international interaction, permitting one to observe the different patterns of behavior exhibited by nations in different situational contexts. By "overlaying" the situational taxonomy on the nation behavior classification system and taking into account other relevant factors, it is possible to produce a probabilistic forecast of future behavior patterns. The paradigm suggested here and one proposed recently by scholars at the University of Michigan are somewhat similar because of their common back ground, but there appears to be some basic differ ences in their respective approaches to the forecasting problem. We believe that these conclusions indicate that the effort expended on this study has been worthwhile and that a modest advance has been made in our understanding of the complex subject of prediction and forecasting in international relations. The contributions made generally 298 are of two types. On the one hand, we have attempted to clarify some often confusing concepts and place others more firmly within the context of the field of inter national relations, while on the other two classification systems have been developed on a pilot study basis. We believe both will prove useful not only in prediction and forecasting but in other related areas of endeavor as well. Finally, we hope that in a general way the analyses presented here will contribute in some way to the increasingly widespread national effort to make some progress in this vital area of endeavor. 299 References for Chapter VI Corson, Walter. "Measuring Conflicts and Cooperation Intensity in East-West Relations: A Manual and Codebook.1 1 Second Draft, Institute for Social Research, University of Michigan, November 1969. (Mimeographed) Fitzsimmons, Barbara. "The Role of Violence in Inter national Conflicts." Support Study No. 1, World Event/Interaction Survey, University of Southern California, March 1968. (Mimeographed) Hays, William L. Statistics for Psychologists. New York: Holt, Rinehart and Winston, 1963. McClelland, Charles A. "International Interaction Anal ysis in the Predictive Mode." Technical Report #3, World Event/Interaction Survey, University of Southern California, 1969. (Mimeographed) Moore, James A. and Robert A. Young. "Some Preliminary Short-Term Predictions of International Inter action." Working Paper No. 1, World Event/Inter action Survey, University of Southern California, September 1969. Sprout, Harold and Margaret S. Sprout. "Explanation and Prediction in International Politics." Inter national Politics and Foreign Policy: A Reader in Research and Theory. Ed. by James A. Rosenau. Glencoe, 111.: The Free Press of Glencoe, 1961. Pp. 60-72. BIBLIOGRAPHY 300 BIBLIOGRAPHY B O O K S Bailey, Thomas A. A Diplomatic History of the American People. N ew York: Appleton-Century-Crofts. Seventh ed. 1964. Carr, L. J. Situational Analysis. N ew York: Harper Brothers. 1948. Cartwright, P. ed. Kurt Lewin's Field Theory in Social Science. N ew York: Harper Brothers. 1951. C a tlin , George E. G . A Study of the Principles of P o litic s , Being an Essay Towards P o litic a l R ationalization. London: George Allen & Unwin, Ltd. 1930. Dakin, A. Calvinism. Philadelphia: The Westminster Press, 1946. Dozer, Donald M . The Monroe Doctrine: Its Modern Significance. N ew York: Alfred A. Knopf. 1965. Easton, David. The P o litic a l System. N ew York: Alfred A. Knopf. 1953. Fisher, H.A.L. A History of Europe. Boston: Houghton M iflin Co. 1939. Haney, Lewis H. Business Forecasting. N ew York: Ginn and Company. 1931. Harman, Harry H. Modern Factor Analysis. Chicago: University of Chicago Press. 1960. Hays, William L. S ta tistics fo r Psychologists. N ew York: H olt, Rinehart and Winston. 1963. Herz, John H. International P o litic s in the Atomic Age. N ew York: Columbia University Press. 1959. Hoffman, Stanley ed. Contemporary Theory in International Relations. Englewood C liffs , N ew Jersey. Prentice H all, Inc. 1960. 301 _____ 302 Kaplan, Abraham. The Conduct of Inquiry; Methodology fo r Behavioral Science. San Francisco. Chandler Publishing C o : T9W: Kaplan, Morton A. System and Process in International P o litic s . N ew York: Jonn Wiley & Sons, Inc. 1957. Karsten, Karl. S cie n tific Forecasting. N ew York: Greenbore Publisher, Inc. 1931. Kerner, Robert S. The Urge to Sea, Berkeley: U niversity of C alifornia Press. 1946. Lai 1, Arthur. Modern International Negotiations: Principles and Practice. N ew York: Columbia U niversity Press. 1966. Lasswell, Harold D . World P o litics and Personal Inse curity. N ew York: McGraw-Hill Book Co. 1935. Lerche, Charles.0. Jr. and Abdul A. Said. Concepts of International P o litic s . Englewood C liffs , N ew Jersey: P rentice-H all, Inc. 1963. Lipset, Seymour H. P o litic a l Man. Garden C ity, N ew Jersey: Doubleday & C cf! 1960. Meehan, Eugene J. The Theory and Method o f P o litic a l Analysis. Homewood, Illin o is : The Dorsey Press. 1965. Meehan, Eugene J. Explanation in Social Science: A System Paradigm. Homewood, Illin o is : The Dorsey Press. 1968. M ills , C. Wriqht. The Marxists. N ew York: Dell Publishing Co. 1962. Morgenstern, Oskar. Wirtschaftsprognose: Eine Untersuchung ih re r Voraussetzungen und Moeglichkeiten, Vienna. 1928. Morgenthau, Hans. P o litics Am ong Nations: The Struggle fo r Power and Peace. N ew York: Alfred A. Knopf. 1948. Northrop, F.S.C. The Logic of the Sciences and the Humanities. N ew York: Meridian Books, Inc. 1947. ~ Organski, A.F.K. World P o litic s . N ew York: A lfred A. Knopf. 1961. Perkins, Dexter. A History of the Monroe Doctrine. Revised ed. Boston: L ittle , Brown and Co. 1955. Plato. The Republic. Translated by Benjamin Jowett. N ew York: The World Publishing Company. 1946. 303 Richardson, Lewis F. The S ta tis tic s of Deadl.y Quarrels. Pittsburgh: The Boxwood Press. 1960. Russett, Bruce M . Trends in World P o litic s . N ew York: Macmillan. 1965. _______________ . International Regions and the International System: A Study in P o litic a l Ecology! Yale University. Rand McNally & Company. Chicago. 1967. ________________ , ert al_. World Handbook o f P o litic a l and Social Indicators. N ew Haven: Yale University Press. 1964. Scheffler, Isra e l. The Anatomy of Inquiry. N ew York: A lfred A. Knopf. 1963. Snyder, Richard C ., H.W. Bruck, and Burton Sapin. Foreign Policy Decision-Making: An Approach to the Study of International P o litic s . N ew York: Tne Free Press. 1962. Sprout, Harold and Margaret Sprout. Foundations of International P o litic s . Princeton: D. Van Nostrand Company, Inc. 1962. Thomas, W.I. P rim itive Behavior. N ew York: McGraw H ill Book Co. 1937. ___________ , and Dorothy S. Thomas. The Child in America. N ew York: Alfred A. Knopf. 1932. ___________ , and F. Znaniecki. The Polish Peasant in America. Second ed. (2 vols.) N ew York: A lfred A. Knopf. 1927. Veldman, Donald J. Fortran Programming fo r the Behavioral Sciences. N ew York: H olt, Rinehart and Winston. 1967. Volkart, Edm und H. Social Behavior and Personality: Contributions of W.I. Thomas to Theory and Social Research. N ew York: Social Science Research Council. 1951. Wright, Quincy. The Study of International Relations. N ew York: Appleton-Century-Crofts. 1955. 304 ARTICLES A N D PERIODICALS A ll port, G.W . "Prejudice: A Problem in Psychological and Social Causation." Ed. by T. Parsons and E.A. Shi 1 s. Toward a General Theory of Action. N ew York: Harper & R ow . Torchbook ed. 1962. 365-87. Banks, Arthur S. and P h illip M . Gregg. "Grouping P o litic a l Systems: Q-Factor Analysis o f 'A Cross-Polity Survey."' American Behavioral S cie ntist. 9, 3 (1965) 3-5. Berry, Brian J.L. "Basic Patterns of Economic Development." Ed. by Norton Ginsburg. Atlas o f Economic Development. Chicago: University o f Chicago Press. 1961. Boguslaw, R. "Situational Analysis and the Problem of Action." Social Problems, V III, 3 (1960-1961) 212-219. Bouldinq, Kenneth. "Is Peace Researchable?" Background, 6, 4 (1963) 70-77. C a tte ll, Raymond B. "The Dimensions of Culture Patterns of Factorization of National Characters." Journal of Abnormal and Social Psychology. 44, 2 (1949) 443-469. __________________ , H. Bruel and H.P. Hartman. "An Attempt at More Refined D efinitions of the Cultural Dimensions o f S vn ta lity in Modern Nations." American Sociological Review. 17, 4 (1951) 408-421. Coleman, James S. "The P o litic a l Systems o f the Developing A^eas." Ed. by Gabriel A. Almond and James S. Coleman. l.,e P o litic s o f Developing Areas. Princeton University Press, i960. C o ttre ll, L.S. J r. "Analysis of Situational Fields in Social Psychology." American Sociological Review. 7 (1942) 370-382. ________ . "Some Neglected Problems in Social Psychology." American'Sociological Review. 15 (1950) 705-712. C utright, P h illip s . "National P o litic a l Development: Measurement and Analysis." American Sociological Review. 28, 2 (1963) 253-264. Deutsch, Karl W . and R.L. M e rritt. "Effects of Events on National and International Images." Ed. by H. Kelman 305 International Behavior. N ew York: H olt, Rinehart, Winston. 1965. Dixon, W.J. ed. "B M D Biomedical Computer Programs." University of C alifornia Publications in Automatic Computation. 2. Berkeley: University of C alifornia Press. 1968. _______ "B M D Biomedical Computer Programs, X-Series Supplement." U niversity of C alifornia Publications in Automatic Computation. 3. Berkeley: University of C alifornia Press. 1969. Feierabend, Ivo K. and Rosalind L. Feierabend. "Aggressive Behaviors Within P o litie s , 1948-1962: A Cross-National Study." Journal o f C onflict Resolution. X, 3 (1966) 249-271. Fitzgibbon, R.H. and Kenneth Johnson. "Measurement o f Latin American P o litic a l Chanqe." American P o litic a l Science .Review. 55 (1961) 515-526. Fleming, William G . "Sub-Saharan A frica: Case Studies of International A ttitudes and Transactions of Ghana and Uganda." Ed. by James N. Rosenau. Linkage P o litic s . The Free Press. N ew York. 1969. Fried!ander, Saul. "Forecasting in International Relations." Ed. by Bertrand de Jouvenal. Futuribles: Studies in Conjecture, I I , Geneva: L ib ra irie Droz, S.A. (1965) 1. Gamson, W . and A. M odigliani. "Knowledge and Foreign Policy Opinions." Public Opinion Quarterly. 