¹ The term “Mainlanders” is used in this article to refer to Chinese who have emigrated to Taiwan from the Mainland since the end of the Second World War and to their children. The term “Taiwanese” is used to refer to Chinese whose ancestors emigrated to Taiwan prior to the Second World War—usually prior to Japanese administration of the island in 1895. Together, these two groupings comprise about 98 per cent. of Taiwan's population, the remainder being aborigines, living mainly on the mountainous east coast of the island. Virtually all of the students whose views are reported in this article were born on Taiwan or arrived there before reaching school age.Google Scholar

² The Spearman Rho is a measure of the association between two sets of rankings, computed by adding together the squares of the differences between these two sets of rankings, and converting the result into a score ranging from +1·0 (perfect positive association—both sets of rankings identical) through 0 (no association) to −1·0 (perfect negative association—each set of rankings precisely inverse to the other). It is possible to estimate statistically the probability that associations of the size noted would occur by chance, given the number of ranks involved. With 21 ranks involved, for instance, there is less than one possibility in 1,000 that a Rho of ·87 would occur because of chance alone. Thus we can refer to this association as significant at the ·001 level. By convention, associations are normally referred to by social scientists as statistically significant only when there is no more than one possibility in 20 (·05 level) that they could have been due to chance alone.Google Scholar

³ Two partial exceptions to this general statement appear in Chart 1. The population density of the cities is naturally higher than that in the countryside despite the greater concentration of Mainlanders in the cities. But both urban and rural districts with relatively high proportions of Mainlander residents have lower population densities than urban and rural districts with low proportions of Mainlanders. The death rate for the high-percentage-Mainlander counties is slightly higher than that for low-percentage-Mainlander counties, but this is due to the greater number of aborigines—whose death rate is considerably higher—living in the small high-percentage-Mainlander counties of Hualien and T'aitung on the east coast.Google Scholar

⁴ This Spearman rank order correlation is significant at the ·001 level. That is, there is less than one possibility in 1,000 that the degree of association shown by these two sets of rankings is due merely to chance.Google Scholar

Allan Cole reports that when the Taiwanese Kao Yu-shu was re-elected mayor of Taipei in 1964, “according to one estimate, four out of five Taiwanese and one out of seven mainlander voters cast ballots for Kao”. “Political Roles of Taiwanese Entrepreneurs,” Asian Survey (Berkeley, Calif.), Vol. 7, No. 9 (09 1967), pp. 645–654. The quotation is from p. 651. Since then, Taipei has been made a special municipality, and though Kao remains its mayor, he is responsible to the Central Government rather than to the voters.Google Scholar

Cole also notes the important role in the island's economy played by a group of prosperous Taiwanese businessmen. But this is not inconsistent with the hypotheses advanced here concerning the relative wealth of the masses of Mainlanders and Taiwanese.Google Scholar

⁵ Ministry of Education, Taiwan Provincial Government, T'ai-wan-sheng chiao-yu t'ung-chi, 1965–66 (Taiwan Educational Statistics, 1965–66 Academic Year) (Taichung, 1966), pp. 18–19.Google Scholar

⁶ Ibid. pp. 28–33.

⁷ Ibid. pp. 76–77.

⁸ These questions were adapted from a series used to study the phenomenon of opinion leadership in Decatur, Illinois, in 1945. See Elihu, Katz and Lazars-feld, Paul S., Personal Influence (Glencoe, Illinois: The Free Press, 1955).Google Scholar

⁹ All but the last (·23) of the Spearman rank order correlations given in this paragraph indicate relationships significant beyond the 0·5 level (that is, there is less than one possibility in 20 that the extent of association shown could be due to chance alone). Those higher than ·71 are significant at the ·01 level. The correlation between the Chinese and American students' rankings of “instrumental” values is not significant. The specific content of these sets of values, and the rankings assigned them by Taiwanese and Mainlander students, are discussed below.Google Scholar

¹⁰ These items were adapted from Hadley Cantril, The Pattern of Human Concerns (New Brunswick, New Jersey: Rutgers University Press, 1965).Google ScholarPubMed

¹¹ See Milton Rokeach, “Attitudes, Values and Political Behavior,” paper presented at the 62nd annual meeting of the American Political Science Association, 9 September 1966. Rokeach defines a “terminal value” as “an idealized goal or end-state of existence,” and an “instrumental value” as “an idealized mode of conduct.” Ibid. p. 11. See also Rokeach's book Beliefs, Attitudes and Values—A Theory of Organizational Change (San Francisco: Jossey-Bass, 1968).Google Scholar

¹² Only differences averaging one-quarter of a ranking or more are mentioned here, and these are given in order of the size of the differences.Google Scholar

¹³ Gabriel, Almond and Sidney, Verba, The Civic Culture (Princeton, New Jersey: Princeton University Press, 1963).Google Scholar

¹⁴ Ibid.“The competent citizen has a role in the formation of general policy. … The competence of the subject is more a matter of being aware of his rights under the rules than of participating in the making of the rules. And though the subject may attempt to make the government official responsive, he appeals rather than demands” (p. 214). See also pp. 12–26 and pp. 380–382.Google Scholar

¹⁵ See Herbert, McClosky, “Conservatism and Personality,” The American Political Science Review (Baltimore), Vol. LI, No. 1 (03 1958), pp. 27–45Google Scholar; and the same author's “Consensus and Ideology in American Politics,” The American Political Science Review, Vol. LVIII, No. 2 (06 1964), pp. 361–382.Google Scholar

¹⁶ “Consensus and Ideology in American Politics,” The American Political Science Review, p. 380. Our small sample of Mainlander high school students also had a slightly higher acquiescence score than their Taiwanese counterparts.Google Scholar

¹⁷ Richard W. Wilson's forthcoming study of political socialization in three elementary schools in the Taipei area indicates that Mainlander students at each of these three schools felt a greater degree of identification with President Chiang K'ai-shek than did their Taiwanese fellow students. “Rural-Urban Differences in Childhood Political Socialization on Taiwan,” paper presented to the Research Conference on Urban Society and Political Development in Modern China” (St. Croix, Virgin Islands, 28 December 1968–3 01 1969), p. 42.Google Scholar

¹⁸ The differences between Taiwanese high school and college students on these two items are significant at the ·05 and ·01 levels, respectively, despite the relatively small number (94) of Taiwanese high school students in the sample.Google Scholar

On the differences in values and attitudes between high school and college students on Taiwan, see Sheldon Appleton, “The Political Socialization of College Students on Taiwan,” paper presented at the 65th annual meeting of the American Political Science Association, 6 09 1969.Google Scholar

¹⁹ The New York Times, 14 07 1968, p. 7.Google ScholarPubMed See also Ministry of Education, Government of the Republic of China, Chung-hua, min-kuo chiao-yu t'ung-chi, 1966 (Educational Statistics of the Republic of China, 1966) (Taipei, 1966), pp. 84–93.Google Scholar

²⁰ Wilson, Richard W., “A Comparison of Political Attitudes of Taiwanese Children and Mainlander Children on Taiwan,” Asian Survey, Vol. 8, No. 12 (12 1968), pp. 98–1000. The quotation is from p. 998.CrossRefGoogle Scholar

²¹ Since the first draft of this article was prepared, a study which tends to support this surmise has come to my attention. A survey of 1,500 elementary and junior high school students in Philadelphia and Pittsburgh shows significant differences between black and white students on a number of political dimensions. The black children show greater alienation from the political system than white children, and differences between the two groupings increase with age. Edward S. Greenberg, “Black Children and the Political System: A Study of Socialization to Support,” paper presented at the 65th annual meeting of the American Political Science Association, 4 09 1969.Google Scholar

²² Bernard, Gallin, “Political Factionalism and Its Impact on Chinese Village Social Organization on Taiwan,” in Marc J. Swartz (ed.), Local-Level Politics (Chicago: Aldine, 1968), p. 398.Google Scholar

²³ Support for such a measure was expressed by 89 per cent. of a quota sample of adult residents of Taipei surveyed two years before the measure was adopted. Among those with children currently enrolled in elementary school, an even higher proportion (93 per cent.) approved. See Shu, Ch'ien and Ch'i, Ching-yao, Ta-chung ch'uan-po tsai cheng-fu kung-kung kuan-hsi chung ti kung-neng (The Role of Mass Communications in Government Public Relations) (Taipei: National Cheng-chih University School of Public Administration and Business Management Educational Centre, 1965), pp. 40–41.Google Scholar

Recently, in addition, a decision has been announced to reduce the size of the army and to streamline its operations.Google Scholar