Table 5 reports estimated coefficients, statistical significance, and odds ratio for each of the four logit models. The dependent variable is whether a Taiwan-born immigrant wrote Taiwanese as their race (yes=1) in the two censuses. Models 1 and 2 include temporal variables only, while models 3 and 4 add variables to control for human capital, economic status, English proficiency, and residential location.
Model 1 has no covariates. Overall, the results suggest that Taiwanese identity is highly stratified by cohorts, which is largely consistent with the bivariate analysis. In addition, there is a significant effect associated with the census year, indicating that a large proportion of Taiwan-born immigrants switched from "Chinese" to "Taiwanese" over time. The main effect of birth and immigrant cohorts applies to the reference cohort, which is "ages 25-34 in 1990 that arrived in the 1980s."
In model 2 of Table 5, we added age effect, duration effect, and joint immigrant cohort and birth cohort effects. A chi-square test (i.e. difference in model 2statistic) shows that the interactive model clearly improves over the baseline model, indicating that selected birth cohorts in specific immigrant cohorts have unique histories of identifying with Taiwanese or
"age-at-immigration" effect. It also shows that birth and immigrant cohorts
Table 5. Logistic Coefficients of Taiwanese Identity among Taiwan-born Immigrants
Model 1 Model 2 Model 3 Model 4
Variable Coeff. Odds
Ratio Coeff. Odds
Ratio Coeff. Odds
Ratio Coeff. Odds Ratio
Constant -1.593*** -1.635*** -1.667*** -1.687***
Year (1990 = 0; 2000 = 1) 0.384*** 1.468 0.552*** 1.736 0.546*** 1.726 0.499*** 1.648 Birth cohort (BC)
15-24 in 1990, 25-34 in 2000 (BC1) 0.503*** 1.653 0.709*** 2.031 0.856*** 2.353 0.798*** 2.220 35-44 in 1990, 45-54 in 2000 (BC3) 0.463*** 1.588 0.404*** 1.499 0.399*** 1.491 0.367*** 1.443 45-54 in 1990, 55-64 in 2000 (BC4) 1.461*** 4.311 1.344*** 3.834 1.348*** 3.849 1.261*** 3.528 55-64 in 1990, 65-74 in 2000 (BC5) 1.276*** 3.584 1.060*** 2.887 1.090*** 2.975 1.038*** 2.825 Immigrant cohort (MC)
1970s (MC2) -0.380*** 0.684 -0.550*** 0.577 -0.530*** 0.589 -0.556*** 0.573
Before 1970 (MC3) -0.135* 0.873 -0.358 0.699 -0.237 0.789 -0.212 0.809
Age effect with time (Y x BC)
15-24 in 1990, 25-34 in 2000 (BC1) -0.389 0.678 -0.514*** 0.598 -0.490*** 0.613
35-44 in 1990, 45-54 in 2000 (BC3) -0.030 0.970 -0.021 0.979 -0.020 0.980
45-54 in 1990, 55-64 in 2000 (BC4) -0.103 0.902 -0.096 0.908 -0.049 0.952
55-64 in 1990, 65-74 in 2000 (BC5) -0.212 0.809 -0.211 0.810 -0.163 0.850
Duration effect with time (Y x MC)
1970s (MC2) -0.153 0.858 -0.145 0.865 -0.125 0.882
Before 1970 (MC3) -0.172 0.842 -0.156 0.856 -0.114 0.892
Joint immigrant cohort and birth cohort effect (MC x BC)
1970s immigrant cohort
BC1 -0.014 0.986 -0.010 0.990 0.067 1.069
BC3 0.363** 1.438 0.356** 1.428 0.403** 1.497
BC4 0.379* 1.461 0.381* 1.464 0.489** 1.631
BC5 0.756*** 2.129 0.729** 2.072 0.785*** 2.192
1960s or earlier immigrant cohort
BC3 -0.266 0.767 -0.325 0.722 -0.328 0.720
BC4 0.572** 1.772 0.475* 1.607 0.539* 1.714
BC5 0.803** 2.232 0.669* 1.952 0.695* 2.003
Educational attainment
High School Dropout (Educ1) -0.107 0.898 -0.073 0.930
4-year College Degree (Educ3) 0.155*** 1.167 0.173*** 1.188
English proficiency
Speak English only -0.417*** 0.659 -0.382*** 0.682
Speak English well -0.118* 0.889 -0.118* 0.888
Economic Status
Median High: 4 to 5 times poverty line
(pov3) 0.122 1.129 0.084 1.087
Median Low: 1 to 4 times poverty line
(pov2) 0.124** 1.132 0.103* 1.108
Poor: below poverty line (pov1) -0.124* 0.883 -0.131* 0.877
Local concentration (percent of total PUMA population)
Mainland Chinese immigrants -0.084*** 0.919
Taiwan-born immigrants 0.246*** 1.279
Metropolitan areas
Los Angeles CMSA 0.290*** 1.336
San Francisco CMSA -0.321*** 0.725
Washington DC CMSA -0.025 0.976
New York CMSA 0.008 1.008
df 7 20 27 33
LR chi2 928 1002 1054 1443
Log likelihood -10,184 -10,147 -10,121 -9,926
Pseudo R2 0.044 0.047 0.050 0.068
No. of observations 18,529
Note: The reference group for birth cohort in 1990 is "25-34 in 1990 and 35-44 in 2000"; for duration effect, it is immigrants arrived in the 1980s; for age effect, it is "25-34 in 1990"; for age-at-arrival effect, the reference group is "ages 25-34"; for immigrant cohort in 1990 the reference group is Taiwan-born immigrants came to the United States in the 1980s; for educational attain-ment, it is "High school dip. w/ college"; for English proficiency, it is "speaks English not well or not at all"; for poverty, it is
"high economic status"; for metropolitan areas, it is outside the four metropolitan areas.
Source: Census 5% PUMS, 1990 and 2000
*p < .05; **p < .01; ***p < .001 (two-tailed tests)
have progressed differently over time. However, the presence of the interaction terms (Y x BC, Y x MC, and BC x MC) has little impact on the coefficient estimates of birth and immigrant cohorts.
The interpretation of age effect (Y x BC) and duration effect (Y x MC) has to be related to the reference group. Although all birth and immigrant cohorts have negative signs indicating they have smaller increases than the reference group, only the age effect of BC1 is statistically significant. In other words, the youngest birth cohort has the smallest increase in their likelihood. In comparison with other cohorts, young immigrants appear relatively more "forgetful" than "rebellious" in their identity shift.
Adding the interaction term (BC x MC) reveals that the age-at-immigration effect is strongest in older cohorts who arrived here early. In other words, the two older birth cohorts (BC4 and BC5) in both the 1970s (MC2) and the pre-1970s (MC3) immigrant cohorts are much more likely to identify themselves as Taiwanese as indicated by the additive joint effects of birth cohort and immigrant cohort. The strong age-at-immigration effect may reflect that many old Taiwan-born immigrants were political dissidents and came to the U.S. to seek political freedom.
To address the question of how much the differences in Taiwanese identity are associated with human capital and socioeconomic correlates, we added a set of covariates in model 3. The variables of educational attainment and English proficiency are collected at the individual level, while poverty status reflects the economic condition of the household in which the individual lives. Again, a chi-square test reveals that adding the covariates presents a better goodness of fit over model 1, indicating that the likelihood of Taiwanese identification varies between socioeconomic groups and across metropolitan areas. The inclusion of the covariates, however, has little impact on the coefficient estimates of the temporal
variables presented in model 2.
Human capital and socioeconomic factors are significantly correlated with identity. Compared with the reference group16, Taiwan-born immigrants who have a 4- year college degree are most likely to identify themselves as Taiwanese. In contrast, speaking English only and living below the poverty line lowers the probability. The results show that human capital in general has a positive effect. Rather than advancing Taiwan-born immigrants' socioeconomic interest in the U.S., the acknowledgement of Taiwanese identity appears to be a symbolic move rather than for the purpose of resource mobilization. Evidently, less resourceful Taiwan-born immigrants are no more likely to consider themselves as Taiwanese than well-off Taiwanese. Furthermore, assimilation diminishes the propensity for doing so.
Model 4 adds location variables. The coefficient estimates of temporal variables do not change significantly from those of previous models.
Results show that a higher concentration of Taiwan-born immigrants is positively related to the probabilities of Taiwanese identity. In other words, living in Taiwan-born immigrant communities reinforces Taiwanese identity, possibly due to better access to ethnic politics and media. This finding supports the assimilation perspective in that ethnic concentration slows down assimilation and leads to a stronger ethnic identity.
Surprisingly, Taiwan-born immigrants who live in areas of higher concentration of [mainland] Chinese immigrants are less likely to claim Taiwanese identity. In those areas, Taiwan-born immigrants may have more opportunities to communicate with [mainland] Chinese immigrants. This finding seems to show that they do not use their Taiwanese identity to set up a boundary and separate themselves from [mainland] Chinese immigrants.
Despite their socioeconomic and political differences, there is not a lot of
animosity between Taiwanese immigrants and [mainland] Chinese immigrants. At the same time, living with other Taiwan-born immigrants enhances Taiwanese identity. This finding is consistent with the assimilation theory. However, large geographical variation remains after controlling for other relevant factors. Taiwan-born immigrants in Los Angeles, which is the largest U.S. destination for Taiwan-born immigrants, are much more likely to claim their Taiwanese identity. Taiwan immigrant organizations in Los Angeles have appeared particularly successful in resource mobilization (Tempest 2004; Overseas Compatriot Affairs Commission 2006).
The pattern of effects is sufficiently complex that we evaluate them graphically. Figure 1 depicts the predicted probability of Taiwanese identification by Taiwan-born immigrants' birth and immigrant cohort, while holding other variables at their sample means. The predicted probability is first calculated for each observation based on the coefficient estimates of model 4, averaged, and then visually presented in Figure 1.
Figure 1 shows there are large temporal variations. The tendency of older cohorts (BC4 and BC5) to declare their Taiwanese identity in the census form is significantly higher than that of younger cohorts (BC1, BC2, and BC3). New arrival cohorts have a stronger propensity than earlier arrival cohorts.
There is an overwhelming period effect (year), which overpowers other forces. Noticeably, all birth cohorts have shown significant
16 The reference group for immigrant cohort in 1990 is Taiwan-born immigrants who came to the United States during the 1980s; for educational attainment it is "High school diploma with some college education"; for English proficiency it is "speaks English not well or not at all";
for poverty, it is "low" or those who are below the poverty threshold; for metropolitan areas, it is outside the four metropolitan areas." We choose this demographic group as the reference group because of their large sample size. The large size is useful for the regression model to converge. But the choice of reference group should have no impact on the result.
consistency in their increases in the propensity from 1990 to 2000. The concerted increase reflects the widespread impacts of heightened cross-Strait tension, the identity politics in Taiwan, and other factors.
Interestingly, even Taiwan-born immigrants who came to the U.S. for more than two decades still experienced a large increase in their propensity. The pervasiveness of identity change seems to reflect the power of
Figure 1. Predicted Probability of Taiwanese Identification among Taiwan-born Immigrants, 1990-2000
globalization, which reduces the distance between host and origin countries.
While all birth cohorts increased the propensity, there are large inter-cohort differences. The variation is largest across birth inter-cohorts. Young cohorts seem to start from a relatively low level and increase more slowly.
In addition, there is a joint effect between birth and immigrant cohorts.
Assimilation is present here: Young cohorts (BC2 and BC3 who arrived prior to 1970; BC1 and BC2 who arrived in the 1970s; and BC1 who arrived in the 1980s) have a smaller increase than other birth cohorts. These groups are likely to be educated in the U.S. and are less involved in ethnic politics. In addition, newly arrived immigrants are more likely to identify themselves as "Taiwanese". They also have a larger increase in their Taiwanese identity during the 1990s. As a result, the gap between birth cohorts is largest among early arrivals. There is also a strong joint effect between immigrant cohort and birth cohort. Older cohorts who came in the 1970s or earlier have the strongest affinity with their Taiwanese identity.
VI. Conclusions
This paper has investigated the rise of Taiwanese identity among Taiwan-born immigrants in the United States. Application of the double cohort method enables us to reveal how identity switching occurs within a complex temporal structure. The extent to which Taiwan-born immigrants change their identification is not only affected by the political events in Taiwan, but is also embedded with their year of birth, year of arrival, and age at arrival. The identity formation is subject to the interplay of multiple forces. The study also demonstrates the usefulness of census data in the analysis of identity formation.
This study is couched in extensive literature. To a great extent, the
findings are consistent with the existing literature on ethnic identification.
That is, the identity of Taiwan-born immigrants is fluid and changeable.
There has been a surge in Taiwanese identity even among Taiwan immigrants who have been in the U.S. for more than two or three decades.
Primordial ties play a critical role in the identity formation of Taiwan-born immigrants, evidenced by the overwhelming period effect－a concerted increase in Taiwanese identity among all birth and immigrant cohorts. The shared birthplace seems to have synchronized all Taiwan-born immigrants in the identity shift. In addition, there is a strong evidence of cohort continuity in identity change, as all cohorts have progressed in a similar direction.
Assimilation plays an important role in the identity formation of Taiwan-born immigrants. First, for those who are acculturated or speak only English at home, they are less likely to identify themselves as Taiwanese.
Second, cohort analysis discloses significant variations in the trajectories of Taiwanese identity, showing a rapid rise among older cohorts and recent arrivals while indicating a relatively small increase among young people and more settled immigrants. Third, age-at-arrival is a strong determinant.
Evidently the BC2 and BC3 cohorts who came to the U.S. before the 1970s were least likely to claim Taiwanese identity and make changes over time.
Fourth, Taiwan-born immigrants have a lower propensity when they live outside Taiwan-born immigrant communities and outside Los Angeles－
the largest Taiwanese migration destination in the U.S. These findings seem to show that assimilation has started to blur the boundary between Taiwanese and [mainland] Chinese immigrants.
This research also contributes to the ethnicity literature. The existing literature on ethnic identity has primarily focused on the extent to which minority groups accept or reject given identities. The story is largely
oriented around the conflicts between the white majority and selected minority groups. In contrast, there has been little apparent tension between Taiwan immigrants and the white majority or between Taiwan immigrants and [mainland] China immigrants. Avenarius' study of Taiwan sub-ethnic immigrant groups in Orange County shows that growing up in Taiwan may not be a sufficient factor to bring immigrants with different sub-ethnic affiliations together in their new place of settlement, as she wrote: "[The]
Taiwanese socialize mainly with fellow native Taiwanese; Mainlanders spend their free time primarily with other Mainlanders" (Avenarius 2007).
Divisions seem more apparent in the Taiwanese diaspora communities, particularly between the older generation of Mainlanders and native Taiwanese.
For Taiwan-born immigrants, being Taiwanese appears to be a
"rebellious" or an "awakening" identity and a symbolic expression of public solidarity with their compatriots back home, influenced by growing Taiwanese nationalism and energized by heightened cross-Strait tension.
Even though political conflicts occur in places far away, globalization has shortened the distance between origin and host countries and enhanced the Taiwanese identity of Taiwan-born immigrants. Highly educated Taiwan immigrants have higher propensities to identify themselves as Taiwanese, which seems to confirm the facilitating effects of globalization. To the extent that ethnic politics may become a less potent force among immigrant children and second generation immigrants, further study is necessary.
Taiwanese identity appears more symbolic than substantive, because the strength of Taiwanese identity is negatively related to educational attainment and economic status. In addition, living with their mainland counterparts lowers the propensity, even though many less affluent Taiwan immigrants are in direct competition with [mainland] Chinese immigrants
for jobs. Taiwanese identity does not appear to be used for resource competitions with [mainland] Chinese immigrants. Nor is it used for further advancing Taiwan immigrants' economic interests. Unlike other minority groups documented in the literature, Taiwan immigrants have not been a target of socioeconomic oppression in the United States.
It is, however, puzzling to observe the large geographical variations between metropolitan areas in the U.S. Not only is the identification highest in Los Angeles and lowest in San Francisco, but the rate of identity shift is also largest in Los Angeles and smallest in San Francisco. Both metropolitan areas have large ethnic Chinese communities, although the relative size of Taiwan-born immigrants is significantly smaller in San Francisco. While regional differences in political mobilization and migration trends may have played a role, further research is clearly needed to better understand the large geographical variations.
In the absence of cross-strait antagonism, it is unclear whether the trend of rising Taiwanese identity will continue in the future. While Taiwanese identity will remain fluid and changeable, we may observe more mixed marriages between sub-ethnic Taiwanese groups and between Taiwanese and [mainland] Chinese immigrants. As a result of inter-group marriages and further assimilation to the U.S., Taiwan immigrants have the possibility to less emphasise on a country-specific identity and create more inclusive identities. The upcoming 2010 U.S. Census may reveal more interesting patterns, especially given that the KMT returned to power in 2008 and that cross-strait relations have become less confrontational since the last U.S. Census in 2000. A follow up study is necessary to find out if Taiwan-born immigrants have a greater or smaller sense of Taiwanese identity as a result of these changes. A qualitative study will also be helpful in further deciphering the forces behind the identity formation of Taiwan
immigrants in the U.S.
Acknowledgements
The authors would like to thank the financial support from Chiang Ching-kuo Foundation for International Scholarly Exchange. Points of view or opinions in this paper are those of the authors and do not necessarily reflect the official position or policies of the Foundation.
References
Ambrosio, T. (ed.) 2002. Ethnic identity groups and U.S. foreign policy.
Westport, Conn.: Praeger.
Alba, R. D. and J. R. Logan. 1992. "Assimilation and Stratification in the Homeownership Patterns of Racial and Ethnic Groups." International Migration Review 26(4): 1314-1341.
Alba, R. D. 1990. Ethnic Identity: The Transformation of White America.
New Haven: Yale University Press.
Alba, R. D. and V. Nee. 2003. Remaking the American Mainstream: Assimi-lation and Contemporary Immigration. Cambridge, Mass: Harvard Uni-versity Press.
Allen, J. P. and E. J. Turner. 1996. "Spatial Patterns of Immigrant Assimila-tion." Professional Geographer 48(2): 140-155.
Allio, F. 2000. "The Dynamics of the Identity Issue in Taiwan Understan-ding Uniqueness, Diversity and the Whole." China Perspectives 28:
43-50.
Avenarius, C. 2007. "Cooperation, Conflict and Integration among Sub-Ethnic Immigrant Groups from Taiwan." Population, Space and Place 13 (2): 95-112.
Barth, F. (ed.) 1969. Ethnic groups and boundaries: the social organization of culture difference. Boston: Little Brown and Company.
Barth, F. 2000. "Boundaries and Connections." Pp.17-36 in Signifying Iden-tities: Anthropological Perspectives on Boundaries and Contested Values, edited by A. Cohen. London; New York: Routledge.
Baum, J. and A. Sherry. 1999. "Identity Crisis." Far Eastern Economic Re-view 162: 21-23.
Bell, D. 1975. "Ethnicity and Social Change." Pp. 141-177 in Ethnicity:
Theory and Experience, edited by N. Glazer and D. P. Moynihan. Cam-bridge, Mass: Harvard University Press.
Brown, M. J. 2004. Is Taiwan Chinese?: The Impact of Culture, Power, and Migration on Changing Identities. Berkeley: University of California Press.
Carliner, G. 2000. "The Language Ability of U.S. Immigrants: Assimilation and Cohort Effects." International Migration Review 34(1): 158-182.
Chen, H. S. 1992. Chinatown No More: Taiwan Immigrants in Contempor-ary New York. Ithaca: Cornell University Press.
Corcuff, S. 2002. "Taiwan's 'Mainlanders', New Taiwanese?" Pp. 163-195 in Memories of the Future: National Identity Issues and the Search for a New Taiwan, edited by S. Corcuff. NY: Armonk; London: M. E. Sharpe.
Duncan, O. D. and S. Lieberson. 1959. "Ethnic Segregation and Assimila-tion." American Journal of Sociology 64(4): 364-374.
Ellis, M. and R. Wright. 1998. "When Immigrants Are Not Migrants: Coun-ting Arrivals of the Foreign Born Using the U.S. Census." International Migration Review 32(1): 127-144.
Eschbach, K. and C. Gomez. 1998. "Choosing Hispanic Identity: Ethnic Identity Switching among Respondents to High School and Beyond." So-cial Science Quarterly 79(1): 74-90.
Farley, R. 1991. "The New Census Question About Ancestry: What Did It Tell Us." Demography 28(3): 411-429.
Fisher, G. M. (1992). The Development of the Orshansky Poverty Thres-holds and Their Subsequent History as the Official U.S. Poverty Measure. Poverty Measurement Working Papers.
Gans, H. J. 1979. "Symbolic Ethnicity: The Future of Ethnic-Groups and Cultures in America." Ethnic and Racial Studies 2(1): 1-20.
Gans, H. J. 1982. The Urban Villagers: Group and Class in the Life of Italian-Americans. New York: The Free Press.
Geertz, C. 1963. "The Integrative Revolution." Pp. 103-130 in Old Societies and New States; the Quest for Modernity in Asia and Africa, edited by C.
Geertz. New York, Free Press of Glencoe.
Glazer, N. and D. P. Moynihan. 1975. "Introduction." Pp. 1-29 in Ethnicity:
Theory and Experience." edited by N. Glazer and D. P. Moynihan. Cam-bridge: Harvard University Press.
Gordon, M. M. 1964. Assimilation in American Life: The Role of Race, Re-ligion, and National Origins. New York: Oxford University Press.
Greeley, A. M. 1974. Ethnicity in the United States: A Preliminary Recon-naissance. New York: Wiley. Wiley Series in Urban Research.
Holdaway, J. 2007. "China: Outside the People's Republic of China." Pp.
355-370 in The New Americans : A Guide to Immigration since 1965, edi-ted by M. C. Waters, R. Ueda and H. B. Marrow. Cambridge, Mass.:
355-370 in The New Americans : A Guide to Immigration since 1965, edi-ted by M. C. Waters, R. Ueda and H. B. Marrow. Cambridge, Mass.: