Hostname: page-component-cd9895bd7-8ctnn Total loading time: 0 Render date: 2024-12-25T17:01:58.541Z Has data issue: false hasContentIssue false
  • English
  • Français

Political Instability in Latin America: The Cross-Cultural Test of a Causal Model

Published online by Cambridge University Press:  24 October 2022

D. P. Bwy
D. P. Bwy*
Affiliation:
Case Western Reserve University∗
Rights & Permissions [Opens in a new window]

Extract

Core share and HTML view are not available for this content. However, as you have access to this content, a full PDF is available via the ‘Save PDF’ action button.

although at times quite thin, there does appear to be a common thread of agreement running through most of the classic and contemporary literature on theories of revolution—this being the simple proposition that the majority of the participants engaging in such activity are dissatisfied, discontented, and often disaffected individuals. If we can think of “revolution” for the moment in its most general terms—to subsume under such a conceptual label both the simplest manifestation of civil disorder to the most grandiose occurrence of what might be called basic social change—then, it seems, we are in a position to illustrate the emergence of this basic proposition throughout the literature.

Type
Topical Review
Information
Latin American Research Review , Volume 3 , Issue 2 , Spring 1968 , pp. 17 - 66
Copyright
Copyright © 1968 by the University of Texas Press

Footnotes

∗

Acknowledgments are due to the National Science Foundation, which partially supported (GS-789) the research for the paper. Special thanks are also due to Professors R. J. Rummel and Raymond Tanter for the use of their (1955–64) conflict data; to Professor Russell Fitzgibbon for the use of his “democratic attainment” ratings across the 20 Latin American republics for 5 time periods since 1945; to Professor Phillips Cutright for making available his individual data (1940–61) from which he composed the “Political Representativeness Index;” and to Professor George Blanksten for guidance and encouragement.

References

NOTES

1. James Geschwender, “Social Structure and the Negro Revolt: An Examination of Some Hypotheses,” Social Forces, 43 (December 1964), p. 249.

This hypothesis has received some support from Sorokin in his analysis of revolution. See: Pitirim A. Sorokin, The Sociology of Revolution (Philadelphia: J. B. Lippincott Company, 1925), p. 367.

2. Lyford P. Edwards, The Natural History of Revolution (Chicago: University of Chicago Press, 1927). Italics added.

3. Eric R. Wolf, Sons of the Shaking Earth (University of Chicago Press, 1959), pp. 108–109.

4. James C. Davies, “Toward a Theory of Revolution,” American Sociological Review, 27 (February 1962), p. 6.

5. Cole Blasier, “Studies of Social Revolution: Origins in Mexico, Bolivia, and Cuba,” Latin American Research Review, 2 (Summer 1967), p. 49.

6. Merle Kling, “Toward a Theory of Power and Political Instability in Latin America,” Western Political Quarterly, 9 (March 1956), p. 33.

7. Blasier, op. cit., p. 39. Frank Tannenbaum, Mexico, The Struggle for Peace and Bread (New York: Alfred A. Knopf, Inc., 1954), p. 136.

8. Ibid. Eyler N. Simpson, The Ejido: Mexico's Way Out (Chapel Hill: The University of North Carolina Press, 1937), p. 43.

9. Ibid. Robert J. Alexander, The Bolivian National Revolution (New Brunswick, N.J.: Rutgers University Press, 1958), p. 58.

10. Bruce M. Russett, “Inequality and Instability: The Relation of Land Tenure to Politics,” World Politics, 16 (April 1964), pp. 442–454.

11. The inferential statistics reported by Russett was even more impressive—the significance level, or probability of error (in the findings) being .001. In other words, unless there really was a positive relationship between inequitable land distribution and instability, this high a correlation would not occur, purely by chance, as often as one time in a thousand.

By using multiple regression analysis, Russett was able to simultaneously “control” for a number of contributing conditions. He argued that a more complex hypothesis, but one closely related to Kling's, would be that “extreme inequality of land distribution leads to political instability only in those poor, predominantly agricultural societies where limitation to a small plot of land almost unavoidably condemns one to poverty.” The correlation between inequality and instability, controlling for (i) GNP per capita (i.e., in poor nations), (ii) percentage of labor in agriculture (i.e., in poorly industrialized, primarily agrarian societies), and (iii) tenancy (i.e., in societies where farms tended to be rented as opposed to owned), jumped to +.71, a considerable improvement in predictability over the earlier coefficient. The implications of these findings for the present analysis seem to be that inequality (itself undoubtedly giving rise to discontent) obviously is associated with instability. The greater the inequality—that is when inequitable land distribution takes place in poor agrarian-based nations—the greater the discontent, and ultimately the greater the instability.

