A major feature of the correlations in Table 6.18 deals with the signs attached to the correlations. Recall that the variables were all operationalized in a manner that equated positive scores with leftist positions and negative scores with rightist positions. For most of the issues, moreover, leftist was interpreted as favoring "greater governmental activity" in the issue area. According to the assumption built into our scoring, therefore, all thirteen issue orientation variables should correlate positively with the experts' ratings, which were also coded with the high values equated with leftism. Note that the correlations are indeed positive for all but three of the thirteen variables, and only two display negative correlations consistently. "Support of the military," BV506, is consistently and significantly (at the .05 level) correlated negatively with both experts' ratings of party ideology. Although the granting of increased financial support to the military appears to be consistent with a general position favoring greater governmental activity to solve social problems, the parties of the world are able to separate this particular issue area from their basic philosophy of government, with parties of the right and left switching positions, as it were, when the issue of supporting the military is at hand. Thus the "popular image" of rightist parties being promilitary and leftist parties being antimilitary tends to be confirmed, contrary to the logic of our scoring. "Electoral participation," BV511, is inversely related to both ratings but significantly so only for the Soviets. For both groups of experts, high electoral participation is also not a hallmark of leftism.
The other issue that deviates from the consistent pattern of positive correlations with the experts' ratings is "supranational integration," BV509. Its correlation is negative in one instance and positive in the others. All of the correlations, moreover, are so low in magnitude that they fail to be significant at the .05 level. Unlike "support for the military," which appears only to have been scored in the "wrong" direction but does relate to the left-right continuum, "supranational integration" appears to have no place whatsoever in the left-right rankings of parties by either set of experts.
As we begin to examine more carefully the U.S. and Soviet patterns in comparison with one another, some of the minor deviations begin to gain importance, leading us toward an understanding of what we did not know before: precisely which factors tend to influence American and Russian evaluations of political parties as leftist and rightist. Issues with high correlations across the columns in Table 6.18 signal the existence of important factors in the ideological ratings of political parties by both sets of judges. The factor that appears to loom largest in their judgments is the parties' positions on the issue of "government ownership of the means of production," which correlates the highest with the ideological rankings of communists and capitalists alike. Although the next two issues are also economic in character and also have substantial correlations with the Soviet and U.S. ratings, they seem less important than the noneconomic but politically strategic variable, "East/ West alignment," whose correlations are second only to BV501. Thus it appears that a party's position on global politics contributes independently of its economics to its placement on the left-right scale.
If we limit ourselves to accepting as important ingredients of common left-right judgments only those issues that correlate consistently above .45 in the table, our search is satisfied by only "economic planning," "distribution of wealth," and "social welfare"--in addition to BV501 and BV507 already discussed. It seems that the Russian and American experts rely on these five factors in approximately the same extent in making their judgments. But, looking further, we find that the Soviets are somewhat more likely to involve in their ratings, the parties' positions on such issues as "secularization of society," "support of the military," and "national integration." The U.S. State Department, on the other hand, is less likely to reflect any of these political issues in their ratings, which are based more on party economics. Thus there are important similarities in the experts' bases of judgment but significant differences as well. By and large, the Soviet evaluation is more sensitive to a wider range of issues than is that of the United States, which, ironically, tends to classify parties on a more economic--even Marxist--basis.
Despite our earlier observation that the signs of the correlations largely confirmed our assumption in scoring the parties on a left-right basis, careful study of the magnitude of the correlations leads one to suspect that "supranational integration" is not the only variable with a poor fit to the left-right continuum. The correlations of the last five variables are at best only weakly related to the experts' ratings. Although there is some element of commonality among these variables and an underlying left-right dimension, it is not great, and one suspects there is some other dimension that might hold more in common with the errant variables.
A general method for identifying such underlying dimensions is factor analysis, and a factor analysis of the entire matrix of intercorrelations did prove helpful. (See Chapter 14.) In brief, the analysis disclosed four factors accounting for 70 percent of the variance among the thirteen issues but only two major dimensions. The first was the postulated left-right dimension, which encompassed the seven variables noted in Table 6.18. The second dimension extracted high loadings from these four variables: "support of the military," "electoral participation," "protection of civil rights," and "interference