30 (1966) 187- 199. Gleditsch, Nils Petter. "Trends in World A irlin e Patterns." Journal of Peace Research. 4 (1967) 366-408. Gould, Loyal N. "The E N D C and the Press." SIPRI, Stockholm Papers. 3, Stockholm: Almqvist & W iksell. 1969. Harary, Frank. "A Structural Analysis o f the Situation in the Middle East." Journal o f C onflict Resolution, V, 2 (1961). Hempel, Carl G . and Paul Oppenheim. "The Logic of Explanation." Ed. by H. Feigl and M . Brodbeck. Readings in the Philosophy of Science. N ew York: Appleton-Century- C rofts, Inc. 1953. 306 International Behavior: A Social-Psychological Analysis. N ew York: H olt, Rinehart and Winston. 1965. Rosenau, James N. "Pre-Theories and Theories o f Foreign P olicy." Ed. by R. Barry F a rre ll. Approaches to Comparative and International P o litic s . Evanston, Illin o is : Northwestern U niversity Press. 19^6. Rummel, R.J. "Dimensions o f C onflict Behavior Within and Between Nations." General Systems Yearbook. 8 (1963). _________ . "Testing S om e Possible Predictors Within and Between Nations." Peace Research Society Papers. 1964. . "Foreign C onflict Behavior Code Sheet." World P oTitics. X V III, 2 (1966) 283-296. __________ . "Understanding Factor Analysis." Journal of C o n flict Resolution, XI, 4 (1967) 444-480. Singer, J. David and Melvin Small. "Alliance Aggregation and the Onset of War, 1815-1945." Ed. by J. David Singer. Q uantitative International P o litic s . N ew York: The Free Press. 1968. Snyder, Richard C., H.W. Bruck, and Burton Sapin. "Decision- Making As An Approach to the Study of International P o litic s ." Princeton: Foreign Policy Analysis Series. 3 (1954). _________________ j__________ ;________________ . "The Decision-Making Approach to the Study of International P o litic s and Foreign Policy: A Reader in Research arid" Theory. N ew York: The Free Press. 1961. Sprout, Harold and Margaret S. Sprout. "Explanation and Prediction in International P o litic s ." Ed. by James N. Rosenau. International P o litic s and Foreign Policy: A Reader in Research and Theory^ N ew York: The Free Press. 1961. Thompson, Kenneth W . "Toward a Theory o f International P o litic s ." American P o litic a l Science Review. XLIX, 3 (1955) 733-746. Thomas, W.I. " T h e Behavior Pattern and the S ituation." Piblications of the American Sociological Society Papers and Proceedings, 22nd AnnualMeeting. 22 (wy: 307 Hermann, Charles F. "International C risis as a Situational Variable." Ed. by James N. Rosenau. International P o litic s and Foreign Policy. N ew York: The Free Press. 1969. Levy, Marion J. J r. '"Does I t Matter I f He's Naked?' Bawled the C hild." Ed. by Klaus Knorr and James N. Rosenau. Contending Approaches to International P o litic s . Princeton: Princeton University Press. 1969. Lundberg, George A. "The Postulates of Science and Their Implications fo r Sociology." Ed. by M . Natanson. Philosophy o f the Social Sciences: A Reader. N ew York: Random House. 1963. Marget, Arthur W . "Morgenstern on the Methodology o f Economic Forecasting." The Journal o f P o litic a l Economy. 37, 3 (1929) 312-339. Meadows, P. "The D ialectic of the S ituation: S om e Notes on Situational Psychology." Philosophy and Phenomenological Research. 5 (1944-45) 354-364. M e rrill, John C. "U.S. Panel N am es World's Ten Leading 'Q uality D a ilie s .'" Journalism Quarterly. 41, 4 (1964). McClelland, Charles A. "The Access to B erlin: The Quantity and Variety o f Events, 1948-1963." Ed. by J. David Singer. Quantitative International P o litic s : Insights and Evidence! N ew York: The Free Press. 1968. ________ . ''Applications of General Systems Theory in International Relations." Main Currents in Modern Thought. 12 (1955) 27-34. Ed. by James N. Rosenau. International P o litic s and Foreign P olicy. N ew York: The Free Press. 1961. _________ _______ . "The Function o f Theory in International Relations." Journal of C o nflict Resolution. IX, 3 (1960). ____________________ , et al_. "The Communist Chinese Performance in C risis and Non-Crisis: Q uantitative Studies of the Taiwan S tra its Confrontation, 1950-1964." Prepared fo r the Behavioral Sciences Group. Naval Ordnance Test S tation. China Lake, C alifornia: 1965. P ru itt, Dean G . "D efinitions of the Situation as a Determinant of International A ction." Ed. by Herbert C. Kelman. 308 Thomas, W.I. "The Relation of Research to the Social Process." Ed. by W.F.G. Swann, et a]_. Essays in the Social Sciences. Washington, D.C.: The Brookings In s titu tio n . 1931. Ward, J.H. J r. and Marion E. Hook. "Application o f an Hierarchical Grouping Procedure to a Problem of Grouping P ro file s ." Educational and Psychological Measurement. 23, 1 (1963). Wriston, Henry. "Thoughts fo r Tomorrow." Foreign A ffa irs . 40, 3 (1962) 374-391. UNPUBLISHED R E S E A R C H B a ll, Geoffrey H. and David S. H all. "Background Information on Clustering Techniques." Technical paper. Menlo Park: Stanford Research In s titu te . 1968. Bobrow, Davis B. "International Indicators." Paper prepared fo r delivery at the 1969 Annual Meeting o f the American P o litic a l Science Association, N ew York C ity. September 2-6, 1969. (Mimeographed) Corson, Walter. "Measuring C onflicts and Cooperation Intensity in East-West Relations: A Manual and Codebook." Second d ra ft. In s titu te fo r Social Research. University of Michigan. November 1969. (Mimeographed) Denton, Frank H. "A Handbook o f Factor Analysis fo r International Relations." School of International Relations, Univer s ity of Southern C alifornia. November 1965. (Mimeographed) Driver, M.J. "C risis and Reciprocity in Soviet-American Interaction: A Newspaper Headline Analysis." University o f Southern C alifornia. August 1969. (Mimeographed) Fitzsimmons, Barbara. "The Role of Violence in International C o n flicts." Support Study #1. World Event/Interaction Survey. University of Southern C a lifo rnia . January 1969. (Mimeographed) 309 Fitzsimmons, Barbara, Gary Hoggard* Charles McClelland, W ayne Martin and Robert A. Young. "World Event/Interaction Survey Handbook and Codebook." World Event/Interaction Survey Technical Report #1. School o f International Relations. University o f Southern C a lifo rnia . January 1969. (Mimeographed) Hoggard, Gary D. "Indicators o f International In te ra ctio n ." University of Southern C a lifornia . July 1969. (Mimeographed) McClelland, Charles A. "International Interaction Analysis: Basic Research and S om e Practical A pplications." World Event/Interaction Survey Technical Report #2. School o f International Relations. University of Southern C alifornia. 1968. (Mimeographed) . "International Interaction Analysis in the Predictive Mode." World Event/Interaction Survey Technical Report #3. School of International Relations. University of Southern C a lifornia . 1969. (Mimeographed) . World Event/Interaction Survey Quarterly A c tiv itie s Report fo r March-May 1969. U niversity of Southern C alifornia. . "Some Effects on Theory from the International Event Analysis Movement." World Event/Interaction Survey. School o f International Relations. University of Southern C a lifo rnia . 1970 (Mimeographed) ________________ and Anne Ancoli. "An Interaction Survey o f the Middle East." World Event/Interaction Survey Technical Report #7. School of International Relations. University of Southern C alifornia. 1970 (mimeographed) Martin, W ayne R. "A Study o f Limited War: Vietnam and Korea." Unpublished P hD dissertation. U niversity of Southern C alifornia. 1970. ___________, and Robert A. Young. "World Event/Interaction Study: P ilo t Study Report." Center fo r Research on C onflict Resolution. U niversity of Michigan. 1966. (Mimeographed) 310 M artin, W ayne R. and Robert A. Young. "A Review o f Six International Event/Interaction Category and Scaling Methods." School of International Relations. University of Southern C alifornia. 1968. (Mimeographed) Moore, James A. "Attention and Interaction in the International System: A Case Study." Unpublished M.A. Thesis. University o f Southern C a lifo rnia . 1970. . and Robert A. Young. "Some Preliminary Short-Term Predictions of International Intera ctio n ." Working Paper #1. World Event/Interaction Survey. University of Southern C alifornia. September 1969. Nie, Norman H., Dale H. Bent and C. Hadlai H ull. S ta tis tic a l Package fo r the Social Sciences (Provisional Users Manual). November 1968. (Mimeographed) P h illip s , Warren and Dennis H all. "The Importance of Governmental Structure as a Taxonomic Scheme fo r Nations." The Dimensionality o f Nations Project. Research Report #18. Department of P o litic a l Science. University o f Hawaii, n.d. (Mimeographed) Puchala, Donald J. "Operational Handbook fo r a Study of Intergovernmental Interaction During Regional Integration." N ew York: Columbia University. 1967. (Mimeographed) Riggs, Fred W . "The D ialectics of Development." Department o f P o litic a l Science. University o f Hawaii. 1968. (Mimeographed) Rubin, T .J ., E.M. Krass and A.H. Schainblatt. "Projected International Patterns I I . " Santa Barbara: General Electric-Tempo. 1967. Tanter, Raymond. "Explanation, Prediction and Forecasting in International P o litic s ." Ann Arbor: University of Michigan. 1969. (Mimeographed) 311 O T H E R S O U R C E S Crowe, S ir Eyre. "Memorandum on the Present State of B ritish Relations with France and Germany." B ritis h Documents on the Origins of the War: 1898-1914. Ed. by G.P. Gooch and Harold Termperley. London: H.M.S.O. 3 (1928). Heinemann, H.E. "Economics: Taking the Guess Out of Prediction." N ew York Times. October 28, 1969, p. 16. Royal In s titu te of International A ffa irs . Chronology of International Events. London: Royal In s titu te of International. A ffa irs . 1955. U.S. Department of Commerce. Le Monde Index. U.S. Government P rinting O ffice. 1967. APPENDICES 312 APPENDIX A W EIS ALPHABETIC A N D N U M E R IC A C T O R C O D E S 313 TABLE 39 N U M ER IC AL LISTING O F W EIS A C T O R S J U M B E R ALPH A A C T O R N U M B E R ALPH A A C T O R 2 U S A U S A 230 S P N Spain 20 C A N Canada 232 A N D Andorra 40 C U B Cuba 235 P O R Portugal 41 HAI H aiti 255 G M W Germany/Fed.Rep. 42 D O M Dominican Republic 265 G M E Germany/Dem.Rep. 5 1 JA M Jamaica 266 E B E Berlin/East 52 TRI Trinidad-Tobago 267 W B E Berlin/West 53 B A R Barbados 290 P O L Poland 70 M E X Mexi co 90 G U A Guatemala 305 A U S Austria 91 H O N Honduras 310 H U N Hungary 92 E LS El Salvador 315 C Z E Czechoslovakia 93 NIC Nicaragua 325 ITA Ita ly 94 C O S Costa Rica 328 VAT Vatican 95 P A N Panam a 331 S A N San Marino 338 M LT Malta 100 C O L Columbia 339 ALB A 1 bania 101 V E N Venezuela 345 Y U G Yugoslavia 110 G U Y Guyana 350 G R C Greece 130 E C U Ecuador 352 C Y P Cyprus 135 P E R Peru 355 B U L Bulgaria 140 B R A Brazil 360 R U M Rumania 145 B O L B olivia 365 U S R U S S R 150 P A R Paraguay 375 FIN Finland 155 C H L Chile 380 S W D Sweden 160 A R G Argentina 385 N O R Norway 165 U R U Uruguay 390 D E N Denmark 198 A FP A lliance fo r Progress 394 W A R Warsaw Pact 199 O A S O A S 395 ICE Iceland 396 N A T N A T O 200 U N K United Kingdom 397 E E C E E C 205 IRE Ireland 398 E FT E FTA 210 N T H Netherlands 399 U N O Any In t. Org. 211 B E L Belgium (UN, e tc.) 212 LU X Luxemburg 220 F R N France 420 G A M Gam bi a 221 M O C Monaco 432 M L I Mali 223 LIC Liechtenstein 433 S E N Senegal 225 S W Z Switzerland 434 D A H Dahomey 315 N U M B E R ALPH A A C T O R N U M B E R A LP H A A C T O R 435 M A U Mauritania 615 A LG A 1 geria 436 NIR Niger 616 T U N Tunisia 437 I V O Ivory Coast 620 LB Y Libya 438 GUI Guinea 625 S U D Sudan 439 U P P Upper Volta 630 IRN Iran 440 G U E E quitorial Guinea 640 T U R Turkey 450 LBR Liberia 645 IRQ Iraq 451 SIE Sierra Leone 651 U A R U A R (Egypt) 452 G H A Ghana 652 S Y R Syri a 461 T O G Togo 660 LEB Lebanon 471 C A O Cameroun 663 JO R Jordan 475 NIG Nigeria 666 ISR Israel 476 BIA Biafra 670 S A U Saudi Arabia 481 G A B Gabon 678 Y E M Yem en 482 C E N Central A fric . Rep. 681 S Y E South Yem en 483 C H A Chad 690 K U W Kuwait 484 C O N Congo(Brazza.) 697 P L O Arab C om m ando 490 C O P Congo(Kinshasa) groups 698 M O M Muscat and O m a n 500 U G A Uganda 699 A R L Arab League 501 K E N Kenya 510 TAZ Tanzania 700 A F G Afghanistan 516 BUI Burundi 710 C H N China, Peoples 517 R W A Rwanda Rep. 520 S O M Somalia 712 M O N Mongolia 530 E T H Ethiopia 713 C H T China, Republic 551 Z A M Zambia of (Taiwan) 552 R H O Rhodesia 720 H O K Hong Kong 553 M A W Malawi 721 M A C Macao 560 S A F South A frica 731 K O N Korea, North 570 LE S Lesotho 732 K O S Korea, South 571 B O T Botswana 740 JAP Japan 572 S W A Swaziland 750 IND India 580 M A G Malagasy 770 P A K Pakistan 590 M A R Mauritius 775 B U R Burma 599 O A U O A U 780 C E Y Ceylon 782 M A D Mai dive 600 M O R Morocco 790 N E P Nepal ! NUMBER ALPHA ACTOR 800 TAI Thailand 811 C A M Cambodia 812 LA O Laos 813 LA P Pathet Lao 816 V T N Vietnam, North 817 V T S Vietnam, South 818 V C G Vietcong and NLF 820 M A L Malaysia 830 SIN Singapore 840 PHI Philippines 850 INS Indonesia 900 A U L Australia 920 N E W N ew Zealand 921 N A U Nauru 990 W S M Western Sam oa 992 S E A S E A T O 998 M L G Any m u ltila te ra l group 999 N S C Not stated, unidentified target 317 TABLE 40 ALPHABETICAL LISTING O F W EIS A C T O R S N U M B E R A LP H A A C T O R N U M B E R ALPH A Actor 700 A F G Afghanistan 042 D O M Dominican Repub. 339 ALB Albania 615 A LG Algeria 130 E C U Ecuador 232 A N D Andorra 092 E LS El Salvador 160 A R G Argentina 440 G U E E quitorial Guinea 900 A U L Australia 305 A U S Austria 530 E T H Ethiopia 053 B A R Barbados 375 FIN Finland 211 B E L Belgium 220 F R N France 266 E B E Berlin/East 267 W B E Berlin/West 481 G A B Gabon 145 B O L B olivia 420 . • ..G AM ... Gambia 571 B O T Botswana 265 G M E Germany/Dem.Rep. 140 B R A Brazil 255 G M W Germany/Fed.Rep. 355 B U L Bulgaria 452 G H A Ghana 775 B U R Burma 350 G R C Greece 516 BUI Burundi 090 G U A Guatemala 438 GUI Guinea 811 C A M Cambodia 110 G U Y Guyana 471 C A O Cameroun 020 C A N Canada 041 HAI H aiti 482 C E N Central African 091 H O N Honduras Rep. 310 H U N Hungary 780 C E Y Ceylon 720 H O K Hong Kong 483 C H A Chad 155 C H L Chile 395 ICE Iceland 710 C H N China, Peoples Rep. 750 IND India 713 C H T China, Rep. of-Taiwan 850 INS Indonesia 100 C O L Columbi a 630 IRN Iran 484 C O N Congo(Brazzaville) 645 IRQ Iraq 490 C O P Congo(Kinshasa) 205 IRE Ireland 094 C O S Costa Rica 666 ISR Israel 040 C U B Cuba 325 ITA Ita ly 352 C Y P Cyprus 437 IVO Ivory Coast 315 C Z E Czechoslovakia 051 JAM Jamaica 434 D A H Dahomey 740 JAP Japan 390 D E N Denmark 663 JO R Jordan 318 7], D U M B E R ALPH A A C T O R N U M B E R ALPHA A C T O R 501 K E N Kenya 150 P A R Paraguay 731 K O N Korea/North 135 P E R Peru 732 K O S Korea/South 840 PHI Philippines 690 K U W Kuwait 290 P O L Poland 235 P O R Portugal 812 L A O Laos 660 LEB Lebanon 552 R H O Rhodesia 570 LE S Lesotho 360 R U M Rumania 450 LBR Liberia 517 R W A Rwanda 620 LB Y Libya 223 LIC Liechtenstein 331 S A N San Marino 212 LU X Luxemburg 670 S A U Saudi Arabia 433 S E N Senegal 721 M A C Macao 451 SIE Sierra Leone 580 M A G Malagasy 830 SIN Singapore 553 M A W Malawi 520 S O M Somalia 820 M A L Malaysia 560 SAF South A frica 782 M A D Maidive 681 S Y E South Yem en 432 M L I Mali 230 S P N Spain 338 M LT Malta 625 S U D Sudan 590 M A R Mauritius 572 S W A Swaziland 435 M A U Mauritania 380 S W D Sweden 070 M E X Mexi co 225 S W Z Swi tzerland 221 M O G Monaco 652 S Y R Syria 712 M O N Mongolia 600 M O R Morocco 510 TAZ Tanzania 698 M O M Muscat and O m a n 800 TAI Thailand 461 T O G Togo 921 N A U Nauru 052 TRI Trinidad-Tobago 790 N E P Nepal 616 T U N Tunisia 210 N T H Netherlands 640 T U R Turkey 920 N E W N ew Zealand 093 NIC Nicaragua 500 U G A Uganda 436 NIR Niger 365 U S R U S S R 475 NIG Nigeria 651 U A R U A R (Egypt) 385 N O R Norway 200 U N K United Kingdom 002 U S A U S A 770 P A K Pakistan 439 U P P Upper Volta 095 P A N Panam a 165 U R U Uruguay | NUMBER ALPHA ACTOR 328 V A T Vatican 101 V E N Venezuela 816 V T N Vietnam/North 817 V T S Vietnam/South 990 W S M Western S am oa 678 Y E M Yem en 681 S Y E Yemen/South 345 Y U G Yugoslavia 551 Z A M Zambia N O N -G O V E R N M E N T A L A C T O R S 198 A FP Alliance fo r Progress 699 A R L Arab League 476 BIA Biafra 397 E E C E E C 398 E FT EFTA 396 N A T N A T O 199 O A S O A S 599 O A U O A U 697 P LO Arab C om m ando groups 813 LAP Pathet Lao 992 S E A S E A T O 818 V C G Vietcong and N LF 394 W A R Warsaw Pact 399 U N O Any in t i. org. (UN, e tc.) 998 M L G Any m ultila te ra l group 999 N S C Not stated, unidentified target APPENDIX B R A W A N D R A N K E D D A T A O N SIX INDICATOR VARIABLES 320 NATIO N V O L U M E TABLE 4 1 R A W D A T A O N SIX INDICATOR VARIABLES 83 S E LE C TE D N ATIO N S 1966-1969 DIPLOMATIC NON-MIL. MILIT. ~E X C H A N G E CONFLICT CONFLICT S C O P E 321 B A LA N C E U S A 4505 4069 373 63 33.62 1.23 C A N 272 245 26 1 6.33 1.83 C U B 175 110 64 1 6.03 1.80 M E X 52 48 4 0 4.33 1.16 E LS 32 15 1 1 6 6.40 1.10 C O L 50 41 6 3 3.57 2.00 V E N 36 29 6 1 3.00 1.57 P E R 68 50 17 1 5.23 1.17 B R A 80 64 13 3 3.48 1.21 B O L 39 34 5 - 0 3.90 1.86 C H L 53 40 12 1 4.42 1.96 A R G 5 1 40 8 3 3.19 1.42 U N K 907 795 103 9 10.55 1.14 N T H 66 47 lb 4 2.64 1.83 B E L 90 75 15 0 5.00 2.65 F R N 685 614 70 1 10.07 1.43 S W Z 49 45 4 0 2.45 1.40 S P N 135 104 29 2 5.19 1.59 P O R 83 70 13 0 3.9b 1.41 G M W 725 678 44 3 13.18 1.34 G M E 281 177 102 2 9.06 1.64 P O L 198 137 59 2 5.66 1.89 A U S 56 53 3 0 2.95 1.27 H U N 86 65 19 2 4.10 1.72 C Z E 379 306 7 1 2 10.53 1.08 ITA 165 140 22 3 4.02 1.67 ALB 62 34 28 0 4.43 1.82 Y U G 227 191 35 1 5.16 1.47 G R C 131 104 26 1 5.04 1.13 C Y P 43 36 7 0 7.17 1.05 B U L 59 44 14 1 2.95 1.51 R U M 211 194 17 0 5.02 1.49 U S R 2224 1615 594 15 24.17 1.36 FIN 4 1 41 0 0 4.10 1.41 S W D 105 90 15 0 5.00 2.23 N O R 34 27 6 1 2.27 1.70 322 TABLE 41 (Continued) DIPLOMATIC NON-MIL. MILIT. NATIO N V O L U M E E X C H A N G E CONFLICT CO NFLICT S C O P E B A LA N C E D E N 55 45 8 2 2.75 1.96 G U I 36 27 9 0 3.27 .92 G H A 65 52 13 0 3.82 1.38 NIG 232 166 36 30 12.21 1.36 C O P 73 5 1 19 3 3.48 1.09 K E N 54 34 20 0 2.84 3.00 TAZ 53 41 12 0 2.79 1.47 E T H 49 44 4 1 2.23 .94 Z A M 101 66 32 3 5.61 2.24 R H O 97 74 19 4 8.82 1.04 S A F 87 65 22 0 4.14 1.98 M O R 52 46 6 0 2.60 1.30 A LG 131 95 35 1 4.23 1.70 T U N 40 3 1 7 2 2.00 1.43 LB Y 50 43 6 1 3.13 1.11 S U D 59 45 13 1 3.11 1.90 IRN 93 89 2 2 3.32 1.18 T U R 121 90 28 3 5.76 1.49 IR Q 160 126 26 8 5.71 1.70 U A R 939 566 206 167 22.36 1.55 S Y R 318 186 96 36 9.64 1.44 LE B 137 78 34 25 7.21 1.02 JO R 497 271 100 126 17.75 1.19 ISR 1445 722 271 452 34.40 1.27 S A U 89 70 15 4 4.68 1.09 Y E M 48 32 13 3 3.43 .96 S Y E 52 38 1 1 3 2.48 1.44 K U W 40 37 3 0 2.67 2.67 C H N 967 440 486 4 1 14.01 1.58 M O N 33 16 16 1 3.67 2.54 C H T 49 33 13 3 3.27 1.22 K O N 209 90 54 65 12.29 1.06 K O S 207 127 28 52 12.18 1.42 JAP 293 261 29 3 10.46 1.68 IND 341 261 70 10 8.97 1.23 P A K 173 146 22 5 6.18 1.06 TAI 139 119 17 3 6.04 1.58 C A M 228 145 75 8 12.67 2.75 LA O 121 80 25 16 8.07 1.78 V T N 925 562 289 •74 20.11 1.39 V T S 496 396 64 36 21.57 1.35 323 TABLE 41 (Continued) DIPLOMATIC NON-MIL. MILIT. NATION V O L U M E E X C H A N G E CONFLICT CONFLICT S C O P E B A LA N C E M A L 94 85 7 2 7.23 .99 SIN 45 43 2 0 3.75 .98 PHI 156 145 10 1 7.80 1.93 INS 146 122 13 1 1 7.68 1.93 A U L 133 118 1 1 4 4.93 2.11 N E W 52 46 6 0 3.71 4.33 324 TABLE 42 R A N K E D D A T A O N SIX INDICATOR VARIABLES ' 83 S E LE C TE D N ATIO NS 1966-1969 \TION V O L U M E DIPLOMATIC E X C H A N G E NON-MIL. CONFLICT MILIT. CONFLICT S C O P E B A LA N C I U S A 1.0 1.0 3.0 6.0 2,0 25.5 C A N 17.0 15.0 31.0 55.5 29.0 65.5 C U B 25.0 31.0 15.5 55.5 32.0 63.0 M E X 63.5 54.0 77.0 73.5 48.0 18.0 E LS 83.0 83.0 60.0 20.0 28.0 14.0 C O L 67.5 66.0 71.5 32.0 60.0 74.0 V E N 80.0 79.0 71.5 55.5 70.0 51.0 P E R 51.0 53.0 42.0 55.5 37.0 19.0 B R A 49.0 49.0 53.0 32.0 61.5 23.0 B O L 78.0 74.0 75.0 73.5 55.0 67.0 C H L 60.5 68.5 57.5 55.5 47.0 71.5 A R G 66.0 68.5 64.5 32.0 67.0 39.5 U N K 7.0 3.0 7.0 17.0 14.0 17.0 N T H 52.0 55.0 46.5 23.5 77.0 65.5 BE L 44.0 42.0 46.5 73.5 42.5 79.0 F R N 9.0 6.0 13.5 55.5 17.0 41.5 S W Z 70.0 59.0 77.0 73.5 80.0 36.0 S P N 33.0 32.5 25.5 43.0 38.0 54.0 P O R 48.0 44.5 53.0 73.5 54.0 37.5 G M W 8.0 5.0 19.0 32.0 9.0 30.0 G M E 16.0 19.0 . 