Citizenship as Privilege and Social Identity: Implications for Psychological Distress

Abstract

Citizenship is both a system of privilege and a source of social identity. This study examines whether there are disparities in psychological distress between citizens and noncitizens, and whether these disparities may be explained by markers of social disadvantage (e.g., poverty, discrimination) or perceptions of success in the United States (i.e., subjective social status). We analyze data from the Asian subsample (n = 2,095) of the National Latino and Asian American Study. The data show that noncitizens report greater psychological distress compared with naturalized citizens and native-born citizens after accounting for sociodemographics (e.g., age, gender, Asian subgroup), socioeconomic characteristics (education, employment, income-to-poverty ratio), immigration (e.g., interview language, years in the United States, acculturative stress), health care visits, and everyday discrimination. Preliminary evidence suggests that subjective social status may explain some of the disparities between naturalized citizen and noncitizen Asian Americans.

Introduction

Despite being one of the most profound forms of social stratification, citizenship remains severely understudied in the health literature (Gee & Ford, 2011; Gubernskaya, Bean, & Van Hook, 2013). The lack of research on citizenship is glaring given strong evidence that health is patterned along other dimensions of social inequality, such as race, social class, gender, and sexual orientation (Adler & Stewart, 2010; Lindley, Walsemann, & Carter, 2011; Williams, Mohammed, Leavell, & Collins, 2010). The broad literature on structural inequities shows that persons with fewer social resources suffer greater morbidity compared with those with more resources (Diez Roux, 2012; Link & Phelan, 1995). This argument can be extended to consider how citizenship is also related to illness. Indeed, a small literature finds that noncitizens report poorer health and lower access to health services compared with citizens (Bustamante & Chen, 2012; Carrasquillo, Carrasquillo, & Shea, 2000; Gubernskaya et al., 2013; Reyes & Miranda, 2015; Yu, Huang, & Singh, 2004).

In the present study, we examine how patterns of psychological distress vary by citizenship among Asian Americans. The extant literature on citizenship and health is largely descriptive. Herein, we consider two theoretical mechanisms that might explain citizenship disparities in mental health.

First, citizenship constitutes a fundamental dimension of structural inequality, and an important marker of immigrant integration (National Academies of Sciences, Engineering and Medicine, 2015). Citizens enjoy rights and privileges that are often unavailable to noncitizens. In the United States, many federal and state government programs explicitly or implicitly exclude noncitizens from eligibility. For example, in most circumstances, noncitizens are ineligible for federal student loans, jury duty, and service in Congress (Bean & Stevens, 2003). Even when they are eligible, noncitizens face many barriers. For example, many immigrants who are eligible to receive public assistance avoid doing so. This is because they fear that using services will label them as a public charge and invalidate their applications for citizenship (Berk & Schur, 2001; Gee & Ford, 2011; Johnson, 1994).

From this perspective, citizens experience considerable advantages compared with noncitizens. Noncitizens face barriers to employment, educational attainment, and wages (Sumption & Flamm, 2012). A recent study found that U.S. federal courts provided harsher sentences against noncitizens than citizens for similar crimes (Light, Massoglia, & King, 2014). Additionally, noncitizens may encounter more discrimination in their day-to-day encounters, such as when trying to obtain a loan or traveling (Gee & Ford, 2011). All of these factors are related to health. Persons with fewer socioeconomic resources and those who report more discrimination generally have greater morbidity compared with persons with more resources and who report less discrimination (Braveman, Cubbin, Egerter, Williams, & Pamuk, 2010; Gee, Spencer, Chen, Yip, & Takeuchi, 2007; Williams et al., 2010). Hence, it is possible that any disparities between citizens and noncitizens are explained by socioeconomic factors or discrimination.

Additionally, naturalized and native-born citizens are not fully equivalent. Some research suggests that the accumulation of disadvantages over the life course may contribute to disparities between social groups (Ferraro & Shippee, 2009; Umberson, Williams, Thomas, Liu, & Thomeer, 2014; Walsemann, Geronimus, & Gee, 2008). Naturalized citizens do not enjoy citizenship privileges until they are naturalized, whereas native-born citizens enjoy their rights from birth.¹ This motivates us to investigate whether native-born citizens may have lower rates of psychological distress compared with naturalized citizens.

Second, citizenship constitutes a dimension of social identity which is perhaps especially salient for immigrants (Andreouli & Howarth, 2013; Bosniak, 2008). Naturalized citizens must, in a formal public setting, sever their identities to their former countries and simultaneously adopt a new identity as an American. Naturalized citizens in the United States must swear the Oath of Allegiance, which states that they “absolutely and entirely renounce” their loyalties to their home countries and that they will “bear true faith and allegiance” to the United States.

We borrow some concepts from identity control theory (Burke, 1991; Stets & Burke, 2005), which provides a useful framework for considering the social identity aspect of citizenship on health outcomes. An “identity” refers to the meanings that people have about themselves as members of groups or as actors within a role. These meanings are pegged to an “identity standard,” an abstract reference from which individuals compare themselves. Positive emotions arise when one’s current and idealized standards match, while negative emotions arise when they are dissonant. At high levels of dissonance, individuals may feel distress and anxiety (Burke, 1991).