12. Raymond Tanter and Manus Midlarsky, “A Theory of Revolution,” Journal of Conflict Resolution, 11 (September 1967), p. 267.

13. Seymour Martin Lipset, “Some Social Requisites of Democracy: Economic Development and Political Legitimacy,” American Political Science Review, 53 (March 1959), p. 74.

14. Daniel Lerner, The Passing of Traditional Society: Modernizing the Middle East (New York: The Free Press of Glencoe, 1958), p. 50.

15. Phillips Cutright, “National Political Development: Measurement and Analysis,” American Sociological Review, 28 (April 1963), pp. 253–264. Cutright's variable of “communication” corresponds conceptually and in its operationalization to Lerner's “media participation.”

16. Lerner, op. cit., p. 60.

17. Lipset, op. cit., pp. 81, 82, 83.

18. Hayward Alker and Bruce Russett, “Correlations Among Political and Social Indices,” in: Russett, et al. World Handbook of Political and Social Indicators (New Haven: Yale University Press, 1964), p. 272.

Alker and Russett ran another test, the results of which also bear considerably on the present argument. In an attempt to “explain” deaths from domestic group violence (or, in our terms, political instability), they inputted a number of independent variables into a multiple regression analysis. They used GNP per capita, for example, to correspond with the common hypothesis that, all other things being equal, economic development promotes political stability. They also included the annual rate of change in GNP per capita in the analysis to see which, if either, of two contradictory causal hypotheses might be correct: One being that a rapid rate of growth may serve to channel energies, relieve dissatisfaction and promote stability. The opposite idea is that a rapid growth rate implies a rapid rate of social change, dislocation, and potential instability. Alker and Russett found that both an increase in the level of GNP per capita and a rise in its growth rate were associated with domestic tranquility; that is, they were negatively associated with violence. In propositional terms: political instability increases as economic development (in static as well as dynamic terms) decreases.

19. Tanter and Midlarsky, op. cit., p. 272.

20. Ted Gurr, with Charles Ruttenberg, The Conditions of Civil Violence: First Tests of a Causal Model. Research Monograph No. 28, Center of International Studies, Princeton University, April 1967, p. 62.

21. Ibid., p. 28.

22. Ivo K. and Rosalind L. Feierabend, “Aggressive Behaviors Within Polities, 1948–62: A Cross-National Study,” Journal of Conflict Resolution, 10 (September 1966), pp. 256–257.

See also: Betty A. Nesvold, Modernity, Social Frustration, and the Stability of Political Systems: A Cross-National Study. Master's thesis, San Diego State College, June 1964.

23. Alker and Russett, op. cit., p. 306.

24. Bruce M. Russett, Trends in World Politics (New York: The Macmillan Company, 1965), p. 136.

25. Max F. Millikan and Donald L. M. Blackmer (eds.), The Emerging Nations (Boston: Little, Brown and Company, 1961), p. 40.

26. Ibid., p. 41.

27. Russett, Trends in World Politics, op. cit., p. 137.

28. Gabriel A. Almond and Sidney Verba, The Civic Culture: Political Attitudes and Democracy in Five Nations (Princeton: Princeton University Press, 1963), p. 18.

29. Arthur S. Banks and Phillip M. Gregg, “Grouping Political Systems: Q-Factor Analysis of A Cross-Polity Survey,” American Behavioral Scientist, 9 (November 1965), p. 4.

30. Gurr, op. cit., p. 46.

31. Russett, Trends in World Politics, op. cit., p. 137.

32. Russett, “Inequality and Instability,” op. cit., pp. 450–451.

33. Lawrence Stone, “Theories of Revolution,” World Politics, 18 (January 1966), p. 173.

34. Davies, op. cit., p. 6.

35. One survey research instrument which appears to be assessing psycho-social dissatisfactions, and which has been applied to four Latin American populations, is the “self-anchoring striving scale” developed by Hadley Cantril and F. P. Kilpatrick. The primary significance of this technique for cross-cultural work, is that it allows the respondent to be measured and evaluated in terms of his own perspective—or as the designers of the instrument say, in terms of his “own unique reality world.” If discontent, as we have suggested, is a key variable in “explaining” political instability, then one might measure the amount of discontent as the discrepancy between what one wants (aspirations) and what one gets (gratifications). The scale is “self-anchoring” in the sense that the respondent emotes (in an open-ended fashion) on his own version of the best (which represents the 10th rung of a hypothetical ladder) and worst (which represents the bottom rung) possible life. The respondent, then, anchors himself when he is asked to identify where he falls on such a continuum: (i) at the present time, (ii) approximately 5 years in the past, and (iii) 5 years into the future. By avoiding the problem of structure, and talking in terms of each respondent's own perceptions of reality, one has arrived at a remarkably comparable cross-cultural tool for measuring similar phenomena (namely, frustrations or dissatisfactions) across often divergent research populations.

For a detailed description of the instrument itself, see: F. P. Kilpatrick and Hadley Cantril, “Self-Anchoring Scaling: A Measure of Individuals' Unique Reality Worlds,' Journal of Individual Psychology, 16 (November 1960), pp. 158–173.

For a report on the results of the use of the instrument is over 14 national samples, see Hadley Cantril, The Pattern of Human Concerns (New Brunswick, N.J.: Rutgers University Press, 1965), 427 pp.

36. L. P. Kendall and K. M. Wolf, “The Two Purposes of Deviant Case Analysis,” in: Paul Lazarsfeld and M. Rosenberg, The Language of Social Research (Glencoe: The Free Press, 1955), p. 169.

37. Gross National Product per Capita, U.S. $, 1957, World Handbook of Political and Social Indicators, pp. 155–157.

38. Percentage of Expenditures of the Central Government on Public Health and Welfare, 1958–59, 1960 Statistical Abstract of Latin America (Latin American Center: University of California, Los Angeles, published: 1956–58, 1961–67), p. 32. 1961 Statistical Abstract, p. 35.

39. GINI Index of Inequality, World Handbook of Political and Social Indicators, pp. 239–240.

40. Lipset, op. cit., pp. 90–91.

41. Ibid., pp. 86–87.

42. Robert Packenham, “Approaches to the Study of Political Development,” World Politics, 18 (October 1964), p. 117.

43. David Easton, “An Approach to the Analysis of Political Systems,” World Politics, 9 (April 1957), pp. 383–400.

44. These outputs, to Easton, constitute specific inducements which may be either positive (a reward) or negative (a punishment).

45. Lewis A. Coser, “Social Conflict and the Theory of Social Change,” British Journal of Sociology, 8 (September 1957), p. 203.

“Mere frustration,” notes Coser, “will not lead to a questioning of the legitimacy of the position of the vested interests, and hence to conflict. Levels of aspiration as well as feelings of deprivation are relative to institutionalized expectation and are established through comparison.”

46. S. M. Lipset, “Democracy and the Social System,” in: Harry Eckstein (ed.), Internal War: Problems and Approaches (New York: The Free Press of Glencoe, 1964), p. 302.

47. Gabriel Almond and Sidney Verba, The Civic Culture, op. cit., p. 246.

48. Ibid., p. 253. There are obviously some qualifications to this conclusion; for one, the relationship doesn't appear to hold for Germany and Italy. Almond and Verba conclude, however, that “in the United States, Britain, and Mexico those who consider themselves competent to participate in governmental decisions are more likely than those who do not feel this way to express pride [positive affect] in the political aspects of their nation.” Ibid., p. 247.

49. Cutright, “National Political Development …,” op. cit. The PRI index was also in: “Urbanization and Change in National Political Structures: 1928–1961,” paper prepared for the Carnegie IDRC Joint Study Group on Measurement Problems, Indiana University, October 1964.

50. For the most recent publication of the results of these surveys, see: Russell H. Fitzgibbon, “Measuring Democratic Change in Latin America,” The Journal of Politics, 29 (February 1967), pp. 129–166.