8.0 43.0 19.0 55.0 P O L 24.0 25.0 17.0 43.0 35.0 68.0 A U S 57.0 50.0 79.5 73.5 71.5 27.5 H U N 47.0 47.5 39.0 43.0 51.5 61.0 C Z E 12.0 11.0 12.0 43.0 15.0 11.0 ITA 27.0 24.0 35.0 32.0 53.0 56.0 ALB 54.0 74.0 28.0 73.5 46.0 64.0 Y U G 20.0 17.0 21.5 55.5 39.0 45.5 G R C 35.5 32.5 31.0 55.5 40.0 16.0 C Y P 74.0 72.0 67.0 73.5 27.0 8.0 B U L 55.5 61.5 49.0 55.5 71.5 49.0 R U M 21.0 16.0 42.0 73.5 41.0 47.5 U S R 2.0 2.0 1.0 14.0 3.0 32.5 FIN 75.0 66.0 83.0 73.5 51.5 37.5 S W D 39.0 36.0 46.5 73.5 42.5 76.0 N O R 81.0 80.5 71.5 55.5 81.0 59.0 NATIO N DIPLOMATIC V O L U M E E X C H A N G E TABLE 42 (Continued) NON-MIL. MILIT. CONFLICT CO NFLICT S C O P E ' 325 B A LA N C E D E N 58.0 59.0 64.5 43.0 75.0 71.5 G UI 79.0 80.5 63.0 73.5 65.5 1.0 G H A 53.0 51.0 53.0 73.5 56.0 34.0 NIG 18.0 20.0 20.0 11.0 12.0 32.5 C O P 50.0 52.0 39.0 32.0 61.5 12.5 K E N 59.0 74.0 37.0 73.5 73.0 82.0 TAZ 60.5 66.0 57.5 73.5 74.0 45.5 E T H 70.0 61.5 77.0 55.5 82.0 2.0 Z A M 40.0 46.0 24.0 32.0 36.0 77.0 R H O 41.0 43.0 39.0 23.5 21.0 7.0 S A F 46.0 47.5 35.0 73.5 50.0 73.0 M O R 63.5 56.5 71.5 73.5 78.0 29.0 A LG 35.5 34.0 21.5 55.5 49.0 59.0 T U N 76.5 78.0 67.0 43.0 83.0 41.5 LB Y 67.5 63.5 71.5 55.5 68.0 15.0 S U D 55.5 59.0 53.0 55.5 69.0 69.0 IRN 43.0 38.0 81.5 43.0 64.0 20.0 T U R 37.5 36.0 28.0 32.0 33.0 47.5 IR Q 28.0 27.0 31.0 18.5 34.0 59.0 U A R 5.0 7.0 6.0 2.0 4.0 50.0 S Y R 14.0 18.0 10.0 9.5 18.0 43.5 LEB 32.0 41.0 23.0 12.0 26.0 6.0 JO R 10.0 12.0 9.0 3.0 7.0 21.0 ISR 3.0 4.0 5.0 1.0 1.0 27.5 S A U 45.0 44.5 46.5 23.5 45.0 12.5 Y E M 72.0 77.0 53.0 32.0 63.0 3.0 S Y E 63.5 70.0 60.0 32.0 79.0 43.5 K U W 76.5 71.0 79.5 73.5 76.0 80.0 C H N 4.0 9.0 2.0 8.0 8.0 52.5 M O N 82.0 82.0 44.0 55.5 59.0 78.0 C H T 70.0 76.0 53.0 32.0 65.5 24.0 K O N 22.0 36.0 18.0 5.0 11.0 9.5 K O S 23.0 26.0 28.0 7.0 13.0 39.5 JAP 15.0 13.5 25.5 32.0 16.0 57.0 IND 13.0 13.5 13.5 16.0 20.0 25.5 P A K 26.0 21.0 35.0 21.0 30.0 9.5 TAI 31.0 29.0 42.0 32.0 31.0 52.5 C A M 19.0 22.5 11.0 18.5 10.0 81.0 L A O 37.5 40.0 33.0 13.0 22.0 62.0 V T N 6.0 8.0 4.0 4.0 6.0 35.0 V T S 11.0 10.0 15.5 9.5 5.0 31.0 326 TABLE 42 (Continued) DIPLOMATIC NON-MIL. MILIT. N ATIO N V O L U M E E X C H A N G E C O N FLTC T CONFLICT S C O P E B A LA N C E M A L 42.0 39.0 67.0 43.0 25.0 5.0 SIN 73.0 63.5 81.5 73.5 57.0 4.0 PHI 29.0 22.5 62.0 55.5 23.0 70.0 INS 30.0 28.0 53.0 15.0 24.0 22.0 AU L 34.0 30.0 60.0 23.5 44.0 75.0 N E W 63.5 56.5 71.5 73.5 58.0 83.0 327 TABLE 43 R A W D A T A O N SIX INDICATOR VARIABLES G R O U P O N E N ATIO NS 1966-1969 (Aggregated) IT I O N V O L U M E DIPLOMATIC E X C H A N G E NON-MIL. CONFLICT MILIT. CONFLICT S C O P E B A LA N C E C H N 967 440 486 4 1 14.01 1.58 C Z E 379 306 7 1 2 10.53 1.08 F R N 341 261 70 10 8.97 1.23 G M W 1445 722 2 71 452 34.40 1.27 IND 497 271 100 126 17.75 1.19 ISR 209 90 54 65 12.29 1.06 JO R 207 127 28 52 12.18 1.42 K O N 685 614 70 1 10.07 1.43 K O S 725 678 44 3 13.18 1.34 NIG 232 166 36 30 12.21 1.36 P A K 173 146 22 5 6.18 1.06 U A R 939 566 206 167 22.36 1.55 U N K 907 795 103 9 10.55 1.14 U S A 4505 4069 373 63 33.62 1.23 U S R 2224 1615 594 15 24.17 1.36 V T N 925 562 289 74 20.11 1.39 V T S 496 396 64 36 21.57 1.35 328 TABLE 44 R A N K E D D A T A O N SIX INDICATOR VARIABLES G R O U P O N E N ATIO NS 1966-1969 (Aggregated) \TION V O L U M E DIPLOM ATIC E X C H A N G E NON-MIL. CONFLICT MILIT. CONFLICT S C O P E B A LA N C I C H N 4.0 9.0 2.0 8.0 8.0 52.5 C Z E 12.0 11.0 12.0 43.0 15.0 11.0 F R N 9.0 6.0 13.5 55.5 17.0 41.5 G M W 8.0 5.0 19.0 32.0 9.0 30.0 IND 13.0 13.5 13.5 16.0 20.0' 25.5 ISR 3.0 4.0 . 5.0 1.0 1.0 27.5 JO R 10.0 12.0 9.0 3.0 7.0 21.0 K O N 22.0 36.0 18.0 5.0 11.0 9.5 K O S 23.0 26.0 28.0 7.0 13.0 39.5 NIG 18.0 20.0 20.0 11.0 12.0 32.5 P A K 26.0 21.0 35.0 21.0 30.0 9.5 U A R 5.0 7.0 6.0 2.0 4.0 50.0 U N K 7.0 3.0 7.0 17.0 14.0 17.0 U S A 1.0 1.0 3.0 6.0 2.0 25.5 U S R 2.0 2.0 1.0 14.0 3.0 32.5 V T N 6.0 8.0 4.0 4.0 6.0 35.0 V T S 11.0 10.0 15.5 9.5 5.0 31.0 329 TABLE 45 R A I/f D A TA O N SIX INDICATOR VARIABLES G R O U P T W O NATIO NS 1966-1969 (Aggregated) \TI0N V O L U M E DIPLOMATIC E X C H A N G E NON-MIL. CONFLICT MILIT. CONFLICT S C O P E B A LA N C E BE L 90 75 15 0 5.00 2.65 C A M 228 145 75 8 12.67 2.75 C A N 272 245 26 1 6.33 1.83 C U B 175 no 64 1 6.03 1.80 D E N 55 45 8 2 2.75 1.96 FIN 41 41 0 0 4.10 1.41 G M E 281 177 102 2 9.06 1.64 INS 146 122 13 1 1 7.68 1.20 ITA 165 140 22 3 4.02 1.67 L A O 121 80 25 16 8.07 1.78 M A L 94 85 7 2 7.23 .99 N O R 34 27 6 1 2.27 1.70 PHI 156 145 10 1 7.80 1.93 S P N 135 104 29 2 5.19 1.59 S W D 105 90 15 0 5.00 2.23 S Y R 318 186 96 36 9.64 1.44 TAI 139 119 17 3 6.04 1.58 Y U G 227 191 35 1 5.16 1.47 330 TABLE 46 R A N K E D D A TA O N SIX INDICATOR VARIABLES . G R O U P T W O N ATIO N S 1966-1969 (Aggregated) \TION V O L U M E DIPLOMATIC E X C H A N G E NON-MIL. CO NFLICT MILIT. CONFLICT S C O P E B A LA N C I B E L 44.0 42.0 46.5 73.5 42.5 79.0 C A M 19.0 22.5 11.0 18.5 10.0 81.0 C A N 17.0 15.0 31.0 55.5 29.0 65.5 C U B 25.0 31.0 15.5 55.5 32.0 63.0 D E N 58.0 59.0 64.5 43.0 75.0 71.5 FIN 75.0 66.0 83.0 73.5 51.5 37.5 G M E 16.0 19.0 8.0 43.0 19.0 55.0 INS 30.0 28.0 53.0 15.0 24.0 22.0 ITA 27.0 24.0 35.0 32.0 53.0 56.0 L A O 37.5 40.0 33.0 13.0 22.0 62.0 M A L 42.0 39.0 67.0 43.0 25.0 5.0 N O R 81.0 80.5 71.5 55.5 81.0 59.0 PHI 29.0 22.5 62.0 55.5 23.0 70.0 S P N 33.0 32.5 25.5 43.0 38.0 54.0 S W D 39.0 36.0 46.5 73.5 42.5 76.0 S Y R 14.0 18.0 10.0 9.5 18.0 43.5 TAI 31.0 29.0 42.0 32.0 31.0 52.5 Y U G 20.0 17.0 21.5 • 55.5 39.0 45.5 331 TABLE 47 R A W D A T A O N SIX INDICATOR VARIABLES G R O U P T H R E E N ATIO NS 1966-1969 (Aggregated) vnoN V O L U M E DIPLOMATIC E X C H A N G E NON-MIL. CONFLICT MILIT. CONFLICT S C O P E B A LA N C E B R A 80 64 13 3 3.48 1.21 C H T 49 33 13 3 3.27 1.22 IRN 93 89 2 2 3.32 1.18 JAP 293 261 29 3 10.46 1.68 N E W 52 46 6 0 3.71 4.33 R U M 211 194 17 0 5.02 1.49 S W Z 49 45 4 0 2.45 1.40 TAZ 53 4 1 12 0 2.79 1.47 TABLE 48 R A N K E D D A T A O N SIX INDICATOR VARIABLES G R O U P T H R E E N ATIO NS 1966-1969 (Aggregated) \TION V O L U M E DIPLOMATIC E X C H A N G E NON-MIL. CONFLICT MILIT. CONFLICT S C O P E B A LA N C I B R A 49.0 49.0 53.0 32.0 61.5 23.0 C H T 70.0 76.0 53.0 32.0 65.5 24.0 IRN 43.0 38.0 81.5 43.0 64.0 20.0 JAP 15.0 13.5 25.5 32.0 16.0 57.0 N E W 63.5 56.5 71.5 73.5 58.0 83.0 R U M 21.0 16.0 42.0 73.5 41.0 47.5 S W Z 70.0 59.0 77.0 73.5 80.0 36.0 TAZ 60.5 66.0 57.5 73.5 74.0 45.5 332 TABLE 49 R A W D A T A O N SIX INDICATOR VARIABLES G R O U P F O U R NATIO NS 1966-1969 (Aggregated) — 1 I— 1 o V O L U M E DIPLOMATIC E X C H A N G E NON-MIL. CONFLICT MILIT. CONFLICT S C O P E B A LA N C E A R G 5 1 40 8 3 3.19 1.42 A U L 133 118 1 1 4 4.93 2.11 B U L 59 44 14 1 2.95 1.51 C H L 53 40 12 1 4.42 1.96 C O L 50 4 1 6 3 3.57 2.00 C Y P 43 36 7 0 7.17 1.05 G H A 65 52 13 0 3.82 1.38 H U N 86 65 19 2 4.10 1.72 IR Q 160 126 26 8 5.71 1.70 M E X 52 48 4 0 4.33 1.16 M O N 33 16 16 1 3.67 2.54 N T H 66 47 15 4 2.64 1.83 P O L 198 137 59 2 5.66 1.89 S A F 87 65 22 0 4.14 1.98 SIN 45 43 2 0 3.75 .98 T U R 121 90 28 3 5.76 1.49 Z A M 101 66 32 3 5.61 2.24 333 TABLE 50 R A N K E D D A T A O N SIX INDICATOR VARIABLES G R O U P F O U R N ATIO N S 1966-1969 (Aggregated) \TION V O L U M E DIPLOMATIC E X C H A N G E NON-MIL. CONFLICT MILIT. CO NFLICT S C O P E B A LA N C I A R G 66.0 68.5 64.5 32.0 67.0 39.5 A U L 34.0 30.0 60.0 23.5 44.0 75.0 B U L 55.5 61.5 49.0 55.5 71.5 49.0 C H L 60.5 68.5 57.5 55.5 47.0 71.5 C O L 67.5 66.0 71.5 32.0 60.0 74.0 C Y P 74.0 72.0 67.0 73.5 27.0 8.0 G H A 53.0 51.0 53.0 73.5 56.0 34.0 H U N 47.0 47.5 39.0 43.0 51.5 61.0 IR Q 28.0 27.0 31.0 18.5 34.0 59.0 M E X 63.5 54.0 77.0 73.5 48.0 18.0 M O N 82.0 82.0 44.0 55.5 59.0 78.0 N T H 52.0 55.0 46.5 23.5 77.0 65.5 P O L 24.0 25.0 17.0 43.0 35.0 68.0 SAF 46.0 47.5 35.0 73.5 50.0 73.0 SIN 73.0 63.5 81.5 73.5 57.0 4.0 T U R 37.5 36.0 . 28.0 32.0 33.0 47.5 Z A M 40.0 46.0 24.0 32.0 36.0 77.0 TABLE 51 R A W D A T A O N SIX INDICATOR VARIABLES G R O U P FIVE N ATIO N S 1966-1969 (Aggregated) \TION V O L U M E DIPLOMATIC E X C H A N G E NON-MIL. CONFLICT MILIT. CONFLICT S C O P E B A LA N C E A LG 131 95 35 1 4.23 1.70 A U S 56 53 3 0 2.95 1.27 B O L 39 34 5 0 3.90 1.86 G R C 131 104 26 1 5.04 1.13 K U W 40 37 3 0 2.67 2.67 LEB 137 78 34 25 7.21 1.02 M O R 52 46 6 0 2.60 1.30 P E R 68 50 17 1 5.23 1.17 S U D 59 45 13 1 3.11 1.90 T U N 40 3 1 7 2 2.00 1.43 TABLE 52 R A N K E D D A T A O N SIX INDICATOR VARIABLES G R O U P FIVE N ATIO N S 1966-1969 (Aggregated) \TION V O L U M E DIPLOMATIC E X C H A N G E NON-MIL. CONFLICT MILIT. CONFLICT S C O P E B A LA N C I A LG 35.5 34.0 21.5 55.5 49.0 59.0 A U S 57.0 50.0 79.5 73.5 71.5 27.5 B O L 78.0 74.0 75.0 73.5 55.0 67.0 G R C 35.5 32.5 31.0 55.5 40.0 16.0 K U W 76.5 71.0 79.5 73.5 76.0 80.0 LEB 32.0 41.0 23.0 12.0 26.0 6.0 M O R 63.5 56.5 71.5 73.5 78.0 29.0 P E R 51.0 53.0 42.0 55.5 37.0 19.0 S U D 55.5 59.0 53.0 55.5 69.0 69.0 T U N 76.5 78.0 67.0 43.0 83.0 41.5 335 TABLE 53 R A W D A TA O N SIX INDICATOR VARIABLES G R O U P SrX NATIO NS 1966-1969 (Aggregated) DIPLOMATIC NON-MIL. MILIT. NATIO N V O L U M E E X C H A N G E CONFLICT C ONFLICT S C O P E B A LA N C E C O P 73 5 1 19 3 3.48 1.09 E T H 49 44 4 1 2.23 .94 GUI 36 27 9 0 3.27 .92 K E N 54 34 20 0 2.84 3.00 P O R 83 70 13 0 3.95 1.41 R H O 97 74 19 4 8.82 1.04 S A U 89 70 15 4 4.68 1.09 S Y E 52 38 1 1 3 2.48 1.44 V E