1. Introduction
Citizenship is a status and a process (Bueker, Reference Bueker2009). As a status, it establishes an administrative boundary between members and non-members. At the same time, citizenship also entails a process of becoming a citizen through socialization via parents, schooling, and community networks. In a democracy, being a citizen signals official membership in a national community and is wrapped up with notions of active civic engagement and voting rights. It is also often linked to the process of instilling the duties and responsibilities to act and engage in politics, along with imbuing a sense of membership.
Given this, how does citizenship affect political participation in the United States? How can scholars learn about the relative power of citizenship as a status and a process? There are several challenges in addressing this question—particularly a selection problem, which arises from the difficulty of isolating the effect of legal status from socialization when comparing citizenship categories (Hainmueller et al., Reference Hainmueller, Hangartner and Pietrantuono2017). In this paper, we use dual citizenship to help provide leverage to address this question (and also provide the first look at dual citizens in the United States using two large, nationally representative samples). We juxtapose two pairs of citizenship categories to highlight both their similarities and differences. The first pair, born-dual and naturalized-dual American citizens, shares a similar legal status but differs in birthplace and socialization. The second pair, born-dual and born-mono American citizens, has comparable birthplaces and socialization, yet differs in the presence of an additional legal status in another country.Footnote 1 Through these comparisons, along with additional efforts to ensure comparability by matching on demographic and political variables, we aim to disentangle the effects of nativity and an additional legal status.Footnote 2
We find that citizenship’s impact on political participation works via both legal status and process. Among dual citizens, those born in the US tend to participate more in politics than immigrants who naturalized (with the exception of voting). Among US-born citizens, dual citizens tend to engage more across a range of political activities than mono citizens. However, US-born mono citizens voted in the 2020 and 2022 elections more than US-born dual citizens. The findings imply that even in our globalized world and with ongoing debates about postnational citizenship, early socialization and citizenship status are still determining factors for civic engagement in the United States.
2. Citizenship and political participation
Most political activities in the US do not require citizenship. Legal permanent residents, for example, can donate money, and anyone can protest, wear a button, or contact political representatives. However, citizenship establishes official full membership and, in particular, determines whether someone can vote in federal and state elections, which explains why GOTV organizations often attend naturalization ceremonies to register new citizens.
Citizenship as a process also influences political participation. Citizens born in the US benefit from longer socialization processes on average, typically gaining language fluency, cultural familiarity and repeated exposure to democratic norms in public schooling. They are more frequently taught what it means to be “good Americans”, reinforcing the idea that political engagement is both a duty and a right of democratic citizens.Footnote 3 This sustained exposure plays a crucial role in shaping motivation, mobilization, and participation (Verba and Nie, Reference Verba and Nie1972; Brady et al., Reference Brady, Verba and Schlozman1995; Jones-Correa, Reference Jones-Correa1998; Cho, Reference Cho1999).
That said, categorical distinctions based on nativity may be blurred, especially if citizens have immigrant parents or families with mixed legal statuses (Amuedo-Dorantes and Lopez, Reference Amuedo-Dorantes and Lopez2017; Menjívar et al., Reference Menjívar, Simmons, Alvord and Valdez2018; Abrego, Reference Abrego2019; Asad, Reference Asad2020). And, because civics and American history are covered in the naturalization exam, the recency effect may outweigh that of primacy or repetition. In that case, we might expect fewer differences in political participation as a result of nativity.
Nativity is also linked to a sense of American identity, however. Given jus soli, being born in the US is often considered synonymous with being American (Wong, Reference Wong2010; Bloemraad, Reference Bloemraad2013). Bloemraad (Reference Bloemraad2013) finds that birthright citizenship appears to provide—more than naturalization—a sense of legitimate belonging to the US. Even though immigrants made a conscious choice and were accepted as new members of society (Coutin, Reference Coutin2003; Hainmueller et al., Reference Hainmueller, Hangartner and Pietrantuono2015; Reference Hainmueller, Hangartner and Pietrantuono2017), it may not generate as great of a sense of belonging and commitment to engage fully in politics as that of someone born an American. Nevertheless, there is also evidence that birthplace may have no effect on patriotism (De la Garza et al., Reference De la Garza, Falcon and Garcia1996; Huddy and Khatib, Reference Huddy and Khatib2007).
Empirically demonstrating how aspects of citizenship influence political engagement is challenging. One could compare naturalized citizens and non-citizen immigrants—both groups were not born in the United States yet they differ in their citizenship status. However, as Hainmueller et al. Reference Hainmueller, Hangartner and Pietrantuono(2017) explain, there is a double selection problem in making this comparison: some immigrants apply to naturalize and others do not, and some succeed in becoming citizens and others fail in their quest. Non-citizens may have been unsuccessful in previous attempts to naturalize, not yet met the residency requirements, or never applied because of financial or language barriers, for example. We cannot isolate the effect of citizenship status on participation, and furthermore, we would not be able to compare all types of political participation, since non-citizens cannot vote in federal elections (and most local ones as well).
Comparing native-born and naturalized citizens might suggest whether birthplace and time in the U.S. influence participation (Pantoja et al., Reference Pantoja, Ramirez and Segura2001). Still, there remains some ambiguity because these groups differ significantly in their exposure to another country. While naturalized citizens are generally more likely to have some form of cultural, social, or psychological connection to their country of origin, native-born citizens tend to show greater variability in their ties to other countries. This asymmetry makes the comparison less precise and complicates efforts to isolate the effects of nativity and socialization in the U.S. It would be ideal to compare naturalized citizens with native-born citizens who are more likely—personally or via their families—to have some extra-US connection reflected in their legal status. To navigate these challenges, we seek pairs of citizenship categories that vary in one dimension while holding others constant.