The 15 criteria were: (1) An educational level sufficient to give the political process some substance and vitality. (2) A fairly adequate standard of living. (3) A sense of internal unity and national cohesion. (4) Belief by the people in their individual political dignity and maturity. (5) Absence of foreign domination. (6) Freedom of the press, speech, assembly, radio, etc. (7) Free and competitive elections—honestly counted votes. (8) Freedom of party organization; genuine and effective party opposition in the legislature; legislative scrutiny of the executive branch. (9) An independent judiciary—respect for its decisions. (10) Public awareness of accountability for the collection and expenditure of public funds. (11) Intelligent attitude toward social legislation—the vitality of such legislation as applied. (1) Civilian supremacy over the military. (13) Reasonable freedom of political life from the impact of ecclesiastical controls. (14) Attitude toward and development of technical, scientific and honest governmental administration. (15) Intelligent and sympathetic administration of whatever local self-government prevails.

As Charles Wolf of RAND has observed, “there are many shortcomings in the Fitzgibbon method and data, including the ambiguity and heterogeneity of the criteria, the weights applied to the criteria …, and the qualifications and prejudices of the respondents (nearly all of the respondents were from the United States).” See: Wolf, “The Political Effects of Military Programs: Some Indications from Latin America,” Orbis, 8 (Winter 1965), pp. 878–879.

Wolf suggests, however, that some of these difficulties could be overcome by independent work with the original data. Among other things, he mentions: (i) the use of only responses to the more distinctly relevant and unambiguous criteria, (ii) the separation of the responses of the more qualified respondents, and (iii) reliability tests, comparing the subjective estimates of the respondents with objective data relating to education, press circulation, the frequency and character of elections, and so on. Fitzgibbon has attempted some of these tests, and the results are reported in the February 1967 issue of the Journal of Politics, cited above.

51. Arnold H. Buss, The Psychology of Aggression (New York: John Wiley & Sons, 1961), p. 58.

52. Jenifer Walton, Correlates of Coerciveness and Permissiveness of National Political Systems: A Cross-National Study, M.A. thesis, Department of Political Science, San Diego State College, June 1965.

53. Robert A. LeVine, “Anti-European Violence in Africa: A Comparative Analysis,” Journal of Conflict Resolution, 3 (1959), pp. 420–429.

54. The budget for ordinary police forces is usually included in the total military outlay.

55. R. A. Gomez, “Revolution, Violence, Political Morality,” in: Gomez, Government and Politics in Latin America (New York: Random House, 1963), p. 59.

56. John J. Johnson, The Military and Society in Latin America (Stanford: Stanford University Press, 1964), pp. 101–102.

57. James Payne, “Peru: The Politics of Structured Violence,” Journal of Politics, 27 (May 1965), p. 369.

58. Edwin Lieuwen, “The Changing Role of the Military in Latin America,” Journals of Inter-American Studies, 3 (October 1961), p. 564.

59. Wyckoff has presented a systematic scheme for analyzing the quality of loyalty. He notes that in states in which the military always intervenes, that the greater the proportion of junior officers to middle and senior officers, the greater the loyalty of the armed forces. In states where the military occasionally intervenes, loyalty is best gauged by the greater proportion of junior plus middle rank officers, as opposed to senior officers. See: Theodore Wyckoff, “The Role of the Military in Latin American Politics,” Western Political Quarterly, 13 (September 1960), pp. 749–760.

60. John Powell searched the following source for what he considered to be dependable comparative data: Sandberg, Bengt, and others, Comparative Data on Latin American Countries, International Bank for Reconstruction and Development (Washington, D.C., January, 1962). See: John D. Powell, “Military Assistance and Militarism in Latin America,” Western Political Quarterly, 18 (June 1965), pp. 382–392.

61. For a discussion of these measures, see: Bruce M. Russett, “Measures of Miiltary Effort,” American Behaviorial Scientist, 7 (February 1964), pp. 26–29.

62. World Handbook of Political and Social Indicators, op. cit., pp. 79–80.

63. Powell, op. cit., p. 384.

64. Kling, op. cit., p. 21.

65. William Stokes, “Violence as a Power Factor in Latin American Politics,” Western Political Quarterly (September 1952), p. 445.

