Less than one-third of the parties in our study were scored for attraction, concentration, and reflection on this differentiator. As for religion, this was often due to overwhelming homogeneity of the society. Unlike religion, however, neither ethnicity, nor language, nor race easily allow for an alternative "intensity" conceptualization. This leads one to argue that many of the world's parties could be scored 1.0 on attraction and 0.0 on concentration reflecting the complete lack of communalism as a basis of party support. While this argument does not seem entirely unreasonable, it does leave one uncomfortable on realizing that the practice would result in major changes in the parties' statistics in Table 5.4. As the data stand now, parties show less attraction, more concentration, and less reflection of ethnic, language, or racial groupings than do groupings on any other cultural differentiator--and that seems to fit with our expectations concerning communalism as a basis of party support.
Geography becomes an especially important factor in party support when the grouping of people by space coincides with some other division within society, producing a situation of reinforcing cleavages. This confluence of spatial and social division usually harbors potential for bitter politics. The identification of regions was idiosyncratic with each country in the project, and again attention was given to uncovering differences among the parties rather than smoothing them over. Because election returns by regions usually suited our purposes of assessing party support, we were able to score about 85 percent of our parties on this differentiator, which places it second in coverage behind occupation. Its overall pattern of statistics for attraction, concentration, and reflection is also very similar to occupation, suggesting that regional groupings function much like occupational groupings in overall tendency to differentiate among parties.
One of the major bases for the inclusion of urban-rural as an aspect of social attraction is the situation created by the Industrial Revolution in most developed societies. The Industrial Revolution deepened rural-urban conflicts, and, according to Lipset and Rokkan, "in country after country [the Industrial Revolution] produced distinct rural-urban alignments in the national legislatures" (1967, p. 19). In underdeveloped countries, politics of any sort tends to be an urban phenomenon, and rural-urban conflicts have often not yet materialized at the national level. Nevertheless, the urban-rural dimension--like regionalism--reflects the spatial grouping of people and had the potential for reinforcing other social divisions, especially occupation, and thus invites our attention.
In coding parties on the urban-rural basis of their support, we were often able to utilize the same election data that provided our scores for regionalism, which was possible when urban areas were identifiable in the regional returns. As a result, we coded 70 percent or more of the parties for urban-rural divisions. Remembering that the comparisons involve many of the same cases but not all, we might still note that parties tend to be more attractive of urban-rural than occupational groupings but more concentrative of urban-rural groupings. This phenomenon betrays the clustering of farmers within specific parties that also do well in urban areas.
Barnes holds that "there can be little doubt that differences in formal education have political consequences. The evidence is compelling that persons of high education participate more, are more knowledgeable, feel more efficacious, and exhibit greater sensitivity to the ideological dimension of politics" (1966, p. 348). Converse feels that education is the prime predictor among all independent variables reflecting political interest, participation, and mobilization. He also states that "education shows remarkable discriminating power as a status measure in predicting to party position" (1968, p. 274).
Unfortunately, we were able to code only somewhat less than half the parties for their support among educational groupings, so our assessment of the values of the attraction, concentration, and reflection statistics for education must be limited. Despite this constraint on our comparison, we should note that the data in Table 5.4 seriously question the importance of education as a differentiator for the basis of party support. In general, the parties scored are more attractive and reflective of educational groupings than they are for any other cultural differentiator, and they are almost as low in concentration as for any other. Education may serve to stimulate political participation per se, but it does not seem to be a strong basis of party divisions.