3. Citizenship categories in the United States and data
Birthplace (jus soli), parentage (jus sanguinis), and naturalization can combine to yield diverse typologies of citizenship status recorded in the US Census. However, there is another type of citizenship status that is invisible in federal records: dual citizenship. Due to the combination of country-level and individual-level factors (i.e., global acceptance of dual citizenship and the increasing number of migrants), this category is rapidly increasing (Vink et al., Reference Vink, De Groot and Luk2015; Harpaz, Reference Harpaz2019). Incorporating dual citizens, we present citizenship categories in the United States in Table 1.
Citizenship categories in the United States
There are three types of dual citizens: born-dual citizens whose birthplace is in the US (group D), born-dual citizens whose birthplace is outside the US (group E, who have at least one American citizen parent), and naturalized-dual citizens who immigrated and naturalized as American, while retaining their original citizenship (group F).Footnote 4 Oftentimes, the discourse on dual citizenship is limited to immigrants, but dual citizens in the United States are, in fact, not limited to immigrants. Focusing on the immigrant population exclusively misses the first group (Group D) entirely, and there is no evidence that this group is small enough to be ignored. If studies of dual citizens only sample from immigrants, the analyses and conclusions drawn from them could be misleading.
We analyze data from the 2020 and 2022 CES (Cooperative Election Study), which use nationally stratified samples conducted over the internet by YouGov. The Common Content was asked of 61,000 (in 2020) and 60,000 (in 2022) adults interviewed in October (pre-election wave) and November (post-election wave). To our knowledge, these were the first times large, nationally representative surveys asked about the dual citizenship status of American respondents.
There is a surprising dearth of knowledge about dual citizens in the U.S. The U.S. government does not collect this information, and surveys that have asked about dual citizenship so far have been either too small, non-representative, or focused on a single racial/ethnic group, making it difficult to draw generalizations about the country as a whole (Escobar, Reference Escobar2004; Cain and Doherty, Reference Cain and Doherty2006; Staton et al., Reference Staton, Jackson and Canache2007; Jung and Ksiazkiewicz, Reference Jung and Ksiazkiewicz2022). In other words, there is no “gold standard” or baseline for comparison. These data provide the first comprehensive look at who, how many, and from what range of countries are American dual citizens.
4. The effect of nativity/socialization and the effect of having two citizenship
Dual citizenship provides a unique opportunity to examine political participation. Our primary reference group is born-dual citizens (Group D) because they share similarities with both naturalized-dual (Group F) and born-mono citizens (Group A).
First, we observe whether the method of acquiring U.S. citizenship influences participation by comparing born-dual and naturalized-dual citizens. While both groups hold legal ties to two countries, we expect that nativity would provide an advantage in political engagement via a potentially greater sense of American identity and longer socialization. Furthermore, while both groups may experience discrimination (Kwon 2022; Schut and Boen 2022), naturalized-dual citizens may encounter additional nativism that dampens mobilization and civic engagement, and their language skills—while sufficient for naturalization—may not always guarantee comfort with the level required for broader political engagement. Therefore, we hypothesize that born-dual citizens participate in U.S. politics more than naturalized-dual citizens (H1).
Next, we examine differences in the number of legal statuses, while their birthplace and socialization are comparable. Previous findings tend to point to lower levels of participation among dual citizens relative to their mono citizen counterparts, but comparisons vary: Cain and Doherty Reference Cain and Doherty(2006) and Staton et al. (Reference Staton, Jackson and Canache2007) compare naturalized-mono with naturalized-dual citizens, focusing only on immigrants; Jung and Ksiazkiewicz (Reference Jung and Ksiazkiewicz2022) compare mono and dual citizens without specifying birthplace.
Dual citizenship could hypothetically double or halve political activity. On the one hand, it expands both opportunities for engagement and constraints in terms of physical and psychological resources, as individuals navigate multiple political spheres. If politics is Dahl’s “sideshow in the circus of life” (Dahl, Reference Dahl2005), individuals are capable of watching multiple sideshows, and one spectacle can spark interest in another. On the other hand, people cannot reside in two countries simultaneously and are limited in terms of time and energy. Increasing cognitive demands may raise participation costs, potentially reducing participation.
It is possible that claiming dual citizenship is already a sign of greater political interest on the part of dual citizens, relative to individuals who only claim one. However, this would be relevant primarily for immigrants who consciously make the choice both to become American and retain their original citizenship. If so, we would expect naturalized-dual citizens to be more politically engaged than born-dual citizens, making our comparison a harder test. We do not believe that dual citizenship is, in and of itself, an indication of political interest, but this is an empirical question we can answer.
Nonetheless, we expect that the difference between mono-citizens and dual-citizens will be negligible once we make a more focused comparison within the US-born population. We hypothesize that having another citizenship would not hinder engagement in US politics. Since both groups are US-born, we expect they would share similar levels of socialization, exposure to American civic duties, and identity from their birthright citizenship.Footnote 5 Furthermore, we do not anticipate that an additional legal status will affect their political interest or participation, at least within the country of residence, the US.Footnote 6 This aligns with the “postnational” perspective (Spiro, Reference Spiro1997; Reference Spiro1999; Soysal, Reference Soysal2001; Schlenker, Reference Schlenker2016; Schlenker et al., Reference Schlenker, Blatter and Birka2017), which argues dual nationality does not lead to shallower civic engagement. Accordingly, we hypothesize that political participation levels are not different between born-dual citizens and born-mono citizens (H2).
5. Who are dual citizens in the US
5.1 Citizenship Status and Immigrant Generation
The CES asks whether the respondent is a citizen of the US, and whether she was born in the US or a naturalized citizen. If the respondent is a US citizen, a follow-up question asks, “Are you also a citizen of another country besides the United States?” If yes, they specify the country. Using these variables, we code citizens into four categories: (1) Born-Mono Citizens, (2) Born-Dual Citizens, (3) Naturalized-Mono Citizens, and (4) Naturalized-Dual Citizens.
5.2 How Many Dual Citizens Are There in the US?
In 2020, 2,452 respondents (4.0%) stated that they are a citizen of another country in addition to the US, rising to 2,617 respondents (4.4%) in 2022.Footnote 7 This proportion is higher than in Germany (3.5%) but lower than in the Netherlands (7.7%), Canada (11.2%), and Switzerland (19%) (Germany: Federal Statistics Office 2022; Statistics Netherlands 2014; Statistics Canada 2021; Switzerland: Federal Statistical Office 2019). Given that our estimates translate to approximately 15 million Americans, we can cautiously predict that the US has the largest number of dual citizens in the world, considering that the two more populous countries do not allow dual citizenship.
5.3 Nativity
In 2020, the number of US-born dual citizens (
$N = 1,216$) was similar to foreign-born dual citizens (
$N = 1,236$). In 2022, US-born dual citizens (
$N = 1,404$) slightly outnumbered foreign-born dual citizens (
$N = 1,213$). While dual citizenship has been discussed primarily within the context of immigration, this ignores about half of the nation’s dual citizens. In other words, an estimated 2.2% (2020) and 2.5% (2022) of all US-born citizens residing in the US have another citizenship that they acquired on the basis of ancestral lineage, marriage, or naturalization while maintaining their American citizenship.
5.4 Country of Another Citizenship
The list of countries for the other citizenship is long—160 (2020) and 153 (2022)—highlighting the global nature of dual citizenship (Table D1). Among them, 66% (2020) and 60% (2022) were democratic, resulting in most dual citizens (82% in 2020 and 83% in 2022) having ties with democratic nations. Canada, the United Kingdom, and Mexico were the most common countries of another citizenship, with Americans holding citizenship in these countries comprising 33% of dual citizens in 2020 and 31% in 2022.
5.5 Demographics and Socioeconomic Indicators
Table 2 summarizes key demographic, socio-economic, and political indicators across citizenship categories.Footnote 8 Notable differences in age, education, income, and political interest are observed.
Descriptive characteristics of U.S. citizenship categories 2020 and 2022
*indicates median.
6. Dual citizenship and political participation
6.1. Measures of political participation
We examine (1) voting and (2) political participation by creating an additive index from seven indicators. The participation index includes the following 7 variables: Run for elective office at any level of government (local, state or federal); Attend local political meetings (such as school board or city council); Put up a political sign (such as a lawn sign or bumper sticker); Work for a candidate or campaign; Attend a political protest, march or demonstration; Contact a public official; and Donate money to a candidate, campaign, or political organization.
6.2. Research design
In addition to creating comparable citizenship groups, we use a matching-based approach to minimize the influence of demographic characteristics and political indicators that could confound the effect of nativity and additional citizenship on political participation. The first model (Model 1) includes demographic characteristics as covariates: age, sex, race/ethnicity, education, income, area of residence (urban or rural), marital status, and employment. Immigrant generation is included only in the second comparison (between born-mono and born-dual citizens). The second model (Model 2) incorporates both demographic and political variables, including political knowledge, political interest, partisanship and strength of partisanship (see Appendix 6 for more details and Model 2 results).
To identify the closest matches, we used Mahalanobis distance matching with calipers. The group sizes were relatively comparable in the comparison between dual citizens, but not in the second comparison due to the larger number of mono-citizens. To address this imbalance, we applied stricter matching criteria, combining exact matching with calipers, and we achieved strong balance (see Appendices 5 and 6).
6.3. Nativity vs. naturalization (born-dual vs naturalized-dual citizens)
We find supporting evidence for H1. Born-dual citizens score higher on the political participation index than naturalized-dual citizens, though voter turnout shows no significant difference. Table 3 presents results from Model 1, which remain consistent in Model 2 (Table G1). Analyzing seven political activities separately, we observe slightly more variation in 2022, but at no point do naturalized-duals exhibit higher political participation than born-dual citizens (Figure H2). On average, born-dual citizens experience longer socialization in civic duties, norms, and participation logistics than naturalized-dual citizens of comparable age.Footnote 9 They may also have a greater sense of American identity deriving from their birthright citizenship and years of socialization in U.S. schools. They are also less likely to experience discrimination for being perceived as “un-American.”
Comparison of political behavior among dual citizens
Note: Robust standard errors in parentheses. *
$p \lt 0.05$, **
$p \lt 0.01$, ***
$p \lt 0.001$.
The lack of difference in voting may be a result of the fact that voter registration is often enabled by GOTV organizations at naturalization ceremonies, such that new Americans are primed to think about voting as a salient part of their new citizenship. Any additional boost from longer socialization for born citizens may have been matched by the increased salience of the steps required to vote for naturalized citizens. Furthermore, ballots are available in multiple languages, unlike other political activities that lack similar language support.Footnote 10 As a result, naturalized-dual citizens comfortable enough to vote may still avoid activities like campaigning or letter-writing. Finally, voting involves (and is increased by) a great deal of institutional (party) support, whereas other political acts require more individual expenditures and effort and may be more depressed by perceptions of bias (Brady et al., Reference Brady, Verba and Schlozman1995; DeSipio, Reference DeSipio2011).
6.4. One vs two citizenship (born-mono and born-dual citizens)
While we expected no difference in the level of political participation between born-mono and born-dual citizens (H2), the results in Table 4 reveal some differences. On average, born-mono citizens were more likely to turn out to vote than born-dual citizens. However, born-dual citizens scored higher on the political participation index compared to born-mono citizens.Footnote 11 When analyzing the seven political activities separately, we observe some variation (Figure H3 and H4). In both years, born-dual citizens were more involved in running for office and working for a candidate or campaign. This pattern suggests that more born-duals are mobilized for elections, perhaps due to networks that facilitate recruitment (Carlson et al., Reference Carlson, Abrajano and Bedolla2020).
Comparison of political behavior among American citizens born in the U.S. (Model 1)
Note: Robust standard errors in parentheses. *
$p \lt 0.05$, **
$p \lt 0.01$, ***
$p \lt 0.001$.
The higher political engagement of born-dual citizens compared to born-mono citizens, except in voting, is surprising. How can we make sense of this? We explored a few possibilities: vote intention, voter registration, and verified vote. In 2020, both groups had similar levels of vote intention and voter registration (Figure I2). By 2022, voter registration remained similar, but born-dual citizens showed slightly greater voting intent than born-mono citizens. Verified vote perhaps may get around the social desirability issue, but the results were similar (Table J1): born-duals were less likely to vote than born-mono citizens.
We explored contextual factors that might affect the link between behavioral intention and actual behavior, specifically those that could deter born-dual citizens from voting despite their initial intention. First, because voting laws differ by state, we took a closer look at the top 4 states where dual citizens reside: California, Florida, New York, and Texas. When we make matches within the same state, the results remain consistent, albeit with reduced statistical power; differences across states cannot explain why born-duals are less likely to vote (Appendix 10). Second, home-ownership and long-term residency in current homes may have generated a greater or lesser stake in the outcome of elections, but incorporating these variables did not alter the pattern (Table L1). More research is needed to understand why holding an additional citizenship might reduce voter turnout while increasing engagement in other political activities.
7. Discussion
This paper advances our understanding of how various dimensions of citizenship influence political participation amid transformative changes in its conception and practice, including the growing prevalence of dual citizenship. We present the first look at dual citizens in the U.S. using nationally representative samples. About four percent of Americans are citizens of both the U.S. and another country. Among them, in 2020, the number of those who acquired US citizenship at birth and those who were naturalized were similar, but in 2022, born-dual citizens outnumbered naturalized-dual citizens. The substantial numbers of born-dual citizens is an important reminder that studies of dual citizenship focusing only on immigrants are missing half the picture.
We examine the effect of being born in the U.S. and the effect of having two citizenship. The results point to the importance of nativity and socialization in affecting non-voting civic engagement. While naturalized-dual citizens tend to have a slightly higher socioeconomic status (which is predictive of participation), born-duals have an advantage in political engagement, at least outside the ballot box. Moreover, having dual citizenship also plays a relevant role in civic engagement. Compared to born-mono citizens, born-duals voted less in elections but participated more in a wide range of both electoral and non-electoral activities. These findings challenge the notion that political engagement is zero-sum and counter normative concerns that dual citizens disengage from American politics. As dual citizenship continues to expand and evolve, these empirical findings also help reduce speculative prejudice toward this group’s political behavior.
Beyond simply recognizing dual citizens as an increasingly visible category, this paper argues that it broadens citizenship categories, thereby allowing for novel theoretical comparisons. More research is needed to explore citizenship as both a status and a process, taking into account variations in psychological foundations, socialization resources, and political contexts. A deeper understanding of these dimensions will provide greater insight into how individuals become active members of a democratic polity.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/psrm.2025.10053. To obtain replication material for this article, https://doi.org/10.7910/DVN/9XBRQS.
Funding statement
None.
Competing interests
None.
Appendix Table of Contents
(1) Heterogeneous Nature of Dual Citizens
(2) Number of Dual Citizens
(a) Sampling Without Nativity Stratification
(b) Self-report Measurement Issue
(c) Question Misinterpretation
(3) Democracy Status of the Other Country of Citizenship
(4) Descriptive Characteristics by Citizenship Status
(5) Balance Test
(6) Matching on Political Factors
(7) Analyses for Each Type of Political Activity
(8) Levels of Vote Intention and Voter Registration
(9) Validated Vote
(10) Geographic Residence of Dual Citizens
(11) Home Ownership
Appendix A. Heterogeneous nature of dual citizens
Dual citizenship arises from various combinations of eligibility, applicability, and willingness.
First, eligibility pertains to how individuals acquire citizenship. A US-born person may be eligible for dual citizenship through two main routes: ancestry and naturalization. If at least one parent has citizenship from another country, the child may acquire dual citizenship at birth based on the jus sanguinis principle. Even if the parents did not apply for it at the child’s birth, the individual might claim it later in life. For example, increasingly, people of European descent use ancestry-tracing services to apply for another citizenship. The second route is through naturalization in another country, which involves various categories of admission. For example, one may marry a non-US citizen and apply for citizenship, or live in another country for a certain period and then apply for citizenship.
On the other hand, a foreign-born person may be eligible for dual citizenship through two main routes: ancestry and birthright citizenship. A child with at least one American citizen parent may also acquire another citizenship—either from the other parent, if they hold a different nationality, or through the jus soli principle in countries that grant citizenship by birthright. These categories are interesting cases, but their numbers are very small. According to the 2020 Census, 1.6% of the population is both foreign-born and born-citizen, approximately 3 million Americans total. In the CES data their numbers would be small, but we assume that Group B may be misclassified as Group C, and Group E as Group F in Table 1. We are unable to obtain sufficient data in the survey for mono and dual citizens who were born Americans abroad via jus sanguinis.
Second, even if one is hypothetically eligible for dual citizenship, the country must allow dual citizenship. This complicates inferring the number of potential dual citizens among those born in the U.S., as the data on the citizenship of their parents or grandparents is scarce.
Third, even if one is eligible and the country allows dual citizenship, the individual (or their parents) needs to choose to obtain another citizenship. This decision depends on the individual’s context and their assessment of the costs and benefits of acquiring another citizenship. It may also reflect their level of state attachment. Socioeconomic factors can also influence this decision. Bloemraad Reference Bloemraad(2004) explored who claims dual citizenship by examining Canadian census samples in the 1980s and 1990s. She found that first-generation immigrants with higher human capital, rather than the economically marginalized, are more likely to embrace dual citizenship. Additionally, dual citizenship claims are lower for child migrants and are higher for immigrants who speak a household language other than English and French.
We recognize that for some born-dual US citizens, a great deal of effort was expended to achieve that status. Gaining citizenship via marriage or ancestry requires paperwork and evidence. Even dual citizenship acquired via jus sanguinis requires engaging with another government bureaucracy, although that is probably easier, relatively speaking, and also more common. Unfortunately, there is no available information about how native-born dual-citizens acquired their non-American citizenship. About 12 percent of Americans marry someone from another country (Van Dam 2023), and one estimate of how many Americans could be eligible for dual citizenship via family ancestry was about 40 percent (Bloemraad, Reference Bloemraad2022). That said, it is unknown how many Americans take advantage of the possibility and have the resources to achieve dual citizenship. Unfortunately, our data can only tell us that 31 to 44 percent are or have been married, but we do not know the spouse’s nationality. Relatedly, personality type of the respondents may be an omitted variable that explains a higher level of non-voting political participation, albeit only for those who actively chose their second citizenship as adults; it also would not explain why born-dual citizens are less likely to vote than born-mono citizens.
Appendix B. Number of dual citizens
In 2020, for the 2,452 respondents (4.0% of the sample) who reported dual citizenship, the confidence interval for dual citizenship was 3.89%–4.20%, and in 2022, it was 4.25%–4.58%. These non-overlapping intervals indicate a real increase in the proportion of individuals reporting dual citizenship between the two years. However, we cannot determine whether there was actual demographic growth in this time period. According to YouGov, about a quarter of the sample from 2022 was also part of the 2020 sample; as such, some respondents may have learned about or become more comfortable with the survey question or identity with repetition.
This number should be regarded as an approximation rather than a definitive one. Below, we outline four key factors to consider.
B.1. Sampling without nativity stratification
The estimate could be a conservative estimate. CES has a nationally representative sample in terms of demographic and socioeconomic indicators, but it is not stratified by nativity criteria. Therefore, the foreign-born population is underrepresented in CES 2020 data, for example. In the 2020 Census, 85% of the population was born in the U.S., 7% were naturalized, and 6.7% were non-US citizens. In the 2020 CES data, 92.6% of the sample was born in the U.S., 5.4% were naturalized, and 2.0% were non-US citizens. Moreover, CES surveys only adults, and the composition of those over the age of 18 in Census 2020 are: 83.5% born in the U.S.; 8.6% naturalized, and 7.8% non-US citizens. Therefore, it is possible that the number of dual citizens in the country could be greater than what we see in the current data.
B.2. Self-report measurement issue
The estimate could also be an overestimate. This number is based on a self-report measure, and we do not have the means to verify the de jure or de facto legal status. Laws about dual citizenship in the U.S. are implied rather than explicit, and enforcement is different from legislation. The measurement errors pose challenges for all studies employing survey methods. This self-report measurement issue extends beyond the current data assessing dual citizenship status but also applies to other datasets concerning various types of citizenship status. Despite the limitations of relying on self-identification of citizenship status, we think that a subjective understanding of citizenship status is important for our research question on the relationship between citizenship status and political participation. We expect that people’s subjective understanding, more than the de jure or de facto legal status, explains their levels of political participation.
Some of the measurement errors may stem from dishonesty or ignorance. Some may have purposefully claimed or hidden their dual citizenship status, while others may have misunderstandings. For instance, some may think they have two citizenship but the official paperwork may not have actually been filed or completed. Some may not be aware of their dual citizenship status; their parents may have applied for citizenship at their birth but did not necessarily communicate this information to their children. Some may also think one of their citizenship has been renounced without that being the case. For example, some countries do not allow the renunciation of their citizenship (e.g., Estonia, North Korea, Uruguay), so one can have de facto dual citizenship.
Relatedly, some people may believe they have dual citizenship when their country does not allow it (e.g., India). MACIMIDE Global Expatriate Dual Citizenship Dataset reports that in 2020, 76 percent of countries maintained a tolerant approach towards dual citizenship. While the trend is in a broadly similar direction across world regions, dual citizenship acceptance has progressed faster in the Americas and Oceania and slower in Africa and Asia. Over the past two decades, nearly all changes in policy have been towards a tolerant approach.
At the same time, there is a great deal of ambiguity in dual citizenship law. For example, India and China are among the few states that reject dual citizenship. Nevertheless, India has its own category, Overseas Citizenship of India (OCI). China has a law that takes away household registration status when an individual naturalizes in another state, but the law is under-enforced (Shachar et al. (Reference Shachar, Bauböck, Bloemraad and Vink2017)).Footnote 12 Official government statistics reflect this complexity as well: the Canadian Census includes a question about dual citizenship (Canada: Statistics Canada 2021), and there are significant numbers of individuals who claim to be dual citizens of countries that are known to be intolerant of dual citizenship. The following is the list of countries of another citizenship:Footnote 13
Countries known to be tolerant of dual citizenship
Canada, UK, Mexico, Ireland, Italy, Germany, Nigeria, Philippines, France, Dominican Republic, Colombia, Venezuela, Cuba, Jamaica, Brazil, Israel, Argentina, Australia, Peru, Russia, Ecuador, El Salvador, Poland, Spain, Switzerland, Guatemala, Hong Kong, Portugal, Chile, Lebanon, American Samoa, Egypt, Honduras, Pakistan, Panama, Vietnam, Greece, Republic of Korea, Nicaragua, Sweden, Guyana, Kenya, Romania, Albania, Barbados, Bulgaria, New Zealand, South Africa, Belize, Costa Rica, Georgia, Aland Islands, Austria, Bangladesh, Bolivia, Finland, Jordan, Trinidad and Tobago, Belgium, Cape Verde, Croatia, Denmark, Dominica, Ghana, Iraq, Norway, Senegal, Serbia, Montenegro, Turkey, Uruguay, Angola, Antigua and Barbuda, Bosnia and Herzegovina, Cambodia, Hungary, Jersey, Luxembourg, Morocco, Palestinian Territory, Occupied, Sudan, Syrian Arab Republic, Algeria, Anguilla, Armenia, Cyprus, Czech Republic, Guinea, Palau, Saint Helena, Sri Lanka, British Virgin Islands, Aruba, Burkina Faso, Cayman Islands, Democratic Republic of the Congo, Cook Islands, Falkland Islands (Malvinas), French Guiana, French Southern Territories, Gambia, Grenada, Guadeloupe, Latvia, Liberia, Lithuania, Martinique, Moldova, Republic of, New Caledonia, Niger, Norfolk Island, Northern Mariana Islands, Rwanda, Saint Kitts and Nevis, Saint Lucia, Saint Vincent and the Grenadines, Slovakia, Slovenia, Somalia, Timor-Leste, Togo, Turks and Caicos Islands, Uganda, Zambia, Zimbabwe, Macao, Thailand, Haiti, Yemen, Papua New Guinea, Netherlands
Countries known to be intolerant of dual citizenship, but still have been reported by the respondents
Afghanistan, Cameroon, Singapore, India, Japan, China, Ukraine, Iran, Bahamas, Belarus, United Arab Emirates, Indonesia, Democratic People’s Republic of Korea, Kuwait, Andorra, Azerbaijan, Bhutan, Botswana, Kazakhstan, Monaco, Mozambique, Nepal, Oman, San Marino, Swaziland
Appendix C. Question misinterpretation
The CES provided an extensive list of countries for country of citizenship, totaling 243 options, which also included a continent (e.g., Antarctica) and U.S. territories (e.g., Puerto Rico, Guam, and the United States Minor Outlying Islands). Consequently, the data on dual citizens include individuals who claim citizenship of both the U.S. and Puerto Rico, for example. This understanding of dual citizenship differs from our intended purpose of examining legal status in two countries (rather than belonging to one nation and one subnational political entity).
We also need to address the number of respondents who indicated they are a citizen of another country and then selected “United States Minor Outlying Islands.” This is problematic because the numbers of respondents who chose this option in both years of the CES exceed the entire population living on the islands. Our interpretation is that some respondents may not have understood what it means to have a citizenship in addition to that of the U.S. They may have chosen “yes” due to this uncertainty and, after scanning the long list of options, they selected the option closest to “United States” rather than go back and change their response to the previous question in the survey. We also believe a similar uncertainty led respondents to select “Other” given the comprehensive list of countries presented. At the end of reading through the list, they chose “Other” because there was no “Don’t Know” or “I do not have another citizenship” option. Therefore, we created a filtered version of the data in which we do not count as dual citizens those who indicated they are a citizen of another country and then chose either “United States Minor Outlying Islands” or “Other.”
As a result, we have three versions of the dataset. The first version is the unfiltered data, where we solely relied on respondents’ declarations of their citizenship status. The second version excluded those who indicated they are a citizen of another country and then chose either “United States Minor Outlying Islands” or “Other”, which was used for analyses in the main text. Lastly, the third version filtered out countries known to be intolerant of dual citizenship. The number of countries that the dual citizens have citizenship from decreases from 161 to 132 in 2020 and decreases from 154 to 126 in 2022. In 2020, the number of respondents who stated that they are a citizen of another country in addition to the U.S. is 2,024. Among them, the number of dual citizens who were born in the U.S. (
$N = 894$) is similar to foreign-born-dual citizens (N = 1,130). In 2022, the number of respondents who stated that they are a citizen of another country in addition to the U.S. is 1,941. Among them, 823 respondents are born-dual citizens and 1,118 respondents were naturalized-dual citizens. The results of our analyses using the three different versions of the data show similar patterns.
Appendix D. Democracy status of the other country of citizenship
Another caveat concerning countries for our analyses is that the country of the other (potential) citizenship may be a confounding factor in our comparison between US-born mono and dual citizens. Unlike the comparison between born- and naturalized-dual citizens, the born-mono citizens in this comparison may have very different family ties than born-duals. Some may have no connections to another country; others may be eligible for another citizenship but opt out from it; and yet others may have connections to another country, but the other country does not allow dual citizenship. Given that there is a positive correlation between the countries that tolerate dual citizenship and the countries that are democratic, dual citizens may have had more widespread exposure to democratic norms in their socialization.
However, there is very little variation in our data regarding whether the other country of citizenship is a democracy.Footnote 14 Over 80% of the dual citizens in our dataset hold citizenship in democratic countries. This reflects the general tendency for democratic countries to allow dual citizenship. In the analysis that includes a dichotomized variable of democracy, we found no significant changes in our results.
Furthermore, regarding the consideration of extra-US socialization, the current dataset is limited in capturing the level of socialization in both the US and the other country of citizenship, which we hope to address in a future project. Nevertheless, we believe that some of the demographic variables we matched on, particularly immigrant generation, can provide insights. To explore potential differences in the level of extra-US socialization, it would be valuable to compare born-mono and born-dual citizens with immigrant parents, which could reveal whether the influence of the parents’ country of citizenship differs between these groups. By matching on the second-generation variable, we could isolate some elements of extra-US socialization. However, the current dataset does not include information on the parents’ citizenship for born-mono citizens, nor does it specify whether born-dual citizens’ other citizenship is derived from their parents’ citizenship or acquired through naturalization.
Top countries of another citizenship for dual citizens (2020 and 2022)
Note: The numbers in parentheses indicate the number of dual citizens associated with the respective country.
Appendix E. Descriptive characteristics by citizenship status
The born-mono citizen group is the oldest on average, and there are more women in both mono citizen groups compared to the dual citizen groups. The born-mono citizen category has the highest proportion of non-Hispanic Whites, followed by the born-dual citizen category, while the naturalized categories have lower proportions. Born-mono citizens also have longer family histories in the country, with most having all grandparents born in the U.S., compared to born-dual citizens, who often have immigrant parents or grandparents. Born-mono citizens are least likely to live in cities, while naturalized citizens, who more often reside in urban areas, align with literature suggesting immigrants tend to reside in urban areas (Yang, Reference Yang1994; Bloemraad, Reference Bloemraad2002). Naturalized citizens had significantly higher education levels compared to US-born citizens, supporting the link between socioeconomic status and naturalization (Frideres et al., Reference Frideres, Goldenburg, Disanto and Horna1987; Bloemraad, Reference Bloemraad2004). Income levels and employment status varied, with naturalized citizens tending to be wealthier and U.S.-born citizens more frequently employed.
US-born citizens consistently had a higher level of political knowledge and political interest than naturalized citizens, and in 2020, the born-mono citizens had the highest levels of both. Political knowledge is measured combining two questions about which party has a majority of seats in the House of Representatives and the Senate, and political interest is measured with a question about how much the respondent follows what is going on in government and public affairs.
As we conjectured earlier, awareness of dual citizenship as an adult does not seem to be an indicator of political interest or knowledge; naturalized-dual citizens and born-dual citizens express very similar levels of political interest and knowledge. Finally, each citizenship group is moderate or slightly left-leaning in terms of political ideology, but compared to mono citizens, dual citizens were more likely to identify themselves as Democrats.
To see if there are differences among the four citizenship categories in terms of socio-economic status, we conducted an ANOVA test (Table E1). The results indicated that all socio-economic indicators, except family income, were statistically significant at the 0.05 level. As a follow-up, we performed Tukey’s Honest Significant Difference (HSD) test, a post hoc analysis used to assess the significance of differences between pairs of group means. Table E2 presents the mean differences for six pairs of groups across various socio-economic indicators for both 2020 and 2022. The findings reveal substantial differences in age, race/ethnicity, immigration generation, and marital status among the four groups in both years. Additionally, notable differences are observed particularly between born-mono and born-dual citizens, as well as between naturalized-dual and born-dual citizens.
ANOVA results for socio-economic indicators by citizenship status
Mean differences and significance in socio-economic indicators across groups for 2020 and 2022
Appendix F. Balance test
First Comparison (H1)
Given that the numbers of naturalized-dual citizens and born-dual citizens are quite similar in both years, we aimed to use as much of the sample as possible. Among the covariates, age, education, and family income were continuous variables, so we set specific matching criteria to approximate the treated and control groups based on the distances between these covariates. There are some people with missing income and employment, and we matched on their missingness. This improves the overall balance between the two groups. In the analysis of matching naturalized-dual citizens (control) and born-dual citizens (treatment), the results for the 2020 dataset show that the omnibus 
$d^2$ balance test (Hansen and Bowers, Reference Hansen and Bowers2008) yielded a 
$p$-value of 0.98 with an effective sample size of 1005.7, representing the equivalent number of matched pairs (Table F1). For the 2022 dataset, the test resulted in a 
$p$-value of 0.99 with an effective sample size of 1089, also representing the equivalent number of matched pairs (Table F2). These results suggest that the matching processes for both years were successful in creating comparable groups for analysis.
Balance test (H1) for 2020
Balance test (H1) for 2022
Second comparison (H2)
Because we prioritized balance on these same demographic factors and we had many more controls than treated, we restrict the matches to be within certain distances on the pretreatment covariates. For instance, we exact matched on gender and used calipers for age (so that respondents in a pair should have a smaller distance than the caliper width.) Matching was completed using propensity score matching with the Mahalanobis distance and a caliper for age to approximate treated and control groups. After the matching, we had an effective sample size of 1,608 (the number of matched observations). Table F3 below shows balances for the matched mono and dual citizens for 2020. The observed pre-treatment covariates in the two matched groups are well balanced (
$p=.96$ for the omnibus 
$d^2$ balance test(Hansen and Bowers, Reference Hansen and Bowers2008). For the 2022 dataset, the matching procedure resulted in an effective sample size of 1,893, with the omnibus 
$d^2$ balance test indicating good covariate balance (
$p$-value of 0.98) (see Table F4).
Balance test (H2) for 2020
Balance test (H2) for 2022
Appendix G. Matching on political factors (Model 2)
In Model 2, we matched on both demographic and political variables, including political knowledge, political interest, partisanship and partisanship strength. We created the partisan strength variable using the 7-point party identification scale. We coded Strong Democrats (1) and Strong Republicans (7) as 1, and all others as 0.
The results remained consistent with our original findings even after accounting for these political factors. Born-mono citizens continued to exhibit higher voter turnout than born-dual citizens in both the 2020 and 2022 elections (one versus two citizenship). And, born-dual citizens still demonstrated higher engagement in other forms of political participation compared to born-monos during both years. However, in Model 2, which includes partisan strength, the effects of having another citizenship or nativity weaken compared to Model 1, with some losing significance entirely, suggesting that partisan strength absorbs much of the explanatory power. This decrease may come from the difference in political donations between born-dual and born-mono citizens disappearing after adding political variables as covariates.
Comparison of political behavior among dual citizens
Note: Robust standard errors in parentheses. *
$p \lt 0.05$, **
$p \lt 0.01$, ***
$p \lt 0.001$.
Comparison of political behavior among American citizens born in the U.S
Note: Robust standard errors in parentheses. *
$p \lt 0.05$, **
$p \lt 0.01$, ***
$p \lt 0.001$.
Appendix H. Analyses for each type of political activity separately
H.1. Descriptive statistics on political activities by citizenship category
Political participation for each type of activity
H.2. Political participation among dual citizens
In both figures, positive values indicate that born-dual citizens participate at higher levels than naturalized-dual citizens.
The effect of nativity on different types of political participation (Model 1: sets matched on demographic variables).
The effect of nativity on different types of political participation (Model 2: sets matched on demographic and political variables).
H.3. Political participation among born citizens
In both figures, positive values indicate that born-dual citizens participate at higher levels than born-mono citizens.
The effect of dual citizenship on different types of political participation (Model 1: sets matched on demographic variables).
The effect of dual citizenship on different types of political participation (Model 2: sets matched on demographic and political variables).
Appendix I. Levels of vote intention and voter registration
Born-dual and naturalized-dual citizens also expressed very similar levels of vote intention and voter registration in 2020. In 2022, born-duals had slightly higher voter registration levels than naturalized duals.
The effect of nativity on voter registration and intention (sets matched on covariates).
Born-dual and born-mono citizens expressed similar levels of vote intention and voter registration, with born-duals expressing slightly greater intent to vote than born-monos in 2022.
The effect of dual citizenship on voter registration and intention (sets matched on covariates).
Appendix J. Validated vote
For the 2020 CCES data, individual records were matched to the Catalist database of registered voters in the United States, with the matching performed in June 2021. Respondents were coded as one of the followings: absentee, early vote, mail, polling place, and unknown method of voting. For the 2022 CCES data, individual records were matched to the TargetSmart database of registered voters in the United States, with the matching performed in August 2023. Respondents were coded as one of the followings: absentee, early mail, polling place, provisional, voted by unknown method, and did not vote. For both years, we coded respondents as 1 if they voted and 0 if they did not vote or were defined as non-voters.
In 2020, 45,660 respondents reported that they voted, but vote validation showed that only 39,198 actually voted in the 2020 election. In 2022, 40,478 individuals reported that they voted in the last election, but vote validation revealed that only 33,573 did.
We thought it was possible that pressures to say one voted (regardless of whether one actually did so) might differ depending on one’s socialization or citizenship status. However, after replicating our analyses using validated vote rather than self-reports, we see the same pattern (Table J1): born-dual citizens are more likely to have voted than naturalized-dual citizens, and born-dual citizens were less likely to vote than born-mono citizens.
Comparing vote validation in 2020 and 2022
$^{***}p \lt 0.001$; 
$^{**}p \lt 0.01$; 
$^{*}p \lt 0.05$; 
$^{\dagger}p \lt 0.1$.
Appendix K. Geographic residence of dual citizens
Since dual citizens are concentrated in a handful of states, and because states can vary in their voting rules and laws, we explored whether the geographic residence of our respondents could be driving these effects.
Geographical distribution of dual citizens in the U.S. (2022).
Figure K1 shows that dual-citizens tend to cluster in specific states: California, Texas, Florida, and New York (Figure K1). Among born-dual citizens, 41.45% (2020) and 39.39% (2022) resided in four major states. Among naturalized-dual citizens, 48.87% (2020) and 50.07% (2022) resided in these same states.
In Tables A12–A15, we replicated our analyses comparing US-born mono and dual citizens. We obviously have less power to distinguish effects, but we find the same pattern in California, Texas, Florida, and New York: born duals tend to vote less than born monos but they engage more in a wide range of non-voting political actions relative to born monos. In other words, differing institutional requirements across states are not driving our pattern of results.
The effect of dual citizenship on U.S.-born citizens in California
$^{***}p \lt 0.001$; 
$^{**}p \lt 0.01$; 
$^{*}p \lt 0.05$.
The effect of dual citizenship on U.S.-born citizens in New York
$^{***}p \lt 0.001$; 
$^{**}p \lt 0.01$; 
$^{*}p \lt 0.05$.
The effect of dual citizenship on U.S.-born citizens in Texas
$^{***}p \lt 0.001$; 
$^{**}p \lt 0.01$; 
$^{*}p \lt 0.05$; 
$^{\dagger}p \lt 0.1$
The effect of dual citizenship on U.S.-born citizens in Florida
$^{***}p \lt 0.001$; 
$^{**}p \lt 0.01$; 
$^{*}p \lt 0.05$.
Appendix L. Home ownership
We examined whether born-monos and born-duals differ in the stakes they feel they have in their local community, focusing on home ownership and the length of residency at their present address. In 2020, born-mono citizens had the highest home ownership rate at 61%, followed by naturalized-dual citizens at 60%, and naturalized-mono citizens at 59%. Born-dual citizens had the lowest home ownership rate among the groups, at 54%. This trend persisted in 2022, with born-monos again having the highest home ownership rate (61%), while born-duals had the lowest rate (52%). For the length of residencyFootnote 15, born-mono and naturalized-mono citizens tend to have longer tenures at their current addresses (3 to 4 years) compared to born-dual and naturalized-dual citizens, whose average residency is 1 to 2 years in both 2020 and 2022.
We matched born-dual citizens and born-mono citizens again by adding home ownership and the length of residency at their present address to the set of covariates as mentioned above. When we compared both groups, the regression results are consistent with the previous findings that did not include home ownership and the length of residency as covariates (see Table L1). Born-mono citizens voted more than born-dual citizens in both 2020 and 2022 elections, while born-dual citizens are more engaged in other forms of political participation than born-monos in both years.
Comparison of political behavior among American citizens born in the U.S
Note: Robust standard errors in parentheses. *
$p \lt 0.05$, **
$p \lt 0.01$, ***
$p \lt 0.001$.
Open access