66. “If the normal way of life of rotating the executive in a given country is by revolution,” continues Silvert, “and if there have been a hundred such changes in a century, then it is not being facetious to remark that revolutions are a sign of stability—that events are marching along as they always have.” See: Kalman Silvert, The Conflict Society: Reaction and Revolution in Latin America (New Orleans: The Hauser Press, 1961), p. 20.

67. Payne, op. cit., p. 363.

68. George I. Blanksten, “The Politics of Latin America,” in: Gabriel Almond and James Coleman (eds.), The Politics of the Developing Areas (Princeton: Princeton University Press, 1960), p. 497.

69. Cutright, 1963, op. cit., pp. 260–261. The basis for Cutright's jump from notions of political development to those of instability are not precisely clear. This is especially true, when one realizes that his measure of political development consists of a rating across all countries in terms of the quality of (i) minority party representation within the legislative body, and (ii) nations ruled by chief executives elected in open and competitive election. While such things may be the negative correlate of “political instability”—as we have suggested in the section “Psycho-Social Dissatisfaction,” which re-interprets both the Cutright and Lipset work—such interpretations can only be made after the linkages between “competition for equal power sharing” (“democracy”) and political stability or instability are “established,” and then only with extreme caution.

70. George I. Blanksten, “Revolutions,” in: Harold E. Davis (ed.), Government and Politics in Latin America (New York: Ronald Press Company, 1958), p. 119.

71. Stokes, op. cit., p. 445.

72. Kling, op. cit., p. 22.

73. Stokes, op. cit., p. 464.

74. Alberto de Mestas, El Salvador: Pais de Lagos y Volcanos (Madrid: Editorial Cultura Hispánica, 1950).

75. L. L. Thurston, Multiple Factor Analysis: A Development and Expansion of the Vectors of the Mind (Chicago: University of Chicago Press, 1960), pp. 55–56. Thurston further notes that “the exploratory nature of factor analysis is often not understood. Factor analysis has its principal usefulness at the border line of science. It is naturally superseded by rational formulations in terms of the science involved. Factor analysis is useful, especially in those domains where basic and fruitful concepts are essentially lacking and where crucial experiments have been difficult to conceive.”

76. The underlying assumptions involved in the use of factor analysis are (i) the variables used in the analysis are linear, (ii) the data for each observation are assumed to be of equal importance and are thus given equal weight (in terms of the present study, therefore, the number of demonstrations in the Dominican Republic are considered equally important as those occurring in Argentina), and (iii) the assumption that the data are distributed normally. Since the Pearson Product Moment Correlation Coefficient is used as input into the factor analysis, the data will have to meet the assumptions underlying this statistic (two of which appear above) as well.

77. This property of independence becomes quite useful when delimiting domain of numerous variables into a manageable one with few, uncorrelated, or independent, variables. Independence, or orthogonality, between factors will be of prime importance in the analysis section of this paper, when we try to establish the variants of inter-group aggression against government into separate dimensions.

78. That is, the sum of squares (the sum of the squared deviations from the mean) of factor loadings is maximized on each factor.

79. The method also has the advantage of providing a mathematically unique, or least squares, solution for the matrix of correlation coefficients.

80. The nine basic variables used in this part of the analysis were those selected and used in a study of conflict behavior within and between nations, conducted by R. J. Rummel in 1963, and published in the General Systems Yearbook, 8 (1963), pp. 1–50. See also: Rummel, Dimensions of Conflict Behavior Within and Between Nations, paper prepared in connection with research supported by the National Science Foundation, under Contract G24827. June 1963, 108 pp.

The Rummel data (gathered for the years 1955–57, for 78 nations) were richly supplemented by Raymond Tanter, when he replicated the Rummel study after gathering similar data for the following three-year period (1958–60) across 83 nations. It is the Tanter data which were factor analyzed, and are presented in Tables I and II.

See: Raymond Tanter, Dimensions of Conflict Behavior Within and Between Nations, 1938–60, monograph prepared in connection with research supported by the National Science Foundation, Contract GS224. See also: Journal of Conflict Resolution, 10 (March 1966), pp. 41–64.

81. The computer program which monitored these calculations was MESA 1, a 95 X 95 Factor Analytic Program with Varimax Rotation. Special thanks are due Northwestern University's Vogelback Computing Center for their generous allowance of computing time.

The lower limit for eigenvalues (i.e., a proportion of variance which may vary from near zero to n, where n is the number of variables entering a factor matrix) to be included in rotation was 1.00. Rotation is carried out in order to obtain a solution which is not entirely dependent upon each particular variable in the analysis. Orthogonal rotation is the fitting of factors to clusters of variables with the restriction that the correlation between factors is zero. The varimax criterion is used to rotate orthogonally to “simple structure,” that is, the maximization of high and low loadings. Thus, this form of rotation continues to maintain independence among the factors.

82. One possible way to test this assumption, would be to compute the factor scores for each of the 20 republics on this third dimension. If the assumption is accurate, nations experiencing low levels of conflict (e.g., Uruguay and Costa Rica) should come out high on such a distribution.

83. Almond and Coleman, op. cit., p. 34.

84. One less dimension, than emerged for similar data gathered over the 1955–57 time period by Rudolph Rummel. For a detailed discussion of these dimensions, and a comparison of them to both the earlier Rummel study, as well as to the basic dimensions extracted from an analysis of conflict data (1946–59) gathered by Eckstein, see: R. J. Rummel, “Dimensions of Conflict Behavior Within Nations, 1946–59,” Journal of Conflict Resolution, 10 (March 1966), pp. 65–73.

85. Harry Eckstein, “Internal War: The Problem of Anticipation,” in: Ithiel de Sola Pool, et al, Social Science Research and National Security: A Report Submitted to the Research Group in Psychology and the Social Sciences (Washington, D.C.: Smithsonian Institution, 1962), p. 104.

86. With the possible exception of “peasant revolt,” which is generally referred to as machetismo (a sporadic outbreak, generally in rural areas, by peasants or peons, primarily involved in agriculture economies), and which is probably measuring violence more of an anomic nature.

87. Of street riots, Silvert says “… they usually take place to protest governmental actions, such as a rise in bus fares or the arrest of political or labor leaders. University students are very prone to this kind of violence.” See: Silvert, op. cit., pp. 20–23.

88. A common term for indices which take all loadings into account is “factor scores,” which represent the country's overall score (in standard score form) weighted by the factor loadings for each of the indices in the analysis. Such scores are automatically computed for both unrotated and rotated factor solutions by Mesal, but were not used here because a more discriminating index was desired. For a description of some formulas for calculating these “factor” or “component” scores, see the following sources: Benjamin Fruchter and Earl Jennings, “Factor Analysis,” in: Harold Borko (ed.), Computer Applications in the Behavioral Sciences (Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1962), pp. 260–262. Henry Kaiser, “Formulas for Component Scores,” Psychometrika, 27 (1962), pp. 83–88.

89. Purges, for example, might be one of the component parts of a factor which allows a population to perceive a government as repressive, and therefore, should appear as an intervening variable (perhaps helping “cause” potential aggressors to temper feelings of hostility with notions of possible reprisal for “deviant” behavior).

90. S. S. Wilks, “Weighting Systems for Linear Functions of Correlated Variables When There is No Independent Variable,” Psychometrika, 3 (March 1938), pp. 24–43. Harold Hotelling, “Analysis of a Complex of Statistical Variables into Principal Components,” Journal of Educational Psychology, 24, pp. 417–441, 498–520.

91. The computation formulas for both the Organized and Anomic Violence Indices are:

92. Rummel, Dimensions of Conflict Behavior…, 1963, op. cit., Appendix II: Raw and Transformed Data, pp. 77–80.

93. Tanter, Dimensions of Conflict Behavior …, 1964, op. cit., Appendix II: Raw Data and List of Nations, pp. 71–74.

94. Rummel, “A Field Theory of Social Action …,” op. cit., Appendix II: Data Tables and Definitions,“ pp. iii–viii.

95. A. Terry Rambo, “The Dominican Republic,” in: Martin C. Needier (ed.), Political Systems of Latin America (New York: D. Van Nostrand Company, Inc., 1964), p. 175.

96. Robert J. Alexander, “Bolivia: The National Revolution,” in Needier, op. cit., p. 329.

97. See: Donald T. Campbell and Julian C. Stanley, “Experimental and Quasi-Experimental Designs for Research on Teaching,” in: N. L. Gage, Handbook of Research on Teaching (New York: Rand McNally & Company, 1963), p. 239.

98. Donald T. Campbell, “From description to Experimentation: Interpreting Trends as Quasi-Experiments,” in: C. W. Harris (ed.), Problems in Measuring Change (Madison: University of Wisconsin Press, 1963), pp. 212–242.

The panel correlation, originally designed for use with survey research data, where the same respondent was interviewed at more than one point in time, seems directly applicable here, where data for the same nation is analyzed, also at more than one point in time.

99. 1950 + 1955 Fitzgibbon ratings.

100. 1960 + 1965 Fitzgibbon ratings.

Note: Although coefficients were available for all of the data points within the matrix (as, for example, those appearing in all the cells of the first variable, “Change in PRI”), correlations only appear which test the causal model:

101. Satisfaction: “Annual Growth of G.N.P. per Capita,” data for 13 of the Latin American republics from the World Handbook of Political and Social Indicators, pp. 160–161; data for the remaining (including the corrected calculation for Venezuela) were computed from GNP per capita statistics from the following sources: 1952 GNP per capita: Harold Davis, Government and Politics in Latin America, p. 64 (most of these figures came from the 1955 Statistical Abstract of Latin America). 1957(58) GNP per capita: figures here came from two sources: The World Handbook of Political and Social Indicators, pp. 155–157, and the 1960 Statistical Abstract of Latin America, p. 30; the difference between the 1952 and 1957(58) figures was taken, and divided by the number of years involved, to obtain the annual growth figure.

Pearson Product Moment Correlation = -.63.

102. Payne, op. cit., p. 369. Italics added.

103. The high value of 3–5% of GNP allocated to defense expenditures during 1959–60, does not appear to be a data-quality error. When checked against similar data gathered by Bruce Russett for the 1959 period, the Peruvian statistics is placed at 30%.

104. Force: data directly from John D. Powell, “Military Assistance and Militarism in Latin America,” Western Political Quarterly, 18 (June 1965), p. 384. Supplemental data for Uruguay and Cuba taken from: Bruce M. Russett, “Measures of Military Effort,” American Behavioral Scientist, 7 (February 1964), pp. 26, 28.

Data presented in raw form; when Y variable transformed to log10, the curve is smoother and more accentuated; kept in raw score form, here, for comparative purposes.

105. It should be recalled that it was ambivalence, in terms of a permissive and then a coercive colonial policy, which LeVine postulated gave rise to anti-European conflict in Africa.

106. The resulting coefficient, while in the same direction as hypothesized, can only be considered a suggestive test of the relationship. For various reasons, it was felt that the strength of the coefficient would be considerably boosted with more sensitive data. For example, if the change-in-defense data were available between 1960 to 1963, the period immediately following the 1958–60 Organized violence data.

Data for the calculation of the change measure was gathered from the following sources: “1958–59 per cent of central government expenditure on defense,” 1960 Statistical Abstract of Latin America, p. 32; 1961 Statistical Abstract of Latin America, p. 35. “1962–63 per cent of central government expenditure on defense,” 1962 Statistical Abstract of Latin America, p. 66; 1963 Statistical Abstract, p. 76; 1964 Statistical Abstract, p. 104.

107. Legitimacy: Change scores calculated from data published by Russell Fitzgibbon and Kenneth Johnson, “Measurement of Latin American Political Change,” American Political Science Review, 55 (September 1961), p. 518.

Data presented in raw form; when Y variable is transformed to log10, the negative correlation coefficient of -.71 (for the relationship in Figure III) is strengthened.

108. The log10 transformation was used, which increased the -.50 relationship found among the raw data to -.67.

109. It will be recalled that the Organized Violence Index, constructed on the basis of the empirical evidence from the factor analysis of conflict activity, weights guerrilla activities .65, governmental crisis .57, and revolutions .31. It is not surprising, therefore, that Cuba and Venezuela, which have experienced considerable guerrilla-type activity during the period 1958–60, should come out highest on Organized violence in the scatter plot in Figures II and IV.

110. Lipset, “Some Social Requisites of Democracy …,” op. cit., p. 92.

This article was sent for critical commentary to 10 scholars in the fields of political science, sociology, and anthropology. Replies received in time to go to press are included on the following pages in the order in which they arrived. The author's rebuttal will be printed in the forthcoming issue of the Review.