N 36 29 6 1 3.00 1.57 Y E M 48 32 13 3 3.43 .96 TABLE 54 R A N K E D D A T A O N SIX INDICATOR VARIABLES G R O U P SIX N ATIO NS 1966-1969 (Aggregated) DIPLOMATIC NON-MIL. MILIT. NATIO N V O L U M E E X C H A N G E CONFLICT CONFLICT S C O P E B A LA N C E C O P 50.0 52.0 39.0 32.0 61.5 12.5 E T H 70.0 61.5 77.0 55.5 82.0 2.0 G UI 79.0 80.5 63.0 73.5 65.5 1.0 K E N 59.0 74.0 37.0 73.5 73.0 82.0 P O R 48.0 44.5 53.0 73.5 54.0 3/.5 R H O 41.0 43.0 39.0 23.5 21.0 7.0 S A U 45.0 44.5 46.5 23.5 45.0 12.5 S Y E 63.5 70.0 60.0 32.0 79.0 43.5 V E N 80.0 79.0 71.5 55.5 70.0 51.0 Y E M 72.0 77.0 53.0 32.0 63.0 3.0 336 TABLE 55 R A W D A TA O N SIX INDICATOR VARIABLES G R O U P S E V E N N ATIO N S 1966-1969 (Aggregated) DIPLOMATIC NON-MIL. MILIT. NATION V O L U M E E X C H A N G E CONFLICT CONFLICT S C O P E B A LA N C E ALB 62 34 28 0 4.43 1.82 E LS 32 15 1 1 6 6.40 1.10 LBY 50 43 6 TABLE 56 1 3.13 1.11 R A N K E D D A T A O N SIX INDICATOR VARIABLES G R O U P S E V E N N ATIO N S 1966-1969 (Aggregated) NATIO N V O L U M E DIPLOMATIC E X C H A N G E NON-MIL. CONFLICT MILIT. CONFLICT S C O P E ’.L A N C E ALB 54.0 74.0 28.0 73.5 46.0 64.0 E LS 83.0 83.0 60.0 20.0 28.0 14.0 LB Y 67.5 63.5 71.5 55.5 68.0 15.0 APPENDIX C EXAMPLES OF SITUATION CODING 337 010— WAR March 5r 1955 August 2, 1955 May 10, 1955 January 1, 1967 May 25, 1967 June 6, 1967 338 APPENDIX C CHN gunboats assault Kaoten Is. and are driven back by the Nationalists. (710-223-713/713-223-710) KOS gov't said it would ask the Communists to withdraw from areas of the 38th parallel— there was an implied threat that if withdrawal was not made, KOS would act to dispel Communists. (732-094-731) US planes shot down 2 Communist Korean MIGs in battle over inter national waters. (002-223-731/731- 223-002) De Gaulle urges US to quit the Vietnam War. Asks the end of detest able intervention. (220-102-002/220- 094-002) American warplanes resumed attacks on North Vietnam. (002-223-816) Cairo radio announced that UAR had invaded ISR after having "wiped out" many attacks on Egyptian territory, as war breaks out in the Middle East. (651-223-666) 339 020— CRISIS THERE WERE NO ACTS OF CRISIS IN 1955. May 27, 1967 May 28, 1967 June 1, 1967 USR said it was considering joining other, nations to end crisis in Middle East. (365-025-999) UAR deploying troops from Yemen to Sinai Peninsula. (651-182-666) US proposed passage of "breathing spell" resolution for crisis in Middle East. (002-101-399) 030— FORMAL INTERNATIONAL NEGOTIATIONS 031— About War January 26, 19 55 February 1, 1955 January 1, 1967 USR formally ended state of war between USR and Germany. (365-081- 255/365-081-265) UN Security Council invites CHN to help end hostilities in offshore islands— invites CHN to send a representative. (399-092-710) President Johnson said he will be very glad to do more than our part to meet Hanoi more than halfway in any possible ceasefire truce or peace conference negotiations. (002-025-999) 032— About Crisis THERE WERE NO NEGOTIATIONS ABOUT CRISIS IN 1955. May 29, 1967 Levi Eshkol said ISR would continue to negotiate with the major powers to find ways to reopen the Strait 340 of Tiran to ISR shipping. (666-025- 999) May 29, 1967 Nasser said that negotiated peace in Middle East is out of the question until Palestinian Arabs returned to their homeland now what is part of ISR and exercise sovereignty there. (651-112-666/651-150-666) 033— About International Incidents/Disputes April 16, 1955 USR and AUS expressed hope that USA, FRN, UNK will join in accord of treaty. (365-024-002/365-024-200/365- 024-220/305-024-002/305-024-200/305- 024-220) January 1, 1955 USR's Malenkov says he'd welcome diplomatic negotiations to settle disagreements in the Far East. (365-094-999) July 19,.1955 Eden presented today 3 proposals intended to make German unity palat able to the USR. (200-101-999) January 17, 1967 ISR and SYR accepted UN proposal to meet in effort to avert full-scale war. (399-101-666/399-101-652/666- 082-399/652-082-399) 034— About Trade or Monetary Policy June 1, 1955 USR and JAP diplomats met in Japan to prepare ground for peace and trade negotiations. (365-032-740/740-033- 365) March 29, 1967 341 20 nations of Latin America have agreed in principle to create a full functioning common market no later than 1980. (998-082-998) Johnson arrived in Punta del Este for meeting on Latin American common market. (002-031-998/998-031-002) US gov't disclosed that it is planning to cease regular regulations applying to exports of nonstrategic goods to the Soviet Union and its satellites in Europe in a gesture linked to negotiations at the Foreign Ministers Conference. (002-065-998) 035— About Future Diplomatic Relations/Recognition April 12, 1967 November 4, 1955 January 16, 1955 February 18, 1955 April 27, 1955 February 28, 1967 USR renewed demands for 4-power talks on reuniting Germany and offered to "normalize" relations with GMW conditional on rejection of Paris accords for arming GMW. (365-150- 002/365-150-200/365-150-200/365-150- 220/365-101-255) Churchill says once the Paris agreements are ratified, UNK can talk to USR. (200-025-365) USR agreed that Big 4 ambassadors in Vienna should meet Monday to prepare final terms for an early restoration of AUS independence. (365-081-002/ 002-081-365/365-081-200/200-081-365/ 365-081-220/220-081-365) Goldberg, UN delegate, obtained agreement today under which Tokyo will consult with Washington regu larly, in advance on questions O r > w . » . . '" • ' , f 342 likely to come before the General Assembly. The agreement apparently puts Japan in the same diplomatic partnership basis as between the USA, CAN, and UNK. (002-081-740/740-081- 002) March 2, 1967 POL and CZE signed a 20-year treaty of friendship and military assist ance. (351-081-290/290-081-365) April 18, 1967 In response to appeal by GMW for improvement in relations between GME-GMW, Ulbricht proposed a meeting between Chancellor Kiessinger and Premier Stoph to "negotiate an understanding between the two German states." (255-095-265/265-101-255) 036— About Disarmament (including Nuclear Weapons and Space) June 15, 1955 Adenauer urges USA and USR to advance disarmament at the Big 4 conference. (255-102-002/255-102- 365) July 22, 1955 New USR proposal for nonaggression agreement between N. Atlantic and E. European military blocs was pre sented today at Big 4 conference— at same time USR put forth abbrev iated version of its UN plan for reducing armaments and banning nuclear weapons. (365-101-200/365-101-220/ 365-101-002) October 20, 1955 US and USR started negotiations to draft an atoms-for-peace resolution. (002-031-365/365-031-002) 343 January 28, 1967 USR, USA, and 60 other nations signed a treaty to limit military activities in space. (998-081-998) February 18, 1967 President Johnson expressed high hopes today for an agreement on a treaty to ban the spread of nuclear weapons, but he cautioned that the US and USR were being drawn into another futile escalation of an atomic arms race. (002-023-999/002-024-999) May 3, 1967 Reopening of Geneva Disarmament Conference was postponed because of disagreement in talks between US-USR— USR disapproved of US version of treaty and points of arms inspections. (365-191-998/002-191-998/365-121-002) June 24, 1967 Johnson and Kosygin talked for 5 hours in Glasboro, New Jersey about arms control. (002-031-365/365-031- 002) 037— About External Dominance April 15, 1955 AUS and USR have agreed to sign a state treaty ending a 10-year occupa tion of AUS. (305-081-365/365-081- 305) August 27, 1955 Last-minute maneuvers threatened to hold up an accord worked out between French Moroccan leaders. Faure appeared optimistic. (220-024-999) January 29, 1967 UNK in Macao signs agreement accept ing all CHN demands of last month. (200-081-710/710-081-200) o 344 November 16, 1967 USA agrees to begin talks immediately with JAP on return of some Pacific Islands captured by USA in World War II in a joint communique issued by Sato and Johnson. (002-082-740/740- 082-002) 040— INTERNATIONAL INCIDENTS/DISPUTES 041— Land May 9, 1955 PAK and IND clash over Kashmir near border of Kashmir and W. Pakistan— the first serious incident since the truce of Jan. 1949. (770-223-750/750- 223-770) September 1, 1955 ISR's army struck at Egyptian mili tary camp in Gaza last night— military camp served as a base for terrorist squads that had killed 15 Israelis inside ISR. (666-223-651/ 651-223-666) Red China has withdrawn troops from disputed Himalayan area on India- Tibetan border. (750-012-710/710- 012-750) ISR tanks fire on Syrians in growing conflict at border. (666-223-652) Chinese troops in Tibet opened fire across Himalayan pass on IND army position in Sikkim— spokesman said shelling was heaviest for a long time. (710-223-750) November 10, 1955 January 10, 1967 September 12, 1967 Sir-, . " v i . 