A key aspect of identity is membership in social groups, whereby the identity standards may derive from the groups, and such groups may vary in accord with their location within a social hierarchy (Stryker & Burke, 2000). As noted above, citizens are stratified socially, culturally, and legally (Gee & Ford, 2011). The immigrant narrative is replete with stories of aspirations to acquire the “American dream,” which often connotes economic success, as well as cultural and structural acceptance (historically termed “assimilation”); this acceptance is formally and legally demarked via naturalized citizenship (National Academies of Sciences, Engineering and Medicine, 2015).

Some identities may not be based on one’s current situation, but part of their “imagined future” whereby individuals do not yet feel that they can currently claim a given identity, but hope to in the future (Frye, 2012). Such identities that may be sought after, but not yet achieved, may be termed, “aspirational identities” (Thornborrow & Brown, 2009). Frye (2012) documents these aspirational identities using the case study of children who seek ambitious educational goals.² Although identity control theory generally focuses on identity standards that are salient in the present moment, it seems plausible that aspirational identities may serve a similar function. That is, individuals may compare themselves with an aspirational identity, and when one’s present situation is dissonant with that aspiration, negative emotions and distress may arise.

Citizenship may be considered an example of an aspirational identity among some immigrants. Immigrants who have become naturalized citizens may have more positive emotions compared with immigrants who have not been naturalized. One potential way of capturing this comparison process is to examine subjective social status (Jackman & Jackman, 1973), referring to one’s appraisal of her position on the social ladder (Adler, Epel, Castellazzo, & Ickovics, 2000). Noncitizens may perceive themselves lower on the ladder compared with citizens, and failure to rise on the ladder may be related to dissonance and distress. Subjective social status may reflect an immigrant’s view of her current social circumstances as well as her future prospects (Chen, Gee, Spencer, Danziger, & Takeuchi, 2009).

Perceptions of low social status may contribute to a sense of relative deprivation, status insecurity, shame, and anxiety (Adler et al., 2000; Scott et al., 2014). Consistent with this view, persons who report low subjective social status are more likely to suffer from mental health problems in numerous countries even after accounting for more traditional measures of socioeconomic status (Scott et al., 2014). Among Asian American immigrants, subjective social status is related to poor mental health and poor self-rated health (de Castro, Gee, & Takeuchi, 2010; Leu et al., 2008).

These considerations motivate the present study, where we investigate the following hypotheses:

Hypothesis 1:

Citizens will report lower psychological distress compared with noncitizens.

Hypothesis 2:

Among citizens, naturalized citizens will report greater psychological distress compared with native-born citizens.

Hypothesis 3:

According to the structural inequality perspective, socioeconomic factors and discrimination will explain the disparities in psychological distress between citizenship statuses.

Hypothesis 4:

According to identity control theory, higher subjective social status may explain the disparities in psychological distress between citizenship statuses.

Our study focuses on psychological distress for several reasons. First, not only is psychological distress intrinsically important, but it is also related to other health problems. For example, psychological distress has a strong association with mortality (Huppert & Whittington, 1995; Russ et al., 2012). Second, emotional outcomes are part of the propositions developed in identity control theory. Third, the small literature on citizenship has focused almost exclusively on access to health services (Bustamante & Chen, 2012; Yu et al., 2004) or global markers of self-rated health (Gubernskaya et al., 2013), yet there is no reason to suspect that these are the only outcomes related to citizenship. Fourth, although psychological distress has been studied among Asian American immigrants, prior studies have not considered the issues of citizenship (M. C. Lee et al., 2015; Leong, Park, & Kalibatseva, 2013; Mossakowski & Zhang, 2014).

In considering these hypotheses, we recognize that the study of citizenship is intertwined with the study of immigrants. A large literature has shown that a variety of immigration-related factors are associated with mental health (Landale, Hardie, Oropesa, & Hillemeier, 2015; Takeuchi et al., 2007). For example, the process of adaptation to a new country may generate “acculturative stress,” which may then be related to psychological distress (Gee et al., 2007, Sirin, Ryce, Gupta, & Rogers-Sirin, 2013). Similarly, factors such as duration in the United States and English language usage are also related to mental health (Schachter, Kimbro, & Gorman, 2012). Hence, it is important to evaluate whether the findings related to citizenship may actually be confounded by such immigration-related factors.

Furthermore, because research shows that noncitizens generally have lower access to health services (Bustamante & Chen, 2012; S. Lee & Matejkowski, 2012), our analysis considers receipt of health services for both physical ailments and mental health problems. Measuring receipt of services for any health issue is important because some research suggests that Asian Americans may express their mental health problems via physical symptoms (i.e., psychosomaticization) and thus, seek treatment for physical ailments to address underlying psychological issues (Grover & Ghosh, 2014; Maffini & Wong, 2014). Accordingly, we also control for use of a variety of medical health services, making the analyses a more conservative test of the major hypotheses.

Method

Data came from the National Latino and Asian American Study (NLAAS; Alegria, Takeuchi, et al., 2004; Takeuchi, Gong, & Gee, 2012). This nationally representative household survey, conducted between 2002 and 2003, used a three-component sampling procedure: (a) housing units and household members were sampled from a core sampling of primary sampling units (metropolitan statistical areas and counties) and secondary sampling units (continuous groupings of census blocks) with probability proportional to size, (b) ethnic groups were targeted by sampling high-density supplemental census block groups in which the ethnic groups made up more than 5% of the population, and (c) secondary participants were recruited from households where a primary respondent had already been interviewed. The response rates for primary and secondary respondents (calculated using the American Association for Public Opinion research, Response Rate Method 3) were 69% and 74%, respectively (Alegria, Takeuchi, et al., 2004). Sample weights were used to account for the joint probabilities of selection, given these three sampling components, and to provide nationally representative sample estimates (Heeringa et al., 2004).

Respondents were 18 years and older and resided in the United States at the time of their interview. The trained interviewers had linguistic and cultural backgrounds similar to those of the respondents. Survey instruments were translated from English into Spanish, Cantonese, Mandarin, Tagalog, and Vietnamese, and verified with back-translation (Alegria, Vila, et al., 2004). Interviewers administered the survey in the respondent’s preferred language using computer-assisted software. Interviews were conducted face-to-face unless a respondent requested a telephone interview.

The sample for this study was restricted to those who self-identified as Asian, Native Hawaiian, or Pacific Islander. Latinos were excluded from the present analysis because their experiences are qualitatively different from those of Asians. For example, Puerto Ricans are U.S. citizens even though they are also sometimes considered “foreign born.” Of the original NLAAS sample of 2,095 Asians, Native Hawaiians, or Pacific Islanders, 80 people were excluded due to missing values for variables of interest. This resulted in an analytic sample of 2,015.

Psychological Distress, the main dependent variable, was measured using the 10-item Kessler Psychological Distress Scale (K-10) that provides a global assessment of distress over the past 30 days (Kessler et al., 2002). The K-10 is widely used in population-based studies in several countries and has been previously used with Asian American populations (Chatterji, Alegria, Lu, & Takeuchi, 2007; Yip, Gee, & Takeuchi, 2008). Respondents were asked, how often in the past 30 days they felt: depressed, so sad nothing could cheer them up, hopeless, restless or fidgety, so restless they could not sit still, tired out for no good reason, that everything was an effort, worthless, nervous, and so nervous that nothing could calm them down. Responses for each item ranged from 1 none of the time to 5 all of the time. Individual items were summed; higher values indicate greater psychological distress (mean = 13.3; range 10–44).

Citizenship Status, the main independent variable, was constructed from responses to three nested questions that asked if respondents (a) were citizens of the United States, (b) were born U.S. citizens or naturalized, and (c) held citizenship in another country. We categorized respondents as native-born (U.S.-born) citizens, naturalized citizens, noncitizens, or dual citizens. The latter refers to respondents who have joint citizenship in the United States and another country.

Acculturative Stress was measured using a 10-item scale adapted from the Mexican American Prevalence and Services Survey (Vega et al., 1998) that captures strains associated with cultural change. Only immigrants answered questions on acculturative stress. The items asked whether respondents experienced: guilt for leaving behind family and friends, limited contact with family/friends from living in the United States, the same level of respect as they had in their country of origin, difficulties interacting with others because of language, poor treatment because of their language skills, difficulties with finding work because of their ethnicity, questioning about their legal status, concerns of being deported, or avoided seeking health services due to fear of immigration officials. Acculturative stress scores ranged from 0 to 10, with higher scores indicating more acculturative stress (Kuder–Richardson Formula 20 = 0.59).

Language of the Interview was coded as “English,” “Chinese,” “Vietnamese,” or “Tagalog.” Although the NLAAS rigorously translated the instruments using standard translation–back-translation methods (Alegria, Vila, et al., 2004), there may nonetheless be differences in reporting due to the use of translated instruments. Furthermore, language is a major consideration in the study of immigrant health (Gee, Walsemann, & Takeuchi, 2010; Ro & Bostean, 2015). Accordingly, we controlled for language of the interview.

Income-to-Poverty Ratio was the reported household income divided by the poverty threshold, as defined by the U.S. Census. This ratio is calibrated to account for the number of family members and their ages. The higher this ratio, the further away a person is from the poverty line (i.e., more economically advantaged).

Employment denoted whether the respondent was currently employed, unemployed, or out of the labor force (e.g., a homemaker). Education was a categorical variable: less than high school, high school graduate, some college, and college degree or more.

Subjective Social Status was determined using the MacArthur Scale of Subjective Social Status, which asks respondents to rank their socioeconomic status based on their perception of their income, education, and occupation/work status relative to other people in the United States using a graphical representation of a ladder (Adler et al., 2000). The measure ranges from 0 to 10, with 0 representing people at the bottom and 10 representing people at the top.

Everyday Discrimination was a 9-item scale adopted from the Detroit Area Study to measure perceptions of chronic and routine unfair treatment (Williams, Yan, Jackson, & Anderson, 1997). Representative items include how often respondents report: being treated with less courtesy than others, people act as if you are not as good as they are, being called names or insulted, and being threatened or harassed. Responses ranged from 1 to 6, never, less than once a year, a few times a year, a few times a month, at least once a week, and almost every day. Items were summed to create a total score of everyday discrimination, with higher scores indicating more discrimination.

Health Care Visits were the total number of visits in the past 12 months. This included seeing mental health professionals (e.g., psychiatrist, social worker, counselor), physicians (family doctor, urologist), or other professionals (e.g., nurse, occupational therapist).

Asian Subgroup denotes the following subgroups: Chinese, Filipino, Vietnamese, Asian Indian, Japanese, Korean, Native Hawaiian, other Pacific Islander, and other Asian.

Age, Gender, and State of Residence were included as sociodemographic controls. Social Desirability captures reporting bias related to desires to present oneself in a favorable manner, a concern given that all of our measures were self-reported. Social desirability was assessed using a 10-item scale (Zuckerman, Michael, Joireman, Teta, & Kraft, 1993). Representative items included “I never met a person that I didn’t like,” “I always win at games,” “I have never been bored,” “I never get annoyed when people cut ahead of me in line.” Positive responses were summed to create a total score ranging from 0 to 10, with higher scores indicating greater social desirability (Kuder–Richardson Formula 20 = 0.71).

Analysis Strategy

All analyses employ sampling weights to account for the complex survey design using the svy suite of commands within the Stata v.13 software. Missing data were minimal: Only 3.8% of participants had missing data on variables of interest and there did not appear to be systematic patterns of missingness. Accordingly, analyses proceeded with complete cases.

Basic bivariate associations establish the general patterns of psychological distress by citizenship. Next, we employ multivariable linear regression with psychological distress as the dependent variable. Our modeling approach begins with the citizenship variables plus the major potential confounders: age, gender, state, Asian ethnicity, and social desirability. We next added in succession measures related to immigration, socioeconomic characteristics, health care visits, everyday discrimination, and subjective social status.

These analyses were replicated for the subsample of immigrants. This allowed for the inclusion of the acculturative stress variables that were asked only of immigrants.

We formally evaluated mediation using Sobel tests (MacKinnon, 2008). In testing for mediation, we first evaluated whether the independent variable (citizenship) was related to the mediator (e.g., subjective social status). In this evaluation, multivariable models were used to control for potential confounders related to compositional differences between citizenship strata. We also explored the possible factors that may explain differences in subjective social status using the same sequence of models that we used to study psychological distress.

We considered investigating differences among the Asian ethnic groups. However, this was not possible because the cell sizes were too small for reliable analysis. For example, there were only 19 native-born Vietnamese participants. Accordingly, our analyses control for ethnic group, but do not further test moderation by ethnicity.

Results

Table I provides the weighted characteristics of the sample, stratified by citizenship. At first glance, there were no differences in psychological distress by citizenship. The rates for distress for noncitizens, native-born, naturalized, and dual citizens were 13.6, 13.0, 13.0, and 13.8, respectively. However, there were differences in demographic factors by citizenship. For example, noncitizens were younger (37.4 years) and had fewer years in the United States (8.6 years) compared with naturalized citizens (45.4 years in age, 21.7 years in the United States). Native-born citizens also reported higher rates of everyday discrimination, higher subjective social status, and less social desirability than noncitizens. In summary, the citizenship groups did not differ with regard to psychological distress, but did show some differences in other study measures in unadjusted analyses.

Table I.

Weighted Descriptive Statistics by Citizenship. Full Sample: National Latino and Asian American Study (n = 2,015).

	Noncitizen (n = 601)			Native-born citizen (n = 454)			Naturalized citizen (n = 794)			Dual citizen (n = 166)
	Mean or %	SD	Sig.	Mean or %	SD	Sig.	Mean or %	SD	Sig.	Mean or %	SD	Sig.
Age	37.4	12.9	c, d	37.5	16.3	c	45.4	15.6	a, b	47.4	16.4	a
Female	50.3			48.9			56.7			48.0
Ethnic Subgroup
Chinese	32.0		b	15.2		a	30.4			33.6
Filipino	16.2			20.7			23.0			16.8
Vietnamese	10.2		b	1.5		a, c, d	18.6		b	13.9		b
Asian Indian	16.0			3.4			6.9			7.7
Native Hawaiian	0.3			3.9			0.0			0.0
Japanese	4.1		b	25.6		a, c	1.6		b	0.0
Korean	7.4			5.5			8.0			6.8
Other Pacific Islander	0.1			1.2			0.0			0.0
Other Asian	13.6			23.1			11.5			21.3
Social Desirability	2.6	2.1	b	1.7	1.6	a, c	2.4	2.2	b	2.2	2.0
Interview Language
English	64.6		b	99.1		a, c, d	62.6		b	69.1		b
Spanish	0.2			0.2			0			0.6
Chinese	21.8		b	0.3		a, c, d	18.9		b	19.1		b
Tagalog	3.8			0.1			3.4			2.9
Vietnamese	9.6		b	0.3		a, c	15.1		b	8.2
Years Lived in the United States	8.6	6.9	b, c, d	37.2	16.3	a, c, d	21.7	10.5	a, b	20.8	10.7	a, b
Education
11 Years or less	19.8			3.3			18.5			9.8
12 Years	14.2			22.3			17.7			14.2
13−15 Years	17.1			36.6			25.2			24.5
16+ Years	48.9			37.9			38.6			51.5
Income-To-Poverty Ratio	5.9	8.8		6.4	5.6		6.1	6.5		6.1	6.3
Employment
Unemployed	18.5			21.6			18.4			20.3
Employed	61.7			65.3			65.0			64.5
Out of labor force	19.8			13.1			16.5			15.1
Everyday Discrimination	15.9	6.7	b	18.4	6.8	a, c	15.3	6.2	b	17.4	6.2
Subjective Social Status	5.6	2.0	b	6.3	1.4	a	5.9	1.9		5.8	2.2
Health Care Visits	0.3	2.8		1.2	6.0		0.5	5.0		0.5	2.2
Psychological Distress	13.6	4.8		13.0	4.2		13.0	4.5		13.8	4.2

Note. p Value calculated using Šídák correction for multiple comparisons: a = Significantly different than noncitizens (p < .0003). b = Significantly different than U.S. born (p < .0003). c = Significantly different than naturalized citizens (p < .0003). d = Significantly different than dual citizens (p < .0003).

Table 2 provides the results of linear regression analysis with psychological distress as the outcome. Model 1 includes citizenship plus potential confounders, including age, gender, Asian ethnicity, state of residence, and social desirability. This model indicates that native-born citizens (b = −0.76, p < .05) and naturalized citizens (b = −0.59, p < .05) reported significantly less distress compared with noncitizens. Dual citizens were not significantly different from noncitizens (b = 0.18, p < .05). Wald tests showed that naturalized citizens and native-born citizens did not significantly differ from one another with respect to psychological distress in Model 1 or in subsequent models. Social desirability was associated with lower psychological distress (b = −0.21, p < .05). Model 2 includes years in the United States and language of interview. Naturalized citizens were significantly less distressed (b = −0.67, p < .05) than noncitizens. Native-born citizens showed a similar, but nonsignificant trend (b = −0.95, p > .05). Diagnostic tests did not indicate collinearity between duration in the United States and age. Model 3 adds education, employment, and the income-to-poverty ratio. As before, naturalized citizens (b = −0.68, p < .05) reported less distress than noncitizens. Native-born citizens also were significantly less distressed (b = −1.05, p < .05). Models 4 and 5 add health care and discrimination, respectively. The patterns for citizenship remain unchanged. Model 6 includes subjective social status. Here, naturalized citizens were no longer significantly different from noncitizens (b = −0.54, p > .05), although native-born citizens remained significantly less distressed (b = −1.09, p < .05). Furthermore, those who perceive greater subjective social status reported less distress (b = −0.33, p < .05).

Table 2.

Linear Regression Models of Psychological Distress on Covariates. Full Sample: National Latino and Asian American Study (n = 2,015).

	Model la	Model 2a	Model 3a	Model 4a	Model 5a	Model 6a
	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)
Citizenship Status (ref = noncitizen)
U.S. born	−0.76* (0.33)	−0.95 (0.51)	−1.05* (0.49)	−1.12* (0.48)	−1.13* (0.46)	−1.09* (0.46)
Naturalized citizen	−0.59* (0.23)	−0.67* (0.29)	−0.68* (0.29)	−0.67* (0.29)	−0.63* (0.28)	−0.54 (0.27)
Dual citizen	0.18 (0.29)	0.11 (0.31)	0.09 (0.27)	0.09 (0.27)	−0.19 (0.28)	−0.11 (0.29)
Demographics
Female	0.50 (0.31)	0.49 (0.32)	0.37 (0.30)	0.30 (0.29)	0.54* (0.22)	0.62* (0.21)
Age	−0.00 (0.01)	−0.01 (0.01)	−0.01 (0.01)	−0.01 (0.01)	0.01 (0.01)	0.00 (0.01)
Ethnicity (ref = Chinese)
Filipino	−1.01**(0.28)	−0.58 (0.47)	−0.66 (0.51)	−0.72 (0.50)	−0.78 (0.52)	−0.71 (0.52)
Vietnamese	−1.12** (0.30)	−0.64 (0.57)	−0.74 (0.59)	−0.70 (0.56)	−0.64 (0.54)	−0.74 (0.54)
Asian Indian	−1.32* (0.48)	−0.80 (0.61)	−0.78 (0.59)	−0.74 (0.56)	−0.51 (0.61)	−0.36 (0.59)
Native Hawaiian	0.67 (1.29)	1.01 (1.32)	0.95 (1.40)	1.02 (1.31)	1.34 (1.44)	1.15 (1.39)
Japanese	−0.63 (0.55)	−0.23 (0.57)	−0.31 (0.58)	−0.19 (0.57)	−0.03 (0.64)	0.02 (0.63)
Korean	0.92 (0.99)	1.35 (1.11)	1.22 (1.07)	1.31 (1.07)	1.41 (1.00)	1.38 (0.95)
Other Pacific Islander	1.04 (2.40)	1.45 (2.34)	1.07 (2.38)	1.26 (2.4)	0.33 (3.08)	0.47 (3.14)
Other Asian	−0.58 (0.72)	−0.16 (0.85)	−0.30 (0.89)	−0.22 (0.89)	−0.26 (0.88)	−0.25 (0.88)
Social Desirability	−0.21**(0.07)	−0.23** (0.07)	−0.25** (0.075)	−0.25** (0.07)	−0.21** (0.07)	−0.21**(0.07)
Interview Language (ref = English)
Spanish		2.78 (3.78)	2.50 (3.73)	2.6 (3.68)	3.15 (3.48)	3.05 (3.02)
Chinese		0.97 (0.58)	0.88 (0.62)	0.93 (0.61)	1.37* (0.63)	1.05 (0.56)
Tagalog		0.47 (0.61)	0.32 (0.66)	0.33 (0.65)	0.65 (0.57)	0.51 (0.55)
Vietnamese		0.14 (0.53)	0.00 (0.58)	−0.06 (0.57)	0.76 (0.55)	0.52 (0.57)
Years Lived in United States		0.01 (0.01)	0.01 (0.01)	0.01 (0.01)	0.00 (0.01)	0.01 (0.01)
Education (ref = less than high school degree)
High school degree			−0.08 (0.54)	−0.00 (0.53)	0.07 (0.57)	0.17 (0.56)
Some college			0.25 (0.48)	0.26 (0.48)	0.40 (0.52)	0.53 (0.50)
College degree or more			−0.29 (0.56)	−0.26 (0.55)	−0.17(0.55)	0.13 (0.58)
Income-to-Poverty Ratio			−0.03 (0.02)	−0.03 (0.02)	−0.03 (0.02)	−0.02 (0.02)
Employment (ref = unemployed)
Employed			−0.37 (0.25)	−0.33 (0.25)	−0.34 (0.25)	−0.24 (0.24)
Out of labor force			−0.25 (0.34)	−0.24 (0.32)	−0.18 (0.31)	−0.16 (0.29)
Health Care Visits				0.15* (0.06)	0.14* (0.06)	0.14* (0.06)
Everyday Discrimination					0.19*** (0.03)	0.18*** (0.03)
Subjective Social Status						−0.33*** (0.07)
Constant	15.91*** (0.42)	15.77*** (0.52)	16.31***(0.97)	16.26*** (0.96)	10.93*** (1.54)	12.45*** (1.36)

Note. Standard errors (SEs) in parentheses.

^aModels also control for state of residence (coefficients, not shown).

^*p < .05.

^**p < .01.

^***p < .001.

Table 3 provides a subanalysis that excludes native-born citizens and includes the measure of acculturative stress that was asked only of immigrants. Naturalized citizens reported less psychological distress compared with noncitizens after adjustment for age, gender, subgroup, state (Model 1), as well as after adjustment for acculturative stress, language of interview, and years in the United States (Models 2 and 3), education, income-to-poverty ratio, employment (Model 4), health care visits (Model 5), and everyday discrimination (Model 6). However, naturalized and noncitizens no longer significantly differed after inclusion of subjective social status (Model 7). In summary, the results of the immigrant subanalysis mirror the results of the full sample.

Table 3.

Linear Regression Models of Psychological Distress on Covariates. Immigrants Only: National Latino and Asian American Study (n = 1,552).

	Model la	Model 2a	Model 3a	Model 4a	Model 5a	Model 6a	Model 7a
	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)	b (SE)
Citizenship Status (ref = noncitizen)
Naturalized citizen	−0.50* (0.20)	−0.22 (0.19)	−0.64* (0.24)	−0.68** (0.25)	−0.66* (0.26)	−0.55* (0.26)	−0.46 (0.250)
Dual citizen	0.19 (0.29)	0.44 (0.32)	0.05 (0.31)	−0.00 (0.23)	0.01 (0.24)	−0.20 (0.27)	−0.13 (0.28)
Demographics
Female	0.43 (0.34)	0.47 (0.33)	0.51 (0.33)	0.40 (0.31)	0.37(0.31)	0.58* (0.25)	0.66** (0.24)
Age	0.00 (0.01)	−0.00 (0.01)	−0.02 (0.01)	−0.02 (0.01)	−0.02 (0.01)	0.00 (0.01)	0.00 (0.01)
Ethnicity (ref = Chinese)
Filipino	−1.53*** (0.26)	−1.12** (0.30)	−1.01 (0.55)	−1.18* (0.58)	−1.22* (0.58)	−1.33* (0.63)	−1.24 (0.64)
Vietnamese	−1.21*** (0.31)	−1.12** (−0.29)	−0.77 (0.68)	−0.92 (0.71)	−0.92 (0.69)	−0.82 (0.66)	−0.90 (0.66)
Asian Indian	−1.39** (0.49)	−1.13* (0.54)	−0.78 (0.69)	−0.76 (0.66)	−0.76 (0.64)	−0.47 (0.69)	−0.31 (0.68)
Native Hawaiian	9.24*** (0.59)	9.72*** (0.76)	9.88*** (0.98)	10.26*** (1.20)	10.30*** (1.19)	10.93*** (1.14)	10.67*** (1.09)
Japanese	0.84 (1.86)	0.68 (1.92)	0.64 (1.97)	0.57 (1.98)	0.63 (1.96)	0.81 (1.73)	1.04 (1.71)
Korean	0.75 (1.14)	0.81 (1.12)	0.92 (1.22)	0.71 (1.16)	0.75 (1.16)	0.99 (1.14)	0.97 (1.08)
Other Pacific Islander	4.95 (3.17)	4.72 (3.12)	5.15 (3.33)	4.89 (3.11)	4.86 (3.12)	5.11 (2.88)	5.2 (2.90)
Other Asian	−0.63 (0.78)	−0.60 (0.81)	−0.44 (0.99)	−0.57 (1.00)	−0.54 (0.99)	−0.43 (1.06)	−0.41 (1.04)
Social Desirability	−0.24** (0.08)	−0.22** (0.07)	−0.22** (0.07)	−0.25** (0.07)	−0.24** (0.07)	−0.23** (0.08)	−0.24** (0.07)
Acculturative Distress		0.42*** (0.07)	0.45*** (0.07)	0.45*** (0.07)	0.45*** (0.07)	0.26*** (0.06)	0.23** (0.06)
Interview Language (ref = English)
Spanish			4.32 (4.73)	3.45 (4.84)	3.52 (4.82)	4.72 (4.50)	4.55 (3.94)
Chinese			0.50 (0.65)	0.38 (0.68)	0.41 (0.67)	1.02 (0.75)	0.80 (0.69)
Tagalog			1.09* (0.62)	0.99 (0.70)	0.97 (0.70)	1.23 (0.61)	1.05 (0.60)
Vietnamese			0.18 (0.59)	0.07 (0.66)	0.03 (0.65)	0.85 (0.64)	0.65 (0.65)
Years Lived in United States			0.05** (0.01)	0.05** (0.01)	0.04** (0.01)	0.02 (0.01)	0.02 (0.01)
Education (ref = less than high school degree)
High school degree				0.12 (0.57)	0.16 (0.56)	0.23 (0.61)	0.31 (0.60)
Some college				0.63 (0.55)	0.61 (0.55)	0.67 (0.60)	0.79 (0.58)
College degree or more				−0.24 (0.55)	−0.21 (0.54)	−0.14 (0.56)	0.14 (0.60)
Income-to-Poverty Ratio				−0.03 (0.02)	−0.03 (0.02)	−0.03 (0.02)	−0.02 (0.01)
Employment (ref = unemployed)
Employed				−0.24 (0.28)	−0.22 (0.27)	−0.12 (0.28)	−0.02 (0.29)
Out of labor force				−0.12 (0.33)	−0.10 (0.32)	−0.04 (0.34)	0.02 (0.32)
Health care visits					0.09 (0.06)	0.10 (0.06)	0.10 (0.06)
Everyday Discrimination						0.18*** (0.03)	0.18*** (0.03)
Subjective Social Status							−0.30** (0.09)
Constant	10.45*** (0.39)	8.37*** (0.49)	8.78*** (0.39)	9.19*** (0.90)	9.16*** (0.87)	4.81*** (1.26)	6.57*** (1.16)

^aModels also control for state of residence (coefficients, not shown).

^*p < .05.

^**p < .01.

^***p < .001.

Taken together, these analyses provide preliminary evidence that subjective social status, but not socioeconomic status or discrimination may be mediators. We next formally evaluated whether subjective social status mediated the association between citizenship and psychological distress. In testing mediation, one assumes that the key explanatory variable is related to the mediator (i.e., citizenship is related to subjective social status). These additional analyses (Table 4) showed that noncitizens perceived lower social position compared with naturalized citizens, native-born citizens and dual citizens after accounting for age, gender, Asian subgroup, social desirability, and state of residence (Model 1), socioeconomic factors (Model 5), and everyday discrimination (Model 7). What appeared to account for these differences in subjective social status between citizenship categories were years in the United States and language of interview (Models 2, 3, and 8).

Table 4.

Linear Regression Models of Subjective Social Status on Covariates. Full Sample: National Latino and Asian American Study (n = 2,015).

	Model la	Model 2a	Model 3a	Model 4a	Model 5a	Model 6a	Model 7a	Model 8a
Citizenship Status (ref = noncitizen)
U.S. born	0.55*** (0.13)	0.32* (0.12)	−0.08 (0.25)	0.01 (0.25)	0.57*** (0.14)	0.12 (0.22)	0.57*** (0.13)	0.12 (0.22)
Naturalized citizen	0.47** (0.15)	0.39* (0.15)	0.23 (0.16)	0.27 (0.16)	0.43** (0.14)	0.27 (0.15)	0.47** (0.15)	0.26 (0.15)
Dual citizen	0.47** (0.14)	0.35* (0.15)	0.27 (0.15)	0.25 (0.17)	0.36** (0.12)	0.21 (0.15)	0.50** (0.13)	0.24 (0.15)
Demographics
Female	0.05 (0.10)	0.11 (0.10)	0.08 (0.10)	0.12 (0.10)	0.19 (0.09)	0.24* (0.09)	0.03 (0.10)	0.21*(0.09)
Age	−0.01** (0.00)	−0.00 (0.00)	−0.02** (0.01)	−0.01 (0.01)	−0.01 (0.00)	−0.01 (0.01)	−0.01** (0.00)	−0.01 (0.01)
Ethnicity (ref = Chinese)
Filipino	0.59** (0.16)	0.08 (0.15)	0.60*** (0.15)	0.12 (0.15)	0.62*** (0.14)	0.22 (0.16)	0.61*** (0.15)	0.22 (0.16)
Vietnamese	−0.63** (0.23)	−0.41 (0.24)	−0.50* (0.21)	−0.38 (0.24)	−0.31 (0.22)	−0.28 (0.20)	−0.66** (0.22)	−0.29 (0.21)
Asian Indian	1.15*** (0.21)	0.52* (0.21)	1.18*** (0.21)	0.58* (0.22)	0.93*** (0.19)	0.50* (0.21)	1.15*** (0.21)	0.48* (0.21)
Native Hawaiian	−0.05 (0.22)	−0.58* (0.28)	−0.18 (0.27)	−0.60 (0.31)	−0.11 (0.23)	−0.55 (0.30)	−0.06 (0.24)	−0.59 (0.33)
Japanese	0.77*** (0.19)	0.24 (0.16)	0.61** (0.17)	0.19 (0.16)	0.65*** (0.16)	0.19 (0.15)	0.77*** (0.18)	0.18 (0.16)
Korean	0.36 (0.24)	−0.22 (0.24)	0.28 (0.22)	−0.22 (0.24)	0.38 (0.24)	−0.09 (0.24)	0.37 (0.23)	−0.10(0.24)
Other Pacific Islander	0.31 (0.31)	−0.11 (0.33)	0.39 (0.33)	−0.04 (0.35)	0.64* (0.27)	0.34 (0.31)	0.38 (0.28)	0.45 (0.23)
Other Asian	0.35** (0.11)	−0.18 (0.10)	0.32** (0.11)	−0.16 (0.10)	0.45*** (0.10)	0.03 (0.09)	0.36** (0.11)	0.03 (0.09)
Social Desirability	−0.09** (0.03)	−0.06* (0.02)	−0.08** (0.02)	−0.05* (0.02)	−0.03 (0.02)	−0.01 (0.02)	−0.09** (0.03)	−0.02 (0.02)
Interview Language (ref = English)
Spanish		−0.34 (1.61)		−0.40 (1.53)		−0.24 (1.47)		−0.30 (1.45)
Chinese		−1.23*** (0.22)		−1.14*** (0.21)		−0.92*** (0.19)		−0.97***(0.19)
Tagalog		−0.89** (0.32)		−0.78* (0.35)		−0.41 (0.35)		−0.45 (0.36)
Vietnamese		−1.14** (0.37)		−1.02** (0.35)		−0.65 (0.33)		−0.74* (0.33)
Years Lived in United States			0.03** (0.01)	0.01 (0.01)		0.01 (0.01)		0.01 (0.01)
Education (ref = 11 years or less)
12 Years					0.43 (0.23)	0.33 (0.20)		0.32 (0.21)
13−15 Years					0.59** (0.23)	0.42* (0.20)		0.41 (0.20)
16+ Years					1.13*** (0.17)	0.94*** (0.16)		0.93*** (0.16)
Income-to-Poverty Ratio					0.03*** (0.01)	0.03*** (0.01)		0.03*** (0.01)
Employment (ref = unemployed)
Employed					0.23* (0.11)	0.30** (0.10)		0.30** (0.10)
Out of labor force					0.02 (0.16)	0.08 (0.15)		0.07 (0.15)
Everyday Discrimination							−0.01 (0.01)	−0.02** (0.01)
Constant	4.54*** (0.17)	4.90*** (0.17)	4.85*** (0.20)	5.03*** (0.19)	3.53*** (0.27)	4.03*** (0.26)	4.86*** (0.30)	4.64*** (0.32)

^aModels also control for state of residence (coefficients, not shown).

^*p < .05.

^**p < .01.

^***p < .001.

We then tested whether the association between citizenship and psychological distress was mediated by subjective social status. The data showed a marginally significant trend for mediation (Sobel statistic, t = −1.66, standard error [SE] = 0.05, p = .10), suggesting a total effect of citizenship of −0.63 units (SE = 0.28). This total effect corresponds to a direct effect of −0.54 units (SE = 0.27) and an indirect effect of 0.09 units (SE = 0.05). This indirect (mediated) effect accounts for 13.8% of the total effect, and the ratio of the total to the direct effect is 1.16.

As a sensitivity check, models for psychological distress were also estimated using negative binomial regression because psychological distress was highly skewed and had a large number of zero responses (Poisson was not used because of overdispersion). To estimate mediation with negative binomial models, we used the method proposed by Karlson, Holm, and Breen (2010). This method accounts for the rescaling of coefficients across nested models with limited dependent variables using the Stata command khb. This methodology was developed mainly for binary nonlinear probability models (i.e., logit and probit models). The estimate of the percent mediated effect using the khb method was almost identical to the estimate calculated using the product of the coefficients method from our linear regression models (14.2% compared with 13.8%, respectively). Similarly, the ratios of total to direct effects were comparable (1.17 and 1.16, respectively). The linear regression results are reported here because they are easier to interpret than the negative binomial regression results.

Discussion

This study provides novel information showing that citizens report less psychological distress compared with noncitizens after accounting for acculturative stress, health care