The Political and Community Context of Immigrant Naturalization

Abstract

Becoming a citizen is a component of a larger process of immigrant incorporation into U.S. society. It is most often treated as an individual-level choice, associated with such personal characteristics as the duration of residence in the U.S., age, education, and language acquisition. This study uses microdata from Census 2000 in conjunction with other measures to examine aspects of the community and policy context that influence the choices made by individuals. The results confirm previous research on the effects of individual-level characteristics on attaining citizenship. There is also strong evidence of collective influences: both the varied political histories of immigrant groups in their home country and the political and community environment that they encounter in the U.S. have significant impacts on their propensity of naturalization.

The Political and Community Context of Immigrant Naturalization

Becoming a citizen changes an immigrant’s political status in the United States, conferring constitutional rights that significantly affect a person’s economic and civic incorporation into U.S. society. Traditionally, naturalization has been treated as an individual choice, and individual characteristics such as educational attainment, income, ability to speak English, and length of time in the United States are well-established positive predictors of immigrant naturalization (Liang 1994; Jones-Correa 2001a; Yang 1994; Cho 1999). These are similar to the predictors of other dimensions of social incorporation or assimilation of immigrants (Alba and Nee 2003).

This study focuses particularly on influences that people experience collectively in the process of naturalization. Immigrants do not approach the question of citizenship only as individuals, but also through shared experiences with those who came from the same country, who have settled in the same community, and who have their race and ethnic background (Yang 1994; Liang 1994; Jones-Correa 2001a; Bueker 2006). These common characteristics are at the heart of the variations in the context of origin and reception through which Portes and Rumbaut (1990) portray the immigrant experience.

In the broader literature on immigrant assimilation, dealing with behaviors ranging from language acquisition to residential mobility, researchers have often looked for effects of shared experiences. We argue here (see also Pantoja and Gershon 2006; Bloemraad 2002) that the process of obtaining citizenship is especially likely to involve such effects precisely because of its political nature. It is partly an outcome of engagement with political institutions, and it is driven by expectations about political status above and beyond social or economic benefits. As Jones-Correa (1999) puts it, “Clearly, from immigrants’ point of view, the state still matters.”

COLLECTIVE EFFECTS ON IMMIGRANT NATURALIZATION

Following from this perspective we study several explicitly political predictors of naturalization. We show for the first time that the restrictiveness of the electoral system of an immigrant’s state of residence (reflected in voter ID policies) impacts naturalization. We also offer the first test of an empowerment hypothesis for immigrant groups – the expectation that in communities where group members enjoy co-ethnic political representation, they are more likely to choose to become citizens.

Group Membership and Naturalization

A central collective factor is racial/ethnic group membership. Studies of naturalization have been surprisingly silent on the issue of racial/ethnic differences. In other domains, such as electoral participation, race is treated prominently (Uhlaner et al. 1989; Bobo and Gilliam 1990; Verba et al. 1993; Ramaskrishnan and Espenshade 2001; De la Garza 2004; Segura and Rodrigues 2006). These researchers suggest that race matters for political participation beyond what can be explained by compositional differences for a variety of complex reasons. Race may represent the unique political culture or political history of a group; it may represent a kind of collective identity or political consciousness that can be conducive to political activity; it may be reflective of group specific resources; and it may also be a marker of the experience of discrimination (Segura and Rodrigues 2006; Wong et al. 2005; Karpathakis 1999).

Yang (1994), using microdata from the 1980 census, offers competing hypotheses about variation in naturalization rates across racial/ethnic groups. The “discrimination” hypothesis is that minority immigrants (that is, all except non-Hispanic whites) will encounter greater barriers to naturalization and will be deterred from following the steps toward citizenship by a fear of rejection. Yang notes, for example, that until the 1950s Asian immigrants were treated by U.S. law as ineligible for naturalization. The “forced self-protection” hypothesis similarly presumes that minority immigrants encounter discrimination, but posits that they respond by seeking citizenship to improve their legal standing. A more specific “cultural differences” hypothesis is that Asians, due to their general cultural orientation and in particular the influence of Confucianism, will be reluctant to cut ties to their ancestral homelands by becoming U.S. citizens. Yang finds that black, Hispanic, and Asian immigrants are less likely to acquire citizenship than non-Hispanic white immigrants. But after controlling for individual-level differences in education, English language ability, and other factors, they are all significantly and substantially more likely to naturalize. He takes these findings as evidence for the forced self-protection hypothesis.

Community Effects

Several researchers have built on the theoretical proposition that social contacts of various sorts—whether with co-ethnics, immigrants, or predominantly native-born groups—shape attitudes, behaviors, and thus the propensity to naturalize (Liang 1994). They have assessed the role of the following variables at various geographic scales in shaping immigrants’ likelihood of naturalizing: residential location (metropolitan vs. rural location); population size or density; ethnic or racial concentration; racial residential segregation; and the density and population size of the immigrant population (Liang 1994; Yang 1994; Bueker 2006; Portes and Curtis 1987).