042— Sea 3 S5 O r> o • V • * July 11, 1955 Determination of ISR to obtain free entry through the Gulf of Aqaba to be allowed by armed force if neces sary, became ISR governmental policy today. (666-025-999) April 18, 1955 UNK told CHN that it doesn't accept any responsibility for crash of air liner in South China Sea last week. (200-112-710) December 31, 1955 Breach by Syria in Galilee (fighting with ISR). (666-223-652/652-223-666) May 12, 1967 US destroyer Walker scraped today for second time by USSR ship in Sea of Japan. US called on USR to halt harassment of American naval vessels— delivered severe protest. (365-221- 002/002-132-365) May 23, 1967 Nasser announced UAR would not allow ISR ships to pass through Gulf of Aqaba— ISR's only outlet to South and East. He announced this on eve of U Thant's arrival. (651-112-666) 043— Air August 4, 1955 Bulgaria acknowledged that 2 of her fighter planes shot down an ISR air liner. (355-013-666) August 30, 1955 Egypt and ISR jet fighters met over Gaza frontier— believed to be first such encounter since signing of armistice in 1948. (666-223-651/651- 223-666) . * ■ ' ' ' ' ^ ’ " '6 bet)' vut>w ^ 0 « „ ° f * c A - ' - * 0 «... ° _ O O 346 ' r . ' o - n June 30, 1967 Comm. China stepped up its pressure on UNK in HOK with charges British aircraft from HOK had recently violated Chinese air space. (710- 121-200/710-192-720) August 22, 1967 Pentagon said 2 US jets shot down over Comm. China after having veered off course. Official CHN news agency said planes had "flagrantly intruded into CHN Airspace in act of irrevocable provocation." (710-223- 002/710-121-002) September 5f 1967 In mid-August, airfield at Kanow, N. Nigeria, was suddenly closed to commercial flights by Nigeria federal authority— reason became clear when soon 15 Soviet transports landed there and unloaded crates. (365-072- 475) 044— Other or Combination September 2, 1955 ISR announced there would be no further military action against Egypt provided that all attacks by Egyptians in whatever form will cease. (666- 025-651) 050— INTERNAL INCIDENTS 051— Espionage ■ February 28, 1955 2 Amer. students imprisoned by CHN arrive in Hong Kong free and full of praise for CHN, and saying they had been spies for US. (710-066-002) _ r- ' • r> ''34-7- .. September 15, 1955 Australia unfolded a 100,000-word story of far-reaching Soviet espionage imperilling security of whole western world. A Royal Commission on Soviet Espionage reported, "the principal target had been the Dept, of External Affairs." (900-121-365) February 5, 1967 US citizen sentenced to 8 years in prison on grounds of subversion in CZE was expelled and returned to US. (315-066-002/315-201-002) March 24, 1967 For. Minister Chagla told Parliament the Indian gov't would conduct "thorough inquiry" into activities of Amer. CIA in India. (750-025-002) July 9, 1967 KOS-CIA claims arrest of 70 members of Communist espionage network organ ized by KON intelligence in EBE begin ning in 1958. Included in those arrested were 16 students brought home from GMW and others. (732-212- 731) 052— Involving Embassies, Consulates, Official Representatives June 18, 1955 USA withdraws 3 army officers from USR embassy at USR request. (002- 011-365/365-094-002) September 18, 1955 British institute building in center of Cyprus set afire as anti-British riots broke out in support of demands for union with Greece. (352-221-200) January 30, 1967 YUG embassies in USA and CAN bombed; condemend by USA and State Depart ment. (002-221-345/020-221-345/002- 025-999) S3---- O 348 February 2, 1967 CHN demonstrators pulled FRN embassy official from his car outside mission in Peking and forced him to stand in below-freezing cold for nearly 7 hours. (710-181-220) February 9, 1967 Soviet loudspeakers, in Moscow blared complaints at curtained windows of CHN embassy today while Soviet For. Minister accused diplomats inside of rudeness. (365-131-710/365-122- 710) 053— Private Citizens (Non-Espionage) January 2, 1955 February 24, 1955 March 2, 1955 January 31, 1967 February 6, 1967 April 4, 1967 USR offers to free 2 US citizens long held in USR. (365-101-002) CZE skaters asks US asylum. 094-002) (315- USR orders US Roman Catholic priest to leave USR. (365-201-002) VAT meets with USR President to . discuss disappearance of Catholics from USR. (328-031-365/365-031-328) Hundreds of slogan-chanting CHN demonstrators delayed for hours today the departure of Soviet air liner carrying 30 women and 50 chil dren home. (710-181-365/710-212-365) US State Dept, expressed concern that VTN might be brainwashing Amer. prisoners to make propaganda state ments attacking US policy. (002-131- 816) o ca> 349 060— TRADE OR MONETARY POLICY (ECONOMIC) March 2, 1955 October 12, 1955 November 23, 1955 April 1, 1967 October 28, 1967 November 20, 1967 US will set up special fund for economic development in Asia and will look to countries of the area for lead on how mutual aid plans should be organized. (002-071-998) Canada and USR agreed to accord each other "most-favored-nations" privi leges in their trade— means each country give same treatment it accords its most favored trading partner. (020-081-365/365-081-020) Kingdom of Yemen gave US company first oil and mineral concession in nation's history. (678-064-002) British gov't decided to go ahead with purchase of US F-lll's. They put in order for 40 planes. (200- 093-002) Mex. President Ordaz said USA con gress should reflect on possible world repercussions before adopting restrictive trade legislation, adding his voice in rising protest from nations that depend largely on US trade. (070-160-002) FRN and other common market members agreed to contribute to $1.4 billion loan UNK is seeking from IMF. (397-081-200/200-081-397) 070— FORMAL DIPLOMATIC RELATIONS July 1, 1955 GMW accepted Soviet invitation to discuss establishment of diplomatic relations. (255-0 8 2-36 5/365-082^ 255), . , .................... n * .. r . v , ' 350 r . * • October _23v 1955 n Relations between US and ITA were discussed by Dulles and ITA Giovanni Granchi. (325-031-002/002-031-325) September 14, 1955 GMW and USR agreed to establish diplomatic relations on basis and understanding that German prisoners in Russia will be freed. (255-081- 365/365-081-255) January 6, 1967 Spain and Rumania signed in Paris an agreement establishing full con sular and commercial relations. (230-081-360/360-081-230/230-064- 360/360-064-230) February 1, 1967 GMW and Rumania agreed to establish full diplomatic relations. (360-081- 255/255-081-360/255-064-360/360-064- 255) August 5, 1967 Editor of Al Ahram said it essential that Egypt have diplomatic relations with US. (651-102-002) November 1, 1967 CHN withdraws economic aid mission and brings diplomatic staff home from INS. (710-192-775/710-192-850) 080— DISARMAMENT (INCLUDING NUCLEAR AND SPACE February 19, 1955 USR proposes all nations should freeze strength of their armed forces and destroy all nuclear weapons. Also called for international con ference on disarmament to be held under UN this year. (365-094-999/ 365-101-999) April 14, 1955 Eisenhower-'approved agreement between NATO powers to siiate-'-'lrifoma^i^n^bh^'' atomic weapons. (002-042-998) May. 28, 1955 August 12, 1967 October 11, 1967 > I ’ , * r ^ ' ‘ V f r * i n ' r . Adenauer calls on US to take ini tiative in proposing general and controlled disarmament. (255-094-002) Foster, Chief US Disarmament Nego tiator, announced USR had advised him they might be ready to join US in presenting to NATO conference a draft treaty to prohibit spread of nuclear weapons. (002-023-999) Johnson Administration fearful that dilatory tactics among allies in W. Europe could indefinitely postpone or even scuttle proposed international treaty to prevent spread of nuclear weapons. (002-121-998) 090— INTERNATIONAL EXCHANGES 091— Educational/Cultural March 11, 1955 July 30, 1955 November 11, 1955 US agrees to allow 11 editors of USR student/youth publications to visit US. (002-081-365/365-081-002) COMM. China theatrical company and US theatrical company have agreed to appear in each other’s country. (710-081-002/002-081-710) State Dept, beat retreat on cultural exchanges with USR putting a “hold for consideration" classification for Soviet proposal to send second group of agricultural experts to US. (002-191-365) March 2, 1967 or Pravda accused US-CIA of having thrown a "sinister shadow" over by letting it send's ^f^^£^L^I^/; ///? . v (002-121-365) 352 092— Technological 'Mnoluding Nuclear andrSpace)r £.0 ^ , ? n ^ f V'.'U A’V p r" ' , i ft r'-f'CS.p March 11, 1955 November 20,. 1955 March 12, 1955 December 20, 1967 100— AID 101— Nonmilitary February 14, 1955 October 21, 1955 November 17, 1955 USR asked if US willing to exchange visits of agriculture delegations between 2 countries. (365-091-002) Bulganin said USR ready to share its experience in industrial and atomic energy matters with India. Bulganin visiting there. (365-053-750) US will allow USR farmers to visit Iowa. (002-082-365) ISR and RUM agreed to set up joint industrial ventures within framework of an economic and scientific pact between the two countries. (666-081- 360/360-081-666) vox*#1 * , SPN has made clear that it is dis satisfied with the amount of economic aid it has received from the US. US says ready to contribute substan tially to financing of Aswan High Dam in Egypt; also willing to aid unified development of Jordan River Valley. (002-025-651) A request from India for allowing as much as $1.5 billion is considered distinct possibility as policy among high US officials. India may.ohave, sto’ ' 3 . .e- ^ -yp- as o f l ' o a ‘ to . reach this sum:£Lrra'foreign exchange wif^sne is to begin her second 5-year ’ 'plan next April. (750-093-002) 353 ^ - ■-ix.-t-'.' • - - -n- April 6© 1967 «„ . . c . . * , * S - - I . - „ Sxgnxf xc„ant ^ xncreas es xn fxnancxal arid technical support for population control programs in developing coun tries are being made by US and UN agencies. (002-073-998/399-073-998) May 12, 1967 McNamara warned Greek military junta that US may cut off aid unless GRC returned to constitutional government. (002-172-350) September 1, 1967 Arab leaders agreed to aid the UAR and JOR because of June War losses. (998-081-651/998-081-663/651-081- 998/663-081-998) October 25, 1967 USA to expand Food for Freedom pro gram in underdeveloped nations under pressure of sagging farm prices in US and sharp cuts in foreign aid funds. (002-025-999) 102— Military January 16, 1955 USA proposed giving military aid to COS if OAS approved COS request for aid. (002-101-094) April 13, 1955 UAR offers military support to AFG in PAK on her western allies threat ened aggressive interference in AFG demand for a plebiscite in northwest frontier province. (651-052-700) November 17, 1955 Nasser warned that US arms shipments to ISR would plunge mideast into armaments race. (651-160-999) & February^ , ' ^19 6 7"^ " USR signed military aid agreement with Irafi worth^nearly $100 million. (365-081-630/630-081-365),, ir „ , i ) ,0 .-- or, © July 19, 1967 September 24, 1967 354 ISR Ambassador to US, Harman, warned of one-sided arms race in his plea for US military support. (666-160- 002/666-093-002) USR concludes series of agreements with VTN providing for continuing deliveries of military and economic aid to Hanoi in 1968— joint commu nique issued at end of month of talks specified that military material include ground-to-air missiles, planes, and artillery. (365-081- 816/816-081-365) 110— EXTERNAL DOMINANCE July 1, 1955 July 26, 1955 November 19, 1955 June 15, 1967 British gov't invited GRC and TUR for talks on status of Cyprus. (200-062-350/200-062-640) India ordered Portuguese to close their delegation in New Delhi. Nehru said move was taken because Portuguese refused to talk about Goa— India has been demanding that the colony be integrated into the Indian republic. (750-201-235) Sultan Mohammed ben Youssef announced today the 43-yr. old French protec torate in Morocco was coming to an end and a "new era of freedom and independence" dawning. (600-025-999) British gov't announced today it would ask people of Gibraltar to declare whether they want to remain associated with UNK or become part of Spain. (200-025-999) 355 September' 11 "j:\13,67 GRC and TUR ended their conference on;'.Cyp^us^and made little effort to concda-1 1 thelriia%re^ment .*£*$ 3.5fi,^Q^ 1*^5.^ 640/640-031-350) 120— GENERAL DIPLOMATIC ASSESSMENTS May 6, 1955 September 27, 1955 January 1, 1967 May 14, 1967 December 3, 1967 Dulles said "small but perhaps sig nificant signs" suggest more peace able attitudes by Communist powers. (002-024-999) Canada called on general assembly to respond to effort of great powers to relax international tension. (020-094-998) De Gaulle hailed easing of tensions in Europe. (220-041-998) Pope Paul VI says world is in danger and urged men to be worthy of divine gift of peace. (328-022-999/328-102* 999) Dean Rusk says US would like to see European allies develop common posi tion so they could win western alliance as equal partners in quest for peace. (002-102-998) APPENDIX D SITUATION D A T A A C R O S S W EIS C O M B E V E N T S ALL AC TO R S, U S A & U S R 1955, 1967, & 1955-1967 C O M B IN E D c: C ’ c'' t .© <y c.‘ C C cr * c 1 2 3 4 SITUATION 5 TABLE 57 ALL AC TO R S- D A T A A C R O S S 6 7 1955 W EIS C O M B E V E N T S 8 9 10 1 1 12 : ^ - > j ] TO TA LS 1 5 0 7 14 0 0 0 0 0 0 0 0 26 2 19 0 45 45 13 3 4 13 0 13 4 84 243 3 8 0 47 30 3 2 6 6 0 3 2 8 115 4 0 0 13 3 2 0 2 4 1 0 0 17 42 5 1 0 4 7 0 3 0 3 2 3 1 12 36 6 0 0 1 10 13 2 3 0 1 0 4 1 35 7 1 0 1 2 0 1 0 0 0 5 0 0 10 8 3 0 97 34 7 9 4 5 7 12 4 28 210 9 7 0 28 28 6 1 0 1 1 2 5 1 14 103 10 8 0 43 29 3 1 0 17 0 0 1 1 1 113 1 1 3 0 24 -15 5 0 1 7 0 1 4 7 67 12 5 0 4 19 3 0 0 5 0 2 2 19 59 13 1 0 4 6 0 0 0 0 0 0 0 0 1 1 14 K 0 1 2 1 0 0 1 0 0 0 1 3 7 15 4 0 4 3 3 2 0 0 0 0 2 3 3 2 1 16 4 c c . •o 0 6 1 0 0 0 0 4 0 6 2 1 17 3 c : 0 0 0 0 0 1 0 0 0 0 0 4 18 2 ' 0 0 4 3 0 0 0 1 0 3 0 13 19 0 c- 0 2 1 0 0 0 0 1 0 1 0 5 20 0 0 0 0 3 0 0 0 0 0 1 0 4 2 1 0 0 0 2 7 0 0 0 0 0 0 0 9 22 26 \0 4 69 4 0 0 0 0 2 1 2 108 w TO TAL 101 0 329 329 77 24 2 1 72 15 50 3 1 213 U1 1262 r © Cf • » c • c • - TABLE 58 - c r ALL AC TO R S-1967 } L ; r < ? .r SITUATION D A TA A C R O S S W EIS C O M B E V E N T S 1 © i C ' 3 4 5 6 7 3 9 10 12 TO TALS 1 3 0 8 1 0 0 0 0 0 0 0 12 2 63 7 V 5 1 41 9 4 2 7 0 6 1 24 215 3 25 2 \ 34 3 1 7 2 0 0 0 4 2 5 112 4 8 7 9 0 1 0 3 0 0 0 02 30 5 6 4 e . 3 3 0 0 0 1 0 1 0 0 18 6 13 0 5 2 1 0 4 0 0 3 0 3 3 1 7 3 l c 2 9 0 0 0 0 0 6 0 0 2 1 8 12 3» . 49 16 0 6 8 5 2 16 0 0 117 9 13 0 f 20 14 5 3 0 2 0 8 0 6 7 1 10 9 i ; 45 - 14 0 3 2 3 0 r * T 0 5 86 n 1 1 i , 24 10 1 3 1 0 0 0 0 1 52 12 20 2 40 26 6 1 1 1 1 1 0 7 106 13 8 °c 1 12 7 0 0 0 0 0 0 0 28 14 1 o( - 1 4 0 0 0 0 0 0 0 0 6 15 4 2 12 9 1 0 1 0 0 0 t o 0 29 16 4 3j 3 10 0 2 1 0 0 1 0 1 25 17 3 0 3 4 1 0 1 0 0 1 0 0 13 18 1 ° 1 1 12 1 1 0 0 0 0 0 0 0 26 19 6 1 8 5 1 3 3 0 0 2 0 0 29 20 0 0 0 0 6 0 0 0 0 0 0 0 6 2 1 5 1 0 9 8 0 0 0 0 0 0 23 w 22 234 4 o 0 • 43 8 0 0 0 0 0 0 0 289 £ CO TO TAL 452 33 309 291 73 28 24 22 3 53 3 54 1345 c i . TABLE 59 ALL ACTORS-1955 and 1967 SITUATION D A T A A C R O S S W EIS C O M B 'E V E N T S 1 2 c 3 4 5 6 7 8 9 10 1 1 12 TO TA LS 1 8 0 7 22 1 0 0 0 0 0 0 0 38 2 28 7 96 86 22 7 6 20 0 19 5 108 458 3 33 2 81 61 10 4 6 ■ 6 0 7 4 13 227 4 8 0 20 12 2 1 2 7 1 0 0 19 72 5 7 4 7 10 0 3 0 4 2 1 4 1 12 54 6 13 0 6 12 14 2 7 0 1 3 4 4 66 7 4 1 3 1 1 0 1 0 0 0 1 1 0 0 3 1 8 15 3 146 50 7 15 12 10 9 28 4 . 28 327 9 20 0 48 42 1 1 4 0 13 2 13 1 20 174 10 17 1 88 43 3 4 2 20 0 4 1 16 199 1 1 14 1 48 25 6 3 2 7 0 1 4 8 119 12 25 2 44 45 9 1 1 6 1 3 2 26 165 13 9 0 5 18 7 0 0 0 0 0 0 0 39 14 2 0 2 6 1 0 0 1 0 0 0 1 13 15 8 2 16 12 4 2 1 0 0 0 2 3 50 16 8 3 3 16 1 2 1 0 0 5 0 7 46 17 6 0 3 4 1 0 2 0 0 1 0 0 17 18 3 1 1 16 14 0 0 0 1 0 3 0 39 19 6 1 10 6 1 3 3 0 1 2 1 0 34 20 0 0 0 0 9 0 0 0 0 0 1 0 10 21 5 1 0 1 1 15 0 0 0 0 0 0 0 32 22 260 4 4 12 12 0 0 0 0 2 1 2 397 TO TA L 553 33 638 620 150 52 45 94 18 103 34 267 2607 00 < J 1 VO 1 2 3 4 SITUATION 5 TABLE 60 USA-1955 D A T A A C R O S S 6 7 W EIS 8 C O M B E V E N T S 9 10 1 1 12 TO TA LS 1 1 0 0 0 0 0 0 0 0 0 0 0 1 2 10 0 12 10 3 1 0 9 0 4 0 37 86 3 4 0 12 6 1 2 3 1 0 2 0 3 34 4 0 0 5 1 2 0 1 2 0 0 0 10 2 1 5 0 0 2 3 0 1 0 1 0 1 0 5 13 6 0 0 1 4 5 0 0 0 0 0 0 0 10 7 1 0 0 2 0 1 0 0 0 2 0 0 6 8 1 0 14 8 3 1 0 1 3 5 0 6 42 9 3 0 4 2 3 0 0 3 1 1 0 3 20 10 2 0 3 2 0 1 0 9 0 0 0 6 23 1 1 0 0 5 2 4 0 0 1 0 0 1 2 15 12 0 0 0 3 1 0 0 1 0 0 1 3 9 13 0 0 0 1 0 0 0 0 0 0 0 0 1 14 1 0 0 0 0 0 0 0 0 0 0 0 1 15 0 0 0 1 3 1 0 0 0 0 0 0 5 16 2 0 0 0 0 0 0 0 0 1 0 0 3 17 0 0 0 0 0 0 0 0 0 0 0 0 0 18 1 0 0 1 0 0 0 0 0 0 0 0 2 19 0 0 0 0 0 0 0 0 1 0 0 0 1 20 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 22 5 0 0 0 0 0 0 0 0 0 0 0 5 TO TAL 3 1 0 58 46 25 8 4 28 5 16 2 75 360 00 C T l CM e o TABLE 61 ‘‘ USA-1967 SITUATION D A T A A C R O S S W EIS C O M B E V E N T S 1 2 © V 3) 0 4 5 6 7 8 9 10 3 ) 1 1 12 TO TALS 1 2 0 Ov 5 1 0 0 0 0 0 0 0 8 2 44 1 29 1 1 3 0 1 4 0 6 0 O 5 104 3 9 0 5 e 6 0 0 0 0 0 2 0 Q 3 25 4 5 0 2 2 0 0 0 2 0 0 0 0 1 1 5 2 0 2 " 1 - 0 0 0 1 0 : 1 J 0 0 7 6 6 0 0 0 1 0 0 0 0 2 0 1 10 7 2 0 0 « 3 0 0 Q 0 0 3 0 0 8 8 2 0 7 3 0 1 0 1 0 3 0 0 17 9 2 0 9 0 1 0 0 2 0 1 0 0 15 10 4 1 14 7 0 1 0 2 0 j 2 0 2 33 1 1 3 0 0 , 1 0 0 0 0 0 0 0 0 4 12 4 0 3 v 3 0 0 0 1 1 0 0 0 12 13 5 0 1 3 2 0 0 0 0 0 0 0 1 1 14 1 0 0 1 0 0 0 0 0 J 0 0 0 2 15 0 0 2 0 1 0 0 0 0 0 0 0 0 2 16 1 0 1 2 • 0 0 0 0 0 0 0 0 4 17 1 0 2 0 c 0 0 0 0 ‘ J 0 1 0 0 4 18 1 0 0 0 0 0 0 0 0 0 0 0 1 19 0 0 1 0 *1 3 _ 0 0 0 1 0 0 6 20 0 0 0 0 0 0 ' 0 0 0 0 0 0 0 0 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 22 126 0 0 0 1 0 0 0 0 0 0 0 127 TO TAL 220 2 78 48 10 5 1 13 1 22 0 1 1 411 TABLE 62 USA-1955 and 1967 SITUATION D A T A A C R O S S W EIS C O M B E V E N T S 1 2 3 4 5 6 7 . 8 9 10 11 12 TO TA LS 1 3 0 0 I O < ; 5 1 0 0 0 0 0 0 0 9 2 54 1 41 ? 21 6 1 1 13 0 10 0 42 190 3 13 0 17 I 12 1 2 3 1 0 4 0 6 59 4 5 0 7 ; 3 2 0 1 4 0 0 0 10 32 5 2 0 4 ' 4. 0 1 0 0 2 0 5 20 6 6 0 1 i A 6 0 0 0 2 0 1 20 7 3 0 0 ; 5 0 1 0 0 0 5 0 0 14 8 3 0 21 11 3 2 0 2 3 8 0 6 59 9 5 0 13 2 o 4 0 0 5 ■ - . 1 2 0 3 35 10 6 1 17 9 0 2 0 11 0 2 0 8 56 11 3 0 5 3 ; 4 0 0 1 0 0 1 2 19 12 4 0 3 6 1 0 0 2 1 0 1 3 21 13 5 0 1 4 c 2 0 0 0 0 0 0 0 12 14 2 0 0 1 0 0 0 0 0 0 0 0 3 15 0 0 2 1 c 3 1 0 0 0 0 0 0 7 16 3 o • 1 2 0 0 0 0 0 1 0 0 7 17 1 0 2 0 0 0 0 0 0 1 0 0 4 18 2 0 0 1 1 0“ 0 0 0 0 0 0 0 3 19 0 0 1 0 1 3 0 0 1 1 0 0 7 20 0 0 0 0 0 0 0 0 0 0 0 0 0 21 0 0 0 0 0 0 0 0 0 0 0 0 0 22 131 0 0 0 1 0 0 0 0 0 0 0 132 TO TAL 251 2 136 94 35 13 5 41 6 38 : 2 86 709 TABLE 63 363 CO E o r > » r > . i — in co i — coi— oinmoi— co oo o o o co CM OO CM OO r — r — o co i — C\J C \ J OlOOCOi- I — oco^tr- Or— O O i — i — O O O O O O CVJ i — r — i— ' = = 1 - o o o o o o o o o o o o o o o o o o o o o o o O O O O O l — Or— O O O O O O O O O O O O O O O CM CO I — 2 LU QQ s o o CO t —r LU 3: LO LO 00 cr» oo I — o I DC DC O oo o H- <c o CTl O O O O C M i — O I — I — o o o o o o o o o o o o o uo CO o CM ^ CM I — o O I — COUOr— CMOr— O O O O O O O O CM CM fN. O O O l — Or— O C M O O O O O O O O O O O O O O Nj- CO O C M O O i — O O CM i — O O O O O O O O O O O O O CO o I —I h— «c UO O CM i — O O CM O CM i — i — i — i — O O O O O O Oi— i — O CO CO *3- O CO N O i — CM O CM CM I — O O i — CM O O O O O O CO CO CO SCDCTlr- 1— OOOr-Nl'N!— C M O O C O O O O O O O O N t - CM i — i — 00 CM O O O O O O O O O O O O O O O O O O O O O O o Or— O O O O O O i — CM O CM O O i — O O O O O O O Cn cMco<^incoiNCOO^Or-CMco<tincoiNOOcriOr-CM «a; i — i — r — r— i — i — r— i — r — r-CMCMCM I — O TABLE 64 USR-1967 SITUATION D A T A A C R O S S W EIS C O M B E V E N T S 1 2 3 4 5 6 7 8 9 10 1 1 12 TO TALS 1 0 0 0 0 0 0 0 0 0 0 0 0 0 2 1 1 3 1 0 1 0 1 0 0 0 2 10 3 2 0 5 3 3 . 0 0 0 0 1 0 0 14 4 0 0 0 3 0 0 0 0 0 0 0 0 3 5 2 0 0 0 0 0 0 0 0 0 ■ 0 0 2 6 0 0 0 0 0 0 0 0 0 0 0 0 0 7 0 0 0 3 0 0 0 0 0 2 0 0 5 8 1 0 2 2 0 0 0 2 . 0 3 0 0 10 9 2 0 0 3 1 1 0 0 0 0 0 1 8 10 1 0 4 1 0 1 0 1 0 0 0 1 9 1 1 0 0 2 0 0 0 0 0 0 0 0 0 2 12 5 1 10 6 3 0 0 0 0 0 0 2 27 13 2 0 0 2 2 0 0 0 0 0 0 0 6 14 0 0 0 0 0 0 0 0 0 0 0 0 0 15 0 1 5 2 1 0 0 0 0 0 0 0 9 16 2 0 0 4 0 0 0 0 0 0 0 0 6 17 2 0 0 0 1 0 0 0 0 0 0 0 3 18 0 0 0 1 0 0 0 0 0 0 0 0 1 19 1 0 2 0 0 0 0 0 0 0 0 0 3 20 0 0 0 0 1 0 0 0 0 0 0 0 1 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 22 0 0 0 2 0 0 0 0 0 0 0 0 2 TO TA L 2 1 3 33 33 12 3 0 4 0 6 0 6 121 TABLE 65 USR-1955 and 1967 SITUATION D A T A A C R O S S W EIS C O M B E V E N T S 1 2 3 4 1 0 0 7 0 2 2 1 9 9 3 2 0 14 10 4 0 0 1 3 5 2 0 1 1 6 0 0 0 2 7 0 0 0 3 8 1 0 22 4 9 3 0 7 6 10 3 0 2 1 5 1 1 0 0 13 2 12 7 1 12 7 13 2 0 0 2 14 0 0 0 0 15 1 1 8 3 16 2 0 0 6 17 2 0 0 0 18 0 0 0 1 19 1 0 2 0 20 0 0 0 0 2 1 0 0 0 0 22 0 0 0 2 TO TAL 28 3 117 66 5 6 7 8 9 0 0 0 0 0 2 3 0 3 0 4 0 0 4 0 0 0 1 2 0 0 1 0 1 2 2 0 1 0 1 0 0 0 0 0 2 2 2 3 1 2 2 0 3 1 1 1 0 6 0 1 0 0 1 0 4 0 0 2 0 2 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 1 0 0 0 0 0 0 0 0 0 25 9 4 26 5 10 1 1 12 TO TALS 0 0 0 7 0 0 18 47 1 0 0 35 0 0 3 10 1 0 1 10 0 0 1 7 3 0 0 6 3 0 3 43 0 0 5 29 0 0 2 39 0 0 0 17 0 0 13 46 0 0 0 6 0 0 0 1 0 0 1 15 0 0 1 9 0 0 0 3 0 0 0 1 0 0 0 3 0 0 0 2 0 0 0 1 0 0 0 2 8 0 48 339

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

An Analysis Of United States International Relations Before And During Limited War

PDF

Setting The Pattern Of Contemporary International Relations In The Middleeast: 1953-1958

PDF

Patterns In Political Violence And War, 1751-1960

PDF

Structural Balance And The International System: The Middle East Conflict, 1967

PDF

The Relationship Between The Development Of Strategic Nuclear Weapons Systems And Deterrence Doctrine In The Soviet Union And Communist China

PDF

Political Development And Political Parties In Turkey

PDF

Towards The Development Of An International Regime For The Seabed Beyond The Limits Of National Jurisdiction: A Comparative Study Of Four Proposals. (Volumes I And Ii)

PDF

Prediction Of Therapeutic And Intellectual Potential In Mentally Retardedchildren

PDF

The Effectiveness Of A Governmental Agency In Its Relations With The Public: An Ecological Analysis

PDF

Role Conflict Resolution In Complex Social Systems: A Polyarchical Systems Model For Organizational And Social Change

PDF

Hierarchy And Market: A Theory Of Alternation And Mutual Modification, With Special Attention To Revolving Funds

PDF

An Economic Analysis Of The Problems Of Federal Debt Management

PDF

The Impact Of The "Miranda" Decision On Police Procedures And Morale

PDF

Exchange Rate Policy And Export Performance

PDF

Developmental Planning In The United Arab Republic: An Economic Analysis And Evaluation Of The Egyptian Experience

PDF

Technical Assistance Programs Of The United Nations And Of The United States: A Comparative Study

PDF

The Economics Of Foreign Aid: The Uar Experience With The United States And Ussr Programs, 1952-1965

PDF

Role-Taking And Performance In A Complex Organization: A Preliminary Investigation

PDF

Management Of Change In Organizations--A Process Analysis

PDF

Toward A Non-Bureaucratic Organization For Development Assistance: Insights From A Study Of Aid Technical Advisers

Asset Metadata

Creator Young, Robert Arthur (author)

Core Title Prediction And Forecasting In International Relations: An Exploratory Analysis

Degree Doctor of Philosophy

Degree Program International Relations

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,political science, international law and relations

Language English

Contributor Digitized by ProQuest (provenance)

Advisor McClelland, Charles A. (committee chair), Berkov, Robert H. (committee member), Morgner, Aurelius (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-434105

Unique identifier UC11362235

Identifier 7025078.pdf (filename),usctheses-c18-434105 (legacy record id)

Legacy Identifier 7025078.pdf

Dmrecord 434105

Document Type Dissertation

Rights Young, Robert Arthur

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA