¹ It is important to note from the very beginning that I am assuming that differences in community life or involvement between the urban and rural sectors, or between places of different magnitude of population, are of potential importance to political participation in some circumstances. What these circumstances may be, how they relate to the effects of urban-rural differences in other factors such as political competition or social structure—for example, the distribution of educational attainment—and what the effects may be for different kinds of attitudes and behaviors in one nation are the topics to be discussed.

The urban-rural disparities in turnout may also contribute to a particular configuration of party support. This has probably been the situation in postwar Japan, where it is frequently assumed that high rural turnouts favor Conservatives in specific relevant constituencies, a practice which is linked with mantenance of a dominant party system.

² Campbell, Angus, Converse, Philip E., Miller, Warren E. and Stokes, Donald E., The American Voter (New York: Wiley, 1960), p. 412Google Scholar; and Muller, Edward N., “Cross-National Dimensions of Political Competence,” The American Political Science Review 64 (September, 1970), 792–809CrossRefGoogle Scholar, especially 797. Actually, Muller's findings on the United States differed with those of The American Voter but were based on different specific question items. The Campbell et al. finding, however, was supported by Muller's evidence on differences between rural areas and intermediate size cities.

Although evidence from other places is contradictory. as will be seen, earlier survey evidence on sectoral tendencies in voting turnout in the United States from the Survey Research Center parallels their findings on involvement. See Campbell, Angus, Gurin, Gerald, and Miller, Warren, The Voter Decides (Evanston: Row, Peterson, 1954), pp. 128–130.Google Scholar

³ Tingsten, Herbert, Political Behavior: Studies in Election Statistics (London: P. S. King, 1937), pp. 211–214Google Scholar, and Lerner, Daniel, The Passing of Traditional Society: Modernizing the Middle East (Glencoe: The Free Press, 1958), p. 63.Google Scholar For additional information where nation states are the unit of analysis see Russett, Bruce M., Alker, Hayward R. Jr., Deutsch, Karl W., Lasswell, Harold D., World Handbook of Political and Social Indicators (New Haven: Yale University Press, 1964), pp. 279–80.Google Scholar

⁴ Tingsten, , Political Behavior, p. 214.Google Scholar

⁵ For Britain see Gallup Analysis of the Election 1966, pamphlet (London: The Daily Telegraph, 1966), p. 27, and Jean, and Chariot, Monica, “Politisation et Dpolitisation en Grande Bretagne,” Revue Française de Science Politique, 11 (September, 1961), 609–41, especially 634.Google Scholar German trends are reported by Faul, Erwin in Wahlen and Wähler in Westdeutschland, ed. Sternberger, Dolf, Erbe, Friedrich, Molt, Peter, Faul, Erwin, (Villingen: Ring, 1960), p. 156Google Scholar and “Das Wahlverhalten verschiedener Bevolkerungsgruppen bei der Bundestagswahl 1965,” Wirschaft und Statistik 3 (1966), 165–72. French patterns are discussed in Lancelot, Alain, L'Abstentionnisme Electorale en France (Paris: Armond Colin, 1968), pp. 195–97Google Scholar, Kesselman, Mark, “French Local Politics: A Statistical Examination of Grass Roots Consensus,” American Political Science Review, 60 (December, 1966), 963–73CrossRefGoogle Scholar, and Tarrow, Sidney, “The Urban-Rural Cleavage in Involvement: The Case of France,” American Political Science Review, 65 (June, 1971), 341–57CrossRefGoogle Scholar, especially 344–46. Japanese voting turnout is discussed in detail in Kyogoku, Jun'ichi and Ike, Nobutaka, “Urban-Rural Differences in Voting Behavior in Postwar Japan,” Economic Development and Cultural Change, 9 (October, 1960), Part II, 167–85CrossRefGoogle Scholar, especially 170–72.

Comparable tendencies are also found in some American states today, or at earlier points in time. The main sources are Robinson, James A. and Standing, William, “Some Correlates of Voter Participation: The Case of Indiana,” Journal of Politics, 22 (February, 1960), 99–106CrossRefGoogle Scholar, Key, V. O. Jr., Southern Politics in State and Na tion (New York: Alfred Knopf, 1950), pp. 510–13Google Scholar, and Burnham, Walter Dean, “The Changing Shape of the American Political Universe,” American Political Science Review, 59 (March 1965), 7–28Google Scholar, especially 15–16.

⁶ Nie, Norman H., Powell, G. Bingham Jr., and Prewitt, Kenneth, “Social Structure and Political Participation: Developmental Relationships, Part I,” American Political Science Review, 63 (June, 1969), 361–78CrossRefGoogle Scholar, especially 368. The authors conclude that differences in social status and organizational life are the main factors in explaining variations in national participation levels outside of voting.

See also Muller, , “Cross-National Dimensions,” p. 797.Google Scholar

⁷ Nie, Powell and Prewitt, “Social Structure and Political Participation” is the main example of research of this kind.

⁸ For the core writings in this area see Deutsch, Karl, “Social Mobilization and Political Development,” American Political Science Review, 55 (September, 1961), 493–515CrossRefGoogle Scholar, and Daniel Lerner, Passing of Traditional Society, especially Chap. 2.

Deutsch's analysis treats the various components of the social modernization process as a cluster of interrelated forces and looks for their combined effects on political behavior. Lerner presents a sequential model wherein urbanization plays a special function in the early phases of modernization. The discusson that follows addresses itself primarily to the question of the utilizability of the participation elements of developmental models. I am not concerned here with the extensions of the discussion to cover the more elaborate models of democratic performance that have been developed from the Deutsch, Lerner, and other modernization analyses.

⁹ Actually, rural residents might develop an awareness of instrumental stakes through the effects of the commercialization of agriculture. See Deutsch, , “Social Mobilization,” p. 494.Google Scholar

¹⁰ Milbrath, Lester W., Political Participation (Chicago: Rand McNally, 1965), pp. 110–114Google Scholar and 128–30. Lane, Robert also uses the concept of centrality in Political Life: Why People Get Involved in Politics (Glencoe, Illinois: The Free Press, 1959), p. 196.Google Scholar

¹¹ See Tingsten, , Political Behavior, p. 214.Google Scholar Actually, the interpretation is attributed to a study of turnout in the Zurich area of Switzerland.

¹² Tingsten, , Political Behavior, p. 213.Google Scholar

¹³ Lancelot, , L'Abstentionnisme Electorale en France, pp. 198 and 200.Google Scholar

¹⁴ Jurg Steiner's comments are relevant at this point: “The hypotheses must be complex, and in addition to the number of inhabitants other factors must be included, such as form of settlement, social structure and transportation situation.” See Bürger und Politik (Meisenheim am Gian: Anton Haln, 1969), p. 150. For a brilliant consideration of the theoretical confusion inherent in many approaches to the question of urban-rural differences, see Pappi's, Franz Urban criticisms of the urban-rural continuum concept in Wahlverhalten und politische Kultur (Meisenheim am Gian: Anton Haln, 1970), pp. 90–101.Google Scholar

Lancelot's distinction between the effects of compact and dispersed settlement patterns in France is a case in point. Rokkan, Stein and Valen, Henry, in “The Mobilization of the Periphery: Data on Turnout, Party Membership and Candidate Recruitment in Norway,” Acta Sociologica, 6 (1962)CrossRefGoogle Scholar, fase. 1–2, pp. 111–158, also introduce refinements into an analysis of urban-rural participation patterns, and, among other things, pay special attention to the importance of intrafamilial role differences in the two sectors.

¹⁵ For example, higher levels of group memberships are found in small communities in Germany, according to “Die Organisationszugehörigkeit in Westdeutschland,” DIVO Pressedienst, August 1, 1964, p. 11.

¹⁶ The precise meaning of urban and rural designations in Japan should be made clear from the outset. Japanese government publications and survey research groups typically use administrative designations rather than size of population to designate urban and rural locations. The relevant Japanese administrative divisions include a metropolitan prefecture (Tokyo), cities, towns and villages. Tokyo and other large metropolitan centers are also appropriately grouped under the category “large cities” in some official reports and in many survey analyses. The category “city” refers in turn in these cases to intermediate administrative units, whose populations may be largely rural in some cases, as the result of the effects of administrative mergers carried out in the 1950s. Towns and villages are nowadays quite systematically rural in composition. In other words, they contain high ratios of persons employed in agriculture, and population densities are much lower than in particularly the large urban areas.

In the following comments and figures, I have used “urban” where evidence is from samples drawn from metropolitan or city populations, and “rural” for findings on persons living in towns and villages. Information from intermediate sized cities will be omitted for simplicity's sake in most cases where data for metropolitan areas were available. I have indicated this in the table footnotes.

¹⁷ Rural levels of speech attendance also exceeded those in the cities, according to Remmei, Kōmei Senkyo, Sōsenkyo no Jittai, 1961, p. 28.Google Scholar Joji Watanuki has reported similar trends using an index of participation that includes turnout, organizational participation, activity in solution of local problems, attendance at rallies and contacts with political elites. See his “Social Structure and Political Participation in Japan,” The Laboratory for Political Research, Department of Political Science, The University of Iowa, Report Number 32, May 1970, pp. 4–5. Watanuki's work is the major exception to the comments that follow on emphases in the literature on urban-rural differences in Japan.

¹⁸ See, for example, the discussion in Kyogoku, and Ike, , “Urban-Rural Differences,” pp. 171–72.Google Scholar These and other commentators on rural behavior have developed what is tantamount to a mobilizational or social influence model of participation and voting choice.

¹⁹ The Coleman approach, known as effect parameter analysis, is discussed in Coleman, James S., Introduction to Mathematical Sociology (New York: The Free Press, 1964)Google Scholar, chap. 6. It permits multivariate analysis of data that do not meet interval level assumptions. For an application in comparative analysis see Thompson, Kenneth, “Cross National Voting Behavior Research: An Example of Computer Assisted Multivariate Analysis of Attribute Data,” in Comparative Politics Series, ed. Eckstein, Harry and Gurr, Ted Robert (Beverly Hills: Sage Publications, 1970), Vol. I, publication number 01–003.Google Scholar

The Coleman analysis and more familiar contingency techniques were used with findings of several Prime Ministers' Office studies and my own 1964 data set. The Prime Minister's Office surveys, one of which was conducted under the auspices of the Kōmei Senkyo Remmei, were acquired from the Roper Public Opinion Center, Williamstown, Massachusetts.

My own survey of political attitudes was carried out in 1964 in Yokohama and in two rural areas, one in Kanagawa and one in Shimane. For purposes of economy the sample was stratified, but random sampling procedures were used in each district.

²⁰ For a discussion of the advantages and shortcomings of these reports in more detail, see Richardson, Bradley M., The Political Culture of Japan (Berkeley and Los Angeles: The University of California Press, 1973), chap. 1.Google Scholar

²¹ The comments here on these matters draw from many sources, among them notably, Dore, R. P., Land Reform in Japan (London: Oxford University Press, 1959)Google Scholar, chap. 14 and Vogel, Ezra, Japan's New Middle Class (Berkeley and Los Angeles: The University of California Press, 1963)Google Scholar, especially pp. 102–13. There are, in addition, substantial differences between the character of neighborhood life in different sections of the large cities, or there were a few years back. Also, provincial cities are in all appearances from the poll data much more like the rural sector in regard to the character of their community life than like the really large cities.

²² The nature of communications processes in the two sectors was discussed in some detail in Richardson, The Political Culture of Japan, chap. 6 (including notes 22, 24, 25 and 50). The really substantial variations in per-capita newspaper subscriptions between the cities and many rural districts was also noted.

For discussion of the two-stage theory, see, for example, Katz, Daniel and Lazarsfeld, Paul, Personal Influence (New York: The Free Press, 1964), chap. 14.Google Scholar

²³ For evidence on the very substantial impact on at least written media exposure of education, see, for example, Tōkyō to Iikai, Senkyo Kanri, Kōmei Senkyo Yoron Chōsa no Gaiyō, 1958, 12, 17, and 23.Google Scholar Only 20 per cent of primary school graduates and 27 per cent of those who completed middle school, for example, read about politics in the newspapers on a regular basis, in contrast with 78 per cent of college graduates. Differences in television and radio exposure to political content attributed to education were somewhat smaller, however.

²⁴ For the best discussion of the sources of leftist party support in different population groupings, including attention to the special trends within Japan's salaried middle class, see Watanuki, Joji, “Patterns of Politics in Present Day Japan,” in Lipset, Seymour M. and Rokkan, Stein, Party Systems and Voter Alignments (New York: The Free Press, 1967), chap. 9.Google Scholar

²⁵ An initial index of urban-rural differences of .098 for feelings about the relevance of national politics was reduced to .025 by controlling for the effects of urban-rural distributions of educational attainment. A sectoral difference rate of .098 in feelings that national politics was easy to understand was reduced to .043. (Underlining in the form of dashes indicates indexes significant at the .05 level of confidence. Solid lines indicate significance at the .01 level. No line indicates indexes that do not meet the .05 confidence criterion.)

The Coleman effect parameters or indexes are extremely simple to understand. In the bivariate case they represent simply the difference between the left and right cells of a normal two by two contingency table. In the case at hand, they are equivalent to the difference when the “yes” categories for rural respondents in Table 5 are substracted from the “yes” responses for urban residents. The plus and minus signs are the result of the numerical coding system assigned to the different categories; in the analysis here a minus sign will indicate that higher frequencies were found in the rural sector. In the multivariate case, the indexes are based on more complicated calculations, but they can be compared to the mean of the relevant differences in the various tables generated by introduction of additional controls in contingency table analysis. In this case, the residual index for the effects of residence is the weighted average of differences attributable to urban-rural residence with the different educational categories held constant. Unlike contingency table analysis, however, multivariate analysis using the Coleman model involves simultaneous controls for two or more independent variables.

²⁶ Controls for education also substantially diminished urban-rural differences in interest in politics in a secondary analysis of the data from Prime Minister's Office Survey 1898 (1963).

²⁷ These relationships between responses to remote and more proximate aspects of politics are also shown in information on people's knowledge about politics. Urban residents felt that they had known about the candidate they supported in gubernatorial elections prior to the election more commonly than rural people, according to my analysis of Prime Minister's Office Survey 1404, while the reverse was the case in regard to prior knowledge about mayoralty and assembly candidates. Moreover, controls for education substantially decreased the strength of the relationship between residence and knowledge at the gubernatorial level, but length of residence was the underlying contributant to development of prior knowledge about local figures. While this is certainly not surprising, the differences in communications and socialization processes for alternative levels of politics implied by these data are of critical importance. But simple proximity in the form of actual participation in electoral processes can be seen as important, too, permitting less educated persons in especially the rural sector to develop an interest in at least some aspects of national political affairs. Compare Lester Milbrath's proposition that “the more stimuli about politics a person receives, the greater likelihood he will participate in politics and the greater depth of his participation.” See Milbrath, , Political Participation, p. 39.Google Scholar

Elsewhere I have shown that variations in proximity to different levels of politics are the the most plausible cause of differential rates of attitudinal change in the postwar period. Proximity and relevance considerations would thus occupy a central position in a theory of mass attitudinal change. See Richardson, The Political Culture of Japan, chap. 2.

²⁸ Controlling for group memberships reduced an original urban-rural difference of —.181 in interest in local politics to —.106, concern about local election outcomes was reduced from —.044 to —.008 and comprehension of local politics was reduced from —.132 to —.056. The relative importance of the higher group involvement levels found in the rural sector can be seen from these figures, as well as the residual importance of simply living in the compact, rural communities. Interestingly, length of residence, which is a variable of importance in some Japanese studies, did not affect the basic urban-rural differences, despite the fact that far more persons had lived all of their lives in the rural sector than was the case in the cities.

²⁹ A factor analysis of attitude structures in Japan reported elsewhere did show that feelings of wanting something (instrumentalism) loaded on the same dimension as interest in politics, and the loadings were higher where specifically local sentiments were concerned. See Richardson, , The Political Culture of Japan, chap. 4, Table 4–5.Google Scholar

³⁰ It should be remembered at this point that most earlier studies of urban-rural differences in political behavior in Japan have typically relied on only one or two simple measures of attitudes toward national politics; this has resulted in a simplistic view of the nature of urban-rural political involvement.

³¹ Feelings of efficacy regarding national political life showed contrary patterns, as is shown in Richardson, , The Political Culture of Japan, chap. 6.Google Scholar This is largely the result of educational differences, and in the face of the prevailing cynicism can be seen mainly as constituting a formal normative response to postwar emphases on democratization of Japanese society and politics.

³² It should be noted that Robert Ward has forecast the presence of tendencies in political negativism of the kind reported here in his “Japan: The Continuity of Modernization” in Political Culture and Political Development, ed. Pye, Lucian W. and Verba, Sidney (Princeton: Princeton University Press, 1965), pp. 27–82.CrossRefGoogle Scholar

³³ The events of the postwar period in Japanese partisan politics are summarized in Scalapino, Robert A. and Masumi, Junnosuke, Parties and Politics in Contemporary Japan (Berkeley and Los Angeles: University of California Press, 1962), chap. 2.Google Scholar

³⁴ Leftist party supporters are not always more negative in orientation, according to the materials I have studied. But this is the case where national surveys were concerned, as in Remmei, Kōmei Senkyo, Kōmei Senkyo no Jittai, 1962, p. 135Google Scholar; Remmei, Kōmei Senkyo, Kōmei Senkyo no Jittai, 1958, p. 71Google Scholar; and Remmei, Kōmei Senkyo, Dairokkai Tōitsu Chihō Senkyo to Yūkensha, 1967, p. 309.Google Scholar

³⁵ The analysis in question utilized data from Prime Minister's Office Survey 1898.

³⁶ Interviews with household association leaders in Yokohama provided some additional evidence on this phenomenon. To be sure, some of the laments about the declining levels of community feeling from these local leaders could be discounted as merely unsubstantiated nostalgia. But the leaders also stated that increasing proportions of persons within particular neighborhoods were unknown to the leaders themselves and had no contact with local organizations. It is noteworthy that these evaluations were concentrated in Yokohama's shitamachi (downtown area), where neighborhoods are typically assumed to be more cohesive than in other urban settings.

³⁷ Interestingly an initial urban-rural index of — .034 in regard to evaluations of national politicians' responsiveness increased to —.088 when education was controlled; analogously, the index for assessments of local responsiveness increased from a basic. —.184 to —.242 after controls for education were introduced.

Something was clearly different about rural residence in these instances, even though organizational involvement was not the contributing element. Perhaps communitywide feelings about responsiveness, reflecting stable linkages between local influentials and specific politicians, are the basis for the differences here.

³⁸ Evidence on the urban-rural distributions of feelings that accord with democratic postulates—such as “politics moves by our power” or “it's better to pay attention to what's going on in politics than just follow leaders”—supports this observation, and there is clearly some attitudinal ambivalence in the more normative dimensions of efficacy and appraisals of performance. A similar ambivalence can be observed among well educated persons. For relevant findings see Kanri Iinkai, Fukushima ken Senkyo, Chihō Senkyo no Jittai, 1963, p. 176Google Scholar and Remmei, Kōmei Senkyo, Seiji Ishiki to Senkyo Kōdō no Jittai: Yamanashi ken Toshi to Nōson, 1966, p. 35.Google Scholar

³⁹ There is some indication that this popular expectation is found in reality in the reports from the Institute for Mathematic Statistics' national character study, although urban-rural differences were not enormous. See, as one example, Kenkyūjo, Tōkei Sūri, Kokuminsei no Kenkyū: Dai Niji Chōsa (Tōkyō: Tōkei Sūri Kenkyūjo, 1959), pp. 55 and 58.Google Scholar

⁴⁰ See the extensive findings of these matters reported in ray The Political Culture of Japan, chap. 4.

⁴¹ This pragmatism is also shown in evidence I have discussed elsewhere regarding the content of both discussions of politics and specific requests for one's vote; in all instances, the open-ended questions I used to probe these matters resulted in spontaneous references to local problems more than to broad national issues. See my “Political Attitudes and Behavior in Contemporary Japan,” unpublished Ph.D. dissertation, The University of California, Berkeley, 1966, p. 90.

⁴² Deutsch's writings are an exception in part, as observed earlier. But Deutsch's model postulates that the important linkage wil be the commercialization of agriculture, with consequent development of concerns relating to general agrarian problems, while the evidence from Japan indicates that specifically local and parochial interests have precedence over other concerns. For evidence suggesting that these themes in rural outlooks have precedents in prewar days, see Fairbank, John K., Reischauer, Edwin O. and Craig, Albert M., East Asia: The Modern Transformation (Boston: Houghton Mifflin, 1965), p. 520.Google Scholar

The presence of growing sentiments of an instrumental kind in the rural sector has been reported earlier by Watanuki, Joji in his “Political Attitudes of the Japanese People,” The Sociological Review Monograph No. 10 (University of Keele, September 1966), pp. 166–168.Google Scholar

⁴³ There are other interpretations for this trend in sectoral feelings of instrumentalism in addition to those advanced already. Feelings of deprivation may motivate acute instrumental sensitivities in many areas of the country, perhaps especially the less affluent ones. It is also possible to speculate on the effects of the higher levels of property ownership that are found in the rural areas, although the only information available in this area is limited to statistics on home ownership. See Tōkeikyoku, Sōrifu, Nihon Tōkei Nenkan 1967, pp. 434–35.Google Scholar

⁴⁴ The basic index for the effects of residence was —.304. When controls for group membership levels were introduced this was reduced to —.244.

The role of organizational involvement here is some what analogous to that reported for France by Sydney Tarrow in “The Urban-Rural Cleavage …,” 353—355. But the consequence is still locally oriented instrumentalism in Japan in many instances, and this would appear to be different from the more general focus of rural pragmatism in the French example.

⁴⁵ I have followed the precedent set by Sidney Verba, Norman H. Nie. and Jae-on Kim, who differentiate between different forms of participation in “The Modes of Democratic Participation: A Cross-National Comparison,” in Eckstein, and Gurr, , Comparative Politics Series (Beverly Hills: Sage Publications, 1971), Vol. II, publication number 01–013.Google Scholar

⁴⁶ Possession of the various attitudinal perspectives was also a little less instrumental in motivating people to go out and ask others for their vote, a tendency that accords with the familiar mobilizational interpretation of participation in Japan.

⁴⁷ An example is Kanri Iinkai, Aichi ken Senkyo, Senkyo Ishiki to Tohyō Kōdō Henyō (Nagoya: 1963), p. 37.Google Scholar

⁴⁸ It is important to remember that the analysis and interpretation focuses on the differences between urban and rural residents not accounted for by the relevant attitudinal patterns. The reader should not forget the fact that many rural as well as urban voters were in fact psychologically involved. Still, more persons who turned out in national elections in the rural districts were not involved than was the case in the city.

Watanuki has presented a similar interpretation of rural participation in his “Social Structure and Political Participation in Japan.” The findings were based on a path analysis of data from the Verba group crossnational study.

⁴⁹ One of the best examples of this kind of relationship is found in Iinkai, Ōsaka fu Senkyo Kanri, Kōmei Senkyo Undō no Kōka Sokutei (Ōsaka: 1961), 32, 44.Google Scholar Longer residence led in this case to heightened exposure to campaign communications and to higher voting levels in a House of Representatives election. However, the linkages between attitudes responsive to the effects of mobility and various kinds of political behavior were not examined. The implications of the study were also limited by the mainly urban composition of the sampled population. It should also be noted that length of residence in my own study did not prove to have a strong relation to the urban-rural differences in political culture, but this may have been due to the simplicity of the dichotomous measure which tapped only lifelong residence and its absence.

⁵⁰ This tendency is shown by figures reported in my “Political Behavior and Attitudes in Contemporary Japan,” p. 39, and by various Japanese sources. For a qualitative assessment of women's comparatively greater involvement in local community life in some urban districts see Vogel, Ezra, Japan's New Middle Class, pp. 102–13.Google Scholar