Many immigrants live in communities with large shares of foreign-born residents who share their racial or ethnic background. A common view is that residential contact with native whites can reflect or lead to greater social-cultural assimilation and increased access to information about U.S. society, which enhances the likelihood to naturalize. Liang (1994) and Bueker (2006) both present this argument. But an alternate theory (Portes and Curtis 1987, Pantoja 2005) is that contact with native born Americans or whites can result in experiences of discrimination. These could (if they do not lead to withdrawal) stimulate greater interest in naturalization as a means to contest discrimination.

Three studies provide mixed findings on the effect of concentration of the foreign-born or co-ethnic population. Measured at the level of the metropolitan region, the percentage foreign born is negatively related to naturalization (Bueker 2006). Measured at the level of the state, the proportion of the co-ethnic members who have naturalized is positively associated with naturalization (Yang 1994), and the overall share of group members in the state is positively associated with naturalization for Asians (Yang 2002).

Other studies have examined the effects of segregation from whites. Portes and Curtis (1987) found that Mexican immigrants living in communities with lower proportions of non-Hispanic whites were less likely to naturalize than those Mexicans living in communities with a higher proportion of whites. Liang (1994), using 1980 census microdata, also shows that higher segregation from whites generally leads to lower odds of naturalization, consistent with what he refers to as the “social enclosure” hypothesis. This effect is larger for Mexican immigrants than for other groups, but the opposite relationship is found for Chinese. Bueker (2006), however, using Current Population Survey data from around 2000, does not find significant effects of segregation on naturalization for blacks, Hispanics, or Asians.

Political Institutions and Naturalization

Bloemraad (2002) has described the importance of national policies in her comparison of naturalization rates for immigrants in the U.S. and Canada. She argues that both material and symbolic support for immigrants, as expressed in and stipulated by national policies, is responsible for the high levels of immigrant naturalization in Canada as compared to similar immigrant groups in the U.S. Bloemraad pays particular attention to how immigrants are politically mobilized through civic organizations in the ethnic and mainstream communities and how government policies can promote or hamper this local infrastructure.

What are the varying features of the American political system that promote or hinder positive feelings for immigrants in particular? To the extent that naturalization is an instrumental decision geared to position the individual with respect to state-provided social services, a restrictive policy environment may increase the incentive to become a citizen and also motivate immigrants to naturalize in order to protect, defend and assert their right to belong (Borjas 2002, Mazzolari 2009). One recent study (Van Hook et al. 2006) shows that the policy environment toward immigrants at the state level – including specific policies that can restrict public assistance to non-citizens as well as broad measures of public attitudes – influence naturalization. Others examine the impact of federal policies, such as the 1996 federal welfare reforms which, as initially implemented, restricted federal public assistance to immigrants who had not yet naturalized. Gilbertson and Singer’s (2003) ethnographic study of Dominicans in New York shows that these policies led to a growing sense of insecurity among legal residents and an effort to gain what the researchers call “protective citizenship.” On the other hand, a welcoming attitude toward immigrants on the part of local residents could also be an encouragement to citizenship as a step toward assimilation.

A related factor could be an immigrant group’s representation in the political process, through elected or appointed public officials, what Bobo and Gilliam (1990) refer to as political empowerment. A small number of ethnographic studies point out that the visibility of co-ethnic political leaders can spur naturalization by promoting a sense of belonging, political inclusion and empowerment. For example, Karphathakis (1999) argues that major obstacles to naturalization amongst Greek immigrants’ in New York City included their “lack of faith” in the American political system, their feeling of not “deserving” citizenship and the sense that they could not influence electoral politics. Co-ethnic elected officials may also play a direct role in promoting naturalization by linking with and providing economic and political resources to local community leaders who promote citizenship from the top down through mobilization drives (Gilbertson and Singer 2003).

Some authors have suggested that specific rules about voting and registration at the state level can be proxies for the overall openness and accessibility of political institutions, and in turn play an important role in shaping political participation (Jones-Correa 2001a; Ramakrishnan and Espenshade 2001). More liberal electoral rules could affect immigrants’ sense of whether the right to vote makes a difference, and in some contexts immigrant rights groups or other organizations may be the intermediary through which such opinions are molded.

One dimension of voting regulations that has recently drawn considerable attention is requirements for voter identification. Some scholars have argued that stricter voter identification requirements depress voting turnout in 2004 and that this effect is especially pronounced for minority voters (Eagleton Institute 2006). The issue of voter ID has been pressed particularly by the Bush Administration’s effort to draw attention to voter fraud involving non-citizens, undocumented aliens, or other without the right to register to vote, a political strategy that has been described as “vote suppression” (New York Times Editorial Board 6/14/07). If state voter ID requirements depress voter turnout, they may also discourage naturalization, a question addressed here for the first time.

The impacts of social and political context are not confined to immigrants’ experience in the U.S. Bueker (2006) suggests that immigrants coming from countries with repressive regimes are more likely to naturalize because the costs of return are higher for them. Yang (1994) finds that immigrants from socialist and refugee sending countries have higher propensities to naturalize. Woodrow-Lafield et al (2004) and Bueker (2006) show that immigrants from politically restrictive countries also are more likely to become citizens.

An additional institutional feature of an immigrant’s country of origin relevant to naturalization is legal recognition of dual citizenship, since it means immigrants would not forfeit political and economic rights as well as a sense of “cultural belonging” in their home country (Brettell 2006). Bloemraad (2002) believes that this variable is not likely to be significant, because official prohibitions to dual citizenship are often not enforced in practice. Others have pursued this question with mixed results (Yang 1994, Itzigsohn 2000, Jones-Correa 2001a). However a recent study by Mazzaroli (2009) examines the effects of changes in dual citizenship policy for the five countries (Colombia, the Dominican Republic, Ecuador, Costa Rica, and Brazil) that adopted dual citizenship in the 1990s). His study confirms that the adoption of dual citizenship in the 1990s by these countries positively affected the naturalization rates of immigrants from those countries.

RESEARCH DESIGN

The discussion above suggests the possible importance of macro-level or contextual factors such as community composition and political institutions in the country of origin and in the U.S. Few studies (such as Bueker 2006, Yang 1994, Woodrow-Lafield et al. 2004, and Liang 1994) have studied these questions for multiple ethnic/racial groups. Our approach is to pool together immigrants from all backgrounds in order to estimate a general model of naturalization that includes both individual and contextual factors. We then estimate specific models for each major group, which allows us to identify group-specific effects.

Data Source and Sample

The analysis relies on the 5 percent Public Use Microdata Sample (PUMS) of the 2000 U.S. census (IPUMS 2004). This dataset has the advantage of being both nationally representative and large enough for detailed analyses. The PUMS also makes it possible to define contextual variables at a finer geographic scale than the state or metropolitan region, the level of measurement in prior research. In the PUMS data, the place of residence is reported for Public Use Microdata Areas (PUMAs) containing approximately 100,000 persons.

The multivariate analysis is restricted to the immigrants who were 18 years old and above when they arrived in the U.S., the age at which they were required to apply for citizenship on their own behalf and unrelated to their parents’ applications. We also limit our sample to immigrants who would have met the naturalization residency requirement of having lived at least for five years in the U.S. by 2000, though this does not necessarily mean that they could have navigated the full legal process by this time. Hence all persons in the resulting sample were age 23 and older and had entered the U.S. by 1995.

Another sampling restriction is by household. PUMS data include all household members, creating potential problems of autocorrelation in estimating multivariate models. Our procedure is first to select all immigrants of a given racial/ethnic category and then randomly to select one person in that category from each household to include in a given analysis. It would be valuable to study patterns of naturalization and citizenship among household members. Very likely one’s options and choices are influenced by the status of spouses and children. However to do so in this study would press beyond the limits of our data. As noted below we have no information on when people naturalized, nor the duration of marriage, nor the status of spouses prior to marriage. To study interactions among household members’ citizenship status or even to control properly for autocorrelation within households would require longitudinal or retrospective data.

Despite the advantages of the census microdata, they have limitations for the study of naturalization. The main issue is it includes no information on the year in which a person became a citizen. In extreme cases, an elderly foreign born person in the microdata sample may have naturalized decades ago, and yet in this cross-sectional analysis we use as predictors their characteristics (and characteristics of their social context) in 2000. The result of subsequent residential mobility would be to add random error to our measures of contextual variables. This is a potential problem for contextual variables measured at the PUMA level. For other variables measured at the state level (electoral policies, state welfare policies, and public attitudes towards immigrants) the issue of residential mobility is likely to be less important, since more than 80% of population movements within a five-year period are within states.

We deal with this limitation in two ways. Following Fix, Passel and Sucher (2003), we restrict the sample to immigrants who have lived in the U.S. for less than 15 years. The number of years spent in legal permanent residence before naturalizing varies across immigrant groups (Rytina and Caldera 2008), but the average is 10–11 years. Because immigration increased sharply in the 1980s and 1990s, most foreign born persons in the sample are relatively recent immigrants, and we lose few cases by this restriction.

As a further check, we replicated the analysis reported here with a more restricted subsample of immigrants who did not move at all within the five years prior to data collection in 2000. For these persons we might expect a stronger correlation between their situation at the time of naturalization and current residential characteristics. In this restricted sample, all PUMA-level variables had the same effects as in the larger sample (except that the effect of co-ethnic political representation for blacks dropped below a significant level). Therefore we do not impose this restriction in the analyses presented below.

Another concern is that PUMS data don’t distinguish between legal and illegal immigrants and therefore includes undocumented migrants who are not eligible for citizenship.¹ Only through the use of administrative records would it be possible to identify legal immigrants and track their progress toward naturalization (as in the research by Woodrow-Lafield et al 2004). We address this limitation by including for the first time in a study of naturalization an occupation measure that has been found to be strongly associated with legal status (Passel and Clark 1998; see also Kooudji and Cobb-Clark 2000). Although undocumented immigrant workers can be found throughout the workforce, they tend to be over-represented in certain occupations. This control variable reduces but does not solve the problem of eligibility for citizenship.

Variables

Citizenship status (i.e., naturalized citizen or non-citizen) is the dependent variable in this study. The analysis utilizes four broad racial/ethnic categories, constructed from two different variables—race and Hispanic origin. The Census measures Hispanic identity and racial identity with separate questions. We combined these two variables to construct four racial categories that structure much of our analysis: non-Hispanic white, non-Hispanic black, Asian, and Hispanic. A respondent who self-reports within any of the Hispanic origin categories is treated as “Hispanic” regardless of race. Only those individuals reported to be white alone, but did not indicate any Hispanic identity, are considered to be non-Hispanic white. Those individuals reported as black (alone or in combination with any other race) but not Hispanic are considered to be non-Hispanic black. Asians are those reported as non-Hispanic Asian (alone or in combination with another race except for black).

Table 1 reports analyses of the five percent microdata sample from Census 2000 (weighted to yield full population counts) showing citizenship status of adults across racial/ethnic groups. The table reiterates what is already well known about the share of immigrants in the population. Among white and black adults well under 10% are foreign-born, while immigrants are a majority of Hispanics and more than three quarters of Asians. This finding means that naturalization has special importance for Hispanics and Asians. Less than half of the foreign-born population (42.9%) has naturalized. Non-Hispanic white immigrants are most likely to be citizens, followed by blacks and Asians. Only 30.1% of Hispanic adult immigrants are naturalized. One goal of this study is to determine whether these gross differences across groups remain when other individual and collective characteristics, including whether people are eligible for naturalization, are taken into account.

Table 1.

Nativity and Citizenship by Race/Ethnicity, 2000 (age 18 and above)

	Total	Native citizen	Naturalized Citizen	Non-citizen	% Citizen of foreign-born
Non-Hispanic White	150,488,985	144,134,950	3,628,743	2,725,292	57.1%
Non-Hispanic Black	23,934,416	21,991,077	920,830	1,022,509	47.4%
Hispanic	22,956,194	10,560,032	3,726,855	8,669,307	30.1%
Asian	8,751,867	2,016,716	3,497,736	3,237,415	51.9%
Other race	3,168,140	2,690,153	210,074	267,913	43.9%
Total	209,299,602	181,392,928	11,984,238	15,922,436	42.9%

Source: IPUMS (2000)

Individual-level predictors

Indicators of adaptation that are closely associated with assimilation theory include length of residence in the United States, and English speaking ability. Years in U.S. are represented in five categories in order to detect nonlinearity. Responses to this question are the person’s report of how long they have lived in the country, not when they became an “immigrant” in a legal sense. Age (treated as a categorical variable) is added as a control variable. English speaking ability indicates whether the respondent speaks only English at home and also how well (in four categories) the respondents who also speaks another language at home speaks English.

Two other measures represent the concept of rootedness in the United States (Portes and Curits 1987; Bueker 2006): marital status and number of children. Marriage is thought to increase stability and social networks and having children suggests a greater level of commitment to remaining in the United States (Bueker 2006). We also control for gender, since previous studies show that women may be more likely to naturalize than men.

Income, education and home ownership are included as indicators of socioeconomic status. Household income was reconstructed as a categorical variable with five categories in order to assess nonlinearity. Dummy variables were created to reflect categories of education. Occupation variables are included as a partial control for illegal status, since there is no direct measure of immigration status in the census. Following Passel and Clark (1998, Table M) we use the 1998 Standard Occupational Classification (SOC) in the PUMS file to identify 22 specific occupations in which illegal immigrants are highly represented (e.g., drywall/ceiling tile installer, grounds maintenance workers, food preparation workers, etc.) and whose proportion of illegal immigrants exceeds the proportion in the workforce (4.3 percent). High-level professional occupations (such as physicians, lawyers, and engineers) and protective service occupations (such as police and firefighters) that require licensing are classified as having zero probability of being illegal immigrants.

Community context

Some predictors are characteristics of the population based on the PUMA of residence. Many prior studies have measured demographic characteristics at the level of counties, metropolitan areas, or states. By virtue of its smaller size the PUMA is likely to reflect better the social environment of people’s daily lives (though PUMAs themselves are not small, averaging about 100,000 residents).

This study includes three community-level variables measured for PUMAs: percentage of adult naturalized citizens, the isolation index, and household income ratio. These are all group-specific. The percentage of foreign-born group members who are naturalized is calculated from the microdata. Because it is theoretically related to the group’s potential voting power, we limit the calculation to group members age 18 and over. The isolation index is a measure of the extent to which group members are exposed only to one another in the census tract where they live (based on Summary File 1, Census 2000). Hence it reflects both the relative size of the group in the PUMA and the degree of residential segregation from other groups across census tracts within the PUMA. Finally, income ratios (based on median household income) are calculated for each minority group (including immigrants and non-immigrants) relative to that of the non-Hispanic white population in the PUMA. The isolation index and income ratio are substantively meaningful only for the minority groups, and they are only included in those group-specific models.

Country of origin and state policy factors

Some collective variables are tied to the country of origin, which we code based on reported place of birth. Freedom House has developed rankings of countries in terms of civil liberty and political freedoms, combined to create an overall 3-point scale. We use dummy variables indicating whether the country is free, partly free, and not free. Another national-level political variable is provision for dual citizenship based on the reports of national policies provided by the U.S. Office of Personnel Management (2001).

We include two measures of the policy environment at the state level, applying indicators previously found to be associated with naturalization. One is an index of the safety net of welfare services to non-citizens developed by Zimmerman and Tumlin (1999) in a Working Paper of The Urban Institute (used also in Van Hook et al. 2006). This measure reflects immigrant access to benefits in 12 separate social policy areas, including post-1996 access to TANF, Medicaid and Food Stamps. We collapse the original measure into two categories, the least restrictive states (where the safety net is “most” and “somewhat” available) vs. the most restrictive (“less” and “least” available). Another indicator of the state policy environment is “immigrant receptivity,” a measure of public attitudes originally developed for metropolitan areas by DeJong and Tran (2001) and DeJong and Steinmetz (2004) based on data from the General Social Survey in the years 1995 to 1997. This measure was expanded to the state level and converted to standardized scores by Van Hook et al (2006).

This study is the first we are aware of that examines the relationship between minority co-ethnic political empowerment, measured by black, Latino, and Asian co-ethnic office holding, and naturalization. A measure of Latino co-ethnic office holding is available from a directory prepared by the National Association of Latino Elected and Appointed Officials (NALEO 2000), including elected and appointed public officials at all levels and listing their official postal address. Corresponding data for black elected officials in 2000 were provided by the Joint Center for Political and Economic Studies. Data on Asian elected and appointed officials (2001–2002) were obtained from the UCLA Asian American Studies Center. We used a standard source to link zip codes to PUMAs (MABLE Geocorr 2000). PUMA level information for each ethnic/racial group was then incorporated into race-specific sub-samples. We report here the effect of a simple dichotomy distinguishing PUMAs with no co-ethnic office holder from those with at least a share of one office holder. In exploratory analyses we experimented with other ways of operationalizing this variable.

Electoral policies are measured at the state level because most voting requirements are regulated by state governments (see Yang 2002; Jones-Correa 2001a; Ramakrishnan and Espenshade 2001). Indicators of states’ early voting and liberalized absentee voting policy are drawn from Hansen (2001). Another variable not used before as a predictor of naturalization is voter identification requirements. Because of its novelty, we describe it in more detail here.

The question is whether or not a respondent lives in a state requiring prospective voters to show some form of personal identification before casting a ballot. We consider this to be an indicator of overall political openness that may directly or indirectly shape immigrant propensity to naturalize. Forms of identification required or requested may include photo or non-photo ID. The Election Reform Information Project (2006) classified state requirements as of 2000. In 2000 eleven states required voter ID, though there was variation in the form of ID and in policies that governed what a poll worker should do if ID was missing. We collapsed the original classification into a simple dichotomy based only on the maximum requirement: does the state require documentary evidence at the polls of the prospective voter’s identification?

RESULTS

Multivariate logistic regression is employed to analyze the effects of the explanatory variables on the probability of citizenship, a dichotomous variable. We estimated a single pooled model of all racial/ethnic groups, plus separate models for each racial group/ethnic. Table 2 reports the odds ratios for the effects of the collective variables in these models, not the full models.²

Table 2.

Effects of collective variables on naturalization (odds ratios)

		All races		Hispanic		Asian		Black		White
Predictor	Reference category
Race	Non-Hisp white (ref)
	Hispanic	1.328
	Asian	1.310
	Non-Hispanic black	1.545
Repressiveness	Free (ref)
	Partly free	1.256		0.96		0.99	##	0.75		2.87
	Not free	1.856		1.93		1.72		1.08		3.09
	Missing data	1.189		1.72		1.22		0.81		1.83
Dual Citizenship	Not allowed (ref)	0.894		1.31		1.07		0.88		0.95
Co-ethnic
naturalization		1.025		1.02		1.03		1.03		1.02
Isolation Index		NA		0.9980		1.0028		0.9989		0.9976
Income ratio		NA		1.17		0.76		0.90		NA
Safety Net	Less available (ref)	0.943		0.98		0.96		0.78		1.02
Receptivity		1.021		1.03		1.03		1.03		1.01
Co-ethnic
representative	None (ref)	NA		1.01	##	0.94		1.01	##	NA
Voter ID policy	No requirement (ref)	0.901		0.94		0.98		0.96		0.90
Absentee policy	Restricted (ref)	1.004	##	0.90		1.05		1.05		1.09
Early vote policy	No early vote (ref)	1.108		1.29		0.97		1.06		0.91
Constant		0.014		0.02		0.01		0.01		0.02

Note: All individual-level variables are controlled in these models.

^##indicates coefficients that are not statistically significant.

One surprise in the pooled model is that, after controlling for other variables, every minority group is shown to have higher odds of naturalization than do non-Hispanic whites by a factor of 31–55%. Differences between non-Hispanic white and blacks or Asians that were reported in Table 1 were due mainly to differences in other individual-level variables. The low overall rate of naturalization of Hispanics is explained mainly by contextual effects: the relatively low percentage of naturalized co-ethnics in predominantly Hispanic communities. Outside of those contexts, Hispanics are more likely than non-Hispanic whites to naturalize. The revised race/ethnic hierarchy revealed here is consistent with Yang’s “forced self-protection” hypothesis – that seeking citizenship is a means of empowerment for immigrants who find themselves in a minority status in the U.S. but not for white immigrants.

Individual-level predictors

Individual-level predictors have strong effects consistent with those found in prior studies, and we briefly summarize them here. There are several cases where effects differ across groups in ways that have not been noted in prior research.

Age

Older persons are more likely to be naturalized. The effect is in the same direction for all groups but smaller for blacks and Asians.

Gender and Family

For Hispanics and Asians, women’s odds of naturalization are higher than men’s by a factor of 4% (Hispanics) and 13% (Asians). For whites women are less likely to become citizens. For blacks, gender has no significant effect on naturalization. Hispanic and Asian married persons are more likely to naturalize; for whites and blacks, they are slightly less likely. For all racial groups, there is a modest increase in likelihood of naturalization for each co-resident under-18 child in the household.

Years in U.S

For members of all racial/ethnic groups (and especially for Asians), each increase in the number of years of residence in the U.S. is associated with higher rates of naturalization. The differences are strong even between age categories in which all immigrants are likely to have met the time requirements to naturalize.

Language

Among blacks and Asians there is little difference in attaining citizenship between those who speak only English at home and those who speak English well and very well. Significant gains, however, are shown at lower levels of language ability in all groups: compared to those who do not speak English at all, those who speak it not well have twice or more the odds of naturalizing, and the odds for those who speak better are even greater.

Education and Income

Members of all groups show gains in naturalization rates with higher educational achievement. However, those with graduate education are surprisingly less likely to naturalize than those with less education. One possibility worth further exploration is that these people have strong career opportunities even without citizenship, or that their options are more global and independent of U.S. citizenship. There is also a curvilinear effect of income for all groups except blacks: those in the middle income categories are most likely to naturalize. Among blacks there is a positive monotonic effect of income.

Home Ownership

Home ownership is significantly associated with naturalization.

Occupation

Occupational categories have been constructed as a proxy for likelihood of being undocumented immigrants. For all groups, immigrants who are employed in occupations with high odds of being undocumented are less likely to naturalize. However there is evidence that many non-Hispanic white immigrants with high professional standing choose not to become citizens (parallel to our finding for Asians with the highest level of education).

Collective predictors

All of the coefficients for collective characteristics in Table 2 are statistically significant, except those marked with a double ##. Some of these variables have been included in prior studies, but this is the first analysis to examine the effects of restrictiveness of the electoral system (reflected in state-level voter ID policies) or co-ethnic political representation (at the metropolitan level).

The results are puzzling in two ways. First, although several collective variables have significant effects as hypothesized in the all-races model, there are two cases where significant effects are in an unexpected direction. Second, several variables have different effects in one group-specific model than in others.

Two collective variables refer to the immigrants’ country of origin: the character of its political system and whether it allows dual citizenship for emigrants. One motive for attaining U.S. citizenship for those who experienced repressive regimes is protection against the government of the country of origin. In the pooled sample the odds to naturalize are 86% greater if one is from a repressive (not free) country, and we find significant effects for all groups when we compare the most repressive category of regimes vs. non-repressive regimes. It is a very strong effect for all but blacks, and it is strongest for white immigrants.

We expected dual citizenship to promote becoming a U.S. citizen for all groups, since doing so does not require losing one’s original identity. But past studies have had mixed results, and as Yang (1994) pointed out, immigrants might perceive dual citizenship as an added burden or responsibility rather than a benefit. In fact availability of dual citizenship reduces the odds of naturalization by about 11% in the pooled sample. This finding is repeated in the separate models for blacks and whites. However, Hispanics from countries that allow dual citizenship have odds of naturalizing that are 31% higher than the odds for those from countries without this possibility. (This result is heavily influenced by the weight of Mexicans among Hispanic immigrants; Mexico allows dual nationality.) For Asians from countries with the recognition of dual citizenship, the odds of naturalizing are 7% higher than the odds for those from countries that do not allow dual citizenship.

It is puzzling why dual citizenship would have opposite effects for Hispanics and Asians, on one hand, and whites and blacks on the other. For the Asian case, it may be relevant that most Asian countries do not recognize dual citizenship. In our sample only 3.4% of Asians reported that their birthplaces were countries which recognize dual citizenship, and most of these are immigrants from Bangladesh (40.5 % of these cases) or from various Western countries. In other words, the positive effect for Asians may be unique to certain countries of origin, not generally representative of the Asian population in the U.S. On the other hand, the effects for Hispanics are strongly affected by the cases of Mexico and the Dominican Republic, both of which instituted dual citizenship relatively recently. Both Jones-Correa (2001b) and Mazzolari (2009) show that effects of dual citizenship on naturalization are strongest for recent adopters. Possibly, then, what we have found is not a general “Hispanic” effect but an effect of recent adoption that might be found for immigrants from non-Hispanic countries.

Two other variables refer to demographic characteristics of the metropolitan area where immigrants live. There is a strong contextual effect of the share of other immigrants of the same racial/ethnic background who have naturalized. Net of one’s own characteristics, clearly there is an additional pull toward naturalization if many co-ethnic immigrants have done so. In our models, for every percent increase in the share of coethnic immigrants who are naturalized, the respondent’s odds of naturalizing increases by 2.5% overall, and similar effects are found for every group. Ethnic isolation (that is, living in an area with a higher probability that the respondent has many coethnic neighbors and less exposure to other groups) has a significant negative effect for Hispanics, blacks, and whites. But the effect is positive for Asians. The direction of effects is consistent with Liang’s (1994) findings, including the Asian exception (represented by Chinese immigrants in Liang’s work). However the size of the effect is quite small.

Another metropolitan characteristic investigated here is the degree to which immigrant group members’ incomes are on par with those of a standard reference category (we use the incomes of U.S. born non-Hispanic whites for comparison). Where the group is relatively more successful, we would expect group members to be more likely to naturalize. This is the result for Hispanics, but the opposite is found for black and Asian immigrants. (This variable is not included in the equations for white immigrants.) Hence this is another example of contradictory results.

Finally we consider several aspects of the policy or political context. The models for all groups include two measures of the policy environment at the state level. The Hispanic, Asian and black models also include dummy variables for the presence of co-ethnic office holders at the finer geographic level of the PUMA.

In the pooled sample, immigrants in states with less restrictiveness of services to non-citizens are less likely to naturalize, an effect that is statistically significant but small – decreasing the odds of naturalization by 6%. This suggests that there may be an instrumental motive for citizenship that is not as salient in less restrictive settings. The group specific models show that this effect is actually much stronger for blacks (for whom the odds are decreased by 22%) while the effect is small but significant for Hispanics and Asians. The effect is small but positive for whites. In contrast, while a less restrictive policy with respect to services may weaken an inducement to citizenship, a welcoming attitude on the part of the public has a small positive impact for all groups. One standard deviation increase in the receptivity index (a large difference in the absolute value of the scale) increases the odds of naturalizing by about 3 % for Hispanics, Asians, and blacks. This effect is smaller for whites (1.4 %).

Presence of co-ethnic office holders for immigrant groups at the PUMA level has a positive but not significant impact on naturalization for Hispanics and blacks. Asians are exceptional; the presence of co-ethnic office holders for Asians has a significant negative effect on naturalization. We had assumed that co-ethnic office holders would have an interest in encouraging naturalization (which might increase their natural electoral constituency). We also anticipated that communities with higher rates of naturalization would be more likely to be able to elect or secure the appointment of a co-ethnic public official. Both processes would lead to a significant positive effect. The results lead us to speculate that co-ethnic office holders might affect naturalization through a different route than we anticipated: they may make it easier for group member to demand and receive public services even without becoming citizens. Then office holders would reduce the instrumental motive for naturalization, just as do less restrictive social service policies. But because this is the first time that a measure of political representation has been used in a study of naturalization, we feel it is premature to draw a strong conclusion.

We also include three institutional variables to represent whether state laws encouraging electoral participation affect naturalization. In the pooled sample, early voting increases naturalization but there is no effect of easier absentee voting. These findings are partly replicated in the group-specific models. Unexpectedly early voting has a negative effect for Asians and non-Hispanic whites. Easier absentee voting increases naturalization for Asians, blacks and whites, but has a negative effect for Hispanics. Finally, in light of much public debate about the effects of voter ID requirements, it is important to note that these requirements tend to depress the odds of naturalization by nearly 10% in the pooled model. Voter ID requirements reduce whites’ odds of becoming a citizen by about 10%, with smaller but significant effects for the other three groups.

DISCUSSION

This study partly verifies results from previous research about the effects of individual-level predictors of attaining citizenship. Age, years in U.S., English speaking ability, and education have especially strong impacts on the likelihood of becoming a naturalized U.S. citizen, mostly consistent with the expectations of assimilation theory. We have nevertheless found unexpected differences across groups in the size and direction of these effects.

Our principal interest in these individual-level variables is to learn to what extent compositional differences among racial/ethnic groups can account for differences in their rates of naturalization. The result is startling. Non-Hispanic white immigrants are much more likely than the average immigrant to be citizens (over 50%), and Hispanics are by far the least likely (barely above 25%). Yet after controlling for individual background characteristics, Asian and black immigrants have much higher odds of naturalization than do whites, and the addition of further controls (especially the naturalization rates of other coethnics in the same urban area) leaves whites as the least likely to naturalize. This inversion of the naturalization hierarchy offers support for the theory that seeking citizenship not only reflects people’s incorporation into American society as individuals, a process of assimilation, but also is a collective behavior associated with minority status. It is consistent with Yang’s hypothesis that immigrants view citizenship as a resource for self-protection.

We also find significant effects of political and community context. Additional research and theoretical work will be needed to understand the nature of these effects. Our analysis cannot identify the mechanisms behind them, and we cannot explain why some contextual characteristics count and others seem not to. We focus here on those that do make a difference. We found a strong place effect: where a higher share of group members in the local area have naturalized, persons in our sample are also more likely to have become citizens. Evidently something about the context not otherwise measured in our models has an impact on choices made by group members. Yang (1994), cited above, reports a similar finding at the state level, and refers to it as the “effect of the immigrants’ ethnic communities.” Discovering the specific nature of this effect deserves attention in future studies. Like other collective variables it could be interpreted within a network perspective.

The character of the government in the country of origin has consistent effects. All else equal, immigrants from politically repressive countries are much more likely to seek citizenship in the U.S. This could be because these immigrants are more likely to consider the U.S. as a permanent home or because they are more strongly motivated to take advantage of political freedom including the right to vote. More specifically related to voting is the effect of state-level voter identification requirements. Because citizenship is just one step, and an early step, toward political participation, it may seem unlikely that immigrants considering citizenship are looking ahead to the conditions under which they could take part in elections. It would be surprising if many immigrants even knew what the identification requirements are. And yet for every group, not only for minorities, this institutional variable has a significant depressive impact. We do not know enough about electoral procedures to be sure how to interpret the effect. Possibly voter identification rules are a proxy for rates of voter registration or voting, or for the strength of immigrant organizations in state politics, or for another similar phenomenon that could be salient in immigrants’ lives.

Other aspects of the community context may reflect the role of citizenship as a protective response to a hostile environment. Where there are fewer statewide restrictions on public services to immigrants, Hispanic, black, and Asian immigrants are less likely to naturalize. In this case, the underlying explanation could be that in these less hostile settings immigrants feel less need to reinforce their position by becoming citizens. Note that this analysis is based on data from 2000. Restrictive measures regarding access to welfare and other services that were taken in the late 1990s at the federal and state levels (e.g., the 1996 Welfare Reform Act) would not have had much impact on naturalization at this time. This finding suggests that more recent data would reveal a trend toward higher naturalization propensities in response to this legislation. On the other hand, a more welcoming environment as indicated by public attitudes towards immigrants promotes naturalization by all immigrant groups. Also, where there are co-ethnic office-holders, immigrants of Asians are less likely to naturalize while the presence of co-ethnic representatives has no significant effect on naturalization for Hispanic and black immigrants. Therefore we cannot make a blanket statement about this dimension of welcome vs. hostility.

Some of the limitations of this study arise from its cross-sectional nature, so that any causal inferences must be tentative. Other limitations are imposed by our reliance on census data for individual-level variables. We would prefer to have more direct measures of how immigrants experience their community setting. This would be especially helpful in interpreting the negative effect of ethnic isolation for blacks, Hispanics, and whites, but the unexpected positive effect for Asians. Black, Hispanic, and white immigrants all tend to live in PUMAs where they are a majority of the population, so high levels of isolation for these groups imply near-homogeneity of co-residents (the mean levels are 75% white for white immigrants, 46% black for blacks, and 45% Hispanic for Hispanics). Asians are present in much smaller numbers in most places (averaging 18% Asian in the PUMAs where they live), and the average Asian immigrant lives in a majority-white area. We might interpret the black and Hispanic results in the way that Liang and others do – where immigrants live in relative separation from members of other racial/ethnic groups, they may be less aware of discrimination and feel less need for the protection of citizenship. But does this interpretation apply to white immigrants in a majority-white society? And why is there the opposite effect for Asians?

Although questions remain about how to interpret these results, the broader point is that acquisition of citizenship is not simply a matter of an individual’s assimilation into a new homeland. Our findings reinforce a growing recognition that naturalization also has a collective character, that there are processes that systematically and differentially affect immigrants from different racial, ethnic, and national origins, and that citizenship needs to be understood within the political and community context of immigrants’ life in this country.