¹ Eckstein, Harry, “On the Etiology of Internal Wars,” History and Theory, 4 (1965), 133–163 CrossRefGoogle Scholar.

² See, for example, Rummel, R. J., “Dimensions of Conflict Behavior Within Nations, 1946–59,” Journal of Conflict Resolution, 10 (03, 1966), 41–64 CrossRefGoogle Scholar; Rummel, R. J., “A Field Theory of Social Action with Application to Conflict Within Nations,” General Systems Yearbook, 10 (1965), 183–211 Google Scholar; Rummel, R. J., “Dimensions of Conflict Behavior Within and Between Nations,” General Systems Yearbook, 8 (1963), 1–50 Google Scholar; Tanter, Raymond, “Dimensions of Conflict Behavior Within and Between Nations, 1958–60,” Journal of Conflict Resolution, 10 (03, 1966), 41–64 CrossRefGoogle Scholar; Tanter, Raymond, “Dimensions of Conflict Behavior Within Nations, 1955–60: Turmoil and Internal War,” Peace Research Society Papers, 3 (1965), 159–183 Google Scholar. For a summary of the dimensionality studies see Hazelwood, Leo, “Concept and Measurement Stability in the Study of Conflict Behavior Within Nations,” Comparative Political Studies, 6 (07, 1973), 171–195 CrossRefGoogle Scholar.

³ See, for example, Feierabend, Ivo K. and Feierabend, Rosalind L., “Aggressive Behaviors Within Polities, 1948–62: A Cross-National Study,” Journal of Conflict Resolution, 10 (09, 1966), 249–271 CrossRefGoogle Scholar; Russett, Bruce, “Inequality and Instability: The Relation of Land Tenure to Politics,” World Politics, 16 (04, 1964), 442–454 CrossRefGoogle Scholar.

⁴ See, for example, Davies, James, “Toward a Theory of Revolution,” American Sociological Review, 27 (02, 1962), 5–18 CrossRefGoogle Scholar; Galtung, Johan, “A Structural Theory of Aggression,” Journal of Peace Research, 1 (1964), 95–119 CrossRefGoogle Scholar; Gurr, Ted Robert, “Psychological Factors in Civil Violence,” World Politics, 20 (01, 1968), 245–278 CrossRefGoogle Scholar; Johnson, Chalmers, Revolution and the Social System, Hoover Institution Studies #3 (Stanford: The Hoover Institution on War, Revolution, and Peace, Stanford University, 1964)Google Scholar; Olson, Mancur, “Rapid Growth as a Destabilizing Force,” Journal of Economic History, 23 (12, 1963), 529–552 CrossRefGoogle Scholar.

⁵ See, for example, Bwy, Douglas, “Political Instability in Latin America: The Cross-Cultural Test of a Causal Model,” Latin American Research Review, 3 (Spring, 1968), 17–66 Google Scholar; Gurr, T. R., A Causal Model of Civil Strife: A Comparative Analysis Using New Indices,” American Political Science Review, 62 (12, 1968), 1104–1124 CrossRefGoogle Scholar; Midlarsky, Manus and Tanter, Raymond, “Toward a Theory of Political Instability in Latin America,” Journal of Peace Research, 4 (1967), 209–227 CrossRefGoogle Scholar; Morrison, Donald and Stevenson, H. M., “Integration and Instability: Patterns of African Political Development,” American Political Science Review, 66 (09, 1972), 902–927 CrossRefGoogle Scholar.

⁶ Hibbs, Douglas, Mass Political Violence (New York: John Wiley, 1973)Google Scholar.

⁷ Gurr, T. R. and Duvall, Raymond, “Civil Conflict in the 1960's: A Reciprocal Theoretical System with Parameter Estimates,” Comparative Political Studies, 6 (07, 1973), 135–169 CrossRefGoogle Scholar.

⁸ See, in particular, Gurr, Ted Robert, Why Men Rebel (Princeton: Princeton University Press, 1970)Google Scholar. The Gurr and Duvall model has been chosen for this analysis since it is more fully specified and operationalized than others. Although Hibbs's work may be better known, it does not represent a single, fully developed body of theory.

⁹ Gurr and Duvall, p. 135.

¹⁰ Ibid., pp. 141–142.

¹¹ The twenty-five countries are: Burundi, Cameroun, CAR, Chad, Dahomey, Ghana, Guinea, Ivory Coast, Kenya, Liberia, Malawi, Mali, Niger, Nigeria, Rwanda, Senegal, Sierra Leone, Somali, Sudan, Tanzania, Togo, Uganda, Upper Volta, Zaire, Zambia.

¹² Taylor, Charles L. and Hudson, Michael, World Handbook of Political and Social Indicators, 2nd ed. (New Haven: Yale University Press, 1972)Google Scholar.

¹³ Morrison, Donald G., Mitchell, Robert, Paden, John N., and Stevenson, Hugh M., Black Africa: A Comparative Handbook (New York: Free Press, 1972)Google Scholar.

¹⁴ Concept scores are created by calculating the mean standard scores of all indicators of a concept. When a concept consists of several dimensions (e.g., economic, political, and social dimensions of stress), a mean standard score for each dimension is first calculated, and then the mean of the dimensions obtained. Gurr and Duvall use a similar simple additive model for creating RIS, DIS, and TRAD scores. In scoring strain and stress, however, they regress the separate indicators on their measures of conflict and then weight each indicator according to the structure of the regression equation. That procedure was not employed here since it results in several scores for the same concept and also since it seems a bit arbitrary.

¹⁵ Since Gurr and Duvall include political discrimination in their measure of strain it is included in the following analyses. Evidence suggests, however, that political discrimination has different effects than do the other components of strain. Regressions of the strain components on conflict for both data sets reveal that other forms of discrimination promote conflict, while political discrimination tends to reduce conflict. Consider, for example, turmoil:

Gurr and Duvall also report in a footnote that political discrimination seems to reduce conflict, unlike the other components of strain, but they proceed to include it in their concept score. Thus, they give positive weights to the other strain indicators and a negative weight to political discrimination. This procedure seems to deviate from their conceptualization of strain, which is viewed as a positive contributor of conflict. Since the negative impact of political discrimination is most marked in the Black Africa data, two measures of strain are created and employed in later analyses, one including and one excluding political discrimination.

¹⁶ Turmoil data have been collected by the authors of Black Africa but are not published in that book because of questionable reliability. Turmoil data begin with the year of independence. Pre-independence years for countries independent after 1960 have been updated with World Handbook data. Taylor and Hudson code days of rioting, demonstrating, and so forth, rather than the number of events of each type. For comparability, five days were considered roughly equivalent to one event.

¹⁷ Concept scores and rank orders on each concept for the twenty-five nations are available from the author.

¹⁸ See, for example, Campbell, Donald T., “Techniques for Determining Cultural Biases in Comparative Research,” in Comparative Perspectives: Theories and Methods, ed. Etzioni, Amitai and Dubow, Frederick L. (Boston: Little, Brown, 1970), pp. 407–410 Google Scholar; Janda, Kenneth, “Data Quality Control and Library Research on Political Parties,” in A Handbook of Method in Cultural Anthropology, ed. Naroll, Raoul and Cohen, Ronald (Garden City: National History Press, 1970), pp. 962–973 Google Scholar; Naroll, Raoul, “Data Quality Control in Cross-Cultural Surveys,” in Naroll, and Cohen, , pp. 927–945 Google Scholar; Rummel, R. J., “Dimensions of Error in Cross-National Data,” in Naroll, and Cohen, , pp. 946–961 Google Scholar; Naroll, Raoul, Data Quality Control—A New Research Technique (New York: Free Press, 1962)Google Scholar; Russett, Bruce, “Techniques for Controlling Error,” in Statistical and Quantitative Methods, ed. Haas, Michael (Evanston: Northwestern University Press, 1972)Google Scholar.

¹⁹ Rummel, “Dimensions.”

²⁰ By random disturbance I mean that: E(e_i) — 0, all i; E(e_ie_j) = 0, all i, j , i ≠ j; , all i.

²¹ The reason for this logic is that the form of error for the systematically erroneous variables is known to be different for the two data sets and nonlinear in at least one of them, whereas the seemingly reliable variables can contain nonlinear systematic error only if the nonlinear form is similar across data sets. Therefore, shared systematic error is possible for a maximum of one data set and is compatible only with differences in correlations across data sets.

²² Gurr and Duvall (“Civil Conflict in the 1960's”) provide a brief discussion of the techniques used, and Hibbs (Mass Political Violence) provides a helpful appendix on two-stage-least-squares regression. Readers unfamiliar with these techniques might find both discussions useful.

²³ These assumptions were derived from discussions in Johnston, John, Econometric Methods (New York: McGraw Hill, 1963), p. 232 Google Scholar, and Goldberger, Arthur S., Econometric Theory (New York: John Wiley, 1964), pp. 299–300 Google Scholar.

²⁴ Recall that Gurr and Duvall create a couple of stress scores, weighting indicators according to the structure of a regression equation predicting each form of conflict. See footnote 14.

²⁵ The inclusion of traditions of conflict which is scored differently for the African countries and includes as components past turmoil and past rebellion (see section on Measurement of Concepts), could explain the exaggerated effects of traditions and the weak-ened effects of past rebellion in the second set of equations.

²⁶ Przeworski, Adam and Teune, Henry, The Logic of Comparative Social Inquiry (New York: John Wiley, 1970)Google Scholar.

²⁷ See in particular Welfling, Mary, Political Institutionalization: Comparative Analyses of African Party Systems, Sage Professional Papers in Comparative Politics #01–041 (Beverly Hills: Sage Publications)Google Scholar.

²⁸ The author has scored thirteen indicators of party system institutionalization in the black African sample for the period 1945–1970. Previous analyses demonstrated that these indicators fall into four conceptual dimensions. The highest-loading variable on each dimension was selected as representative of that dimension, and the mean standard score of the variables calculated for two time periods, 1955–60 and 1961–65. The four representative variables are percentage of seats held by independents, new party entities, electoral participation, and electoral discrimination (for postindependence years). A fuller explanation of this scoring procedure can be found in Duvall, Raymond and Welfling, Mary, “Social Mobilization, Political Institutionalization and Conflict in Black Africa: A Simple Dynamic Model,” Journal of Conflict Resolution (12, 1973)CrossRefGoogle Scholar; and a full explanation of the scoring of all institutionalization indicators can be found in Welfling, Political Institutionalization.

²⁹ For other suggestions on how different forms of conflict interrelate see Hibbs, Mass Political Violence, and Duvall and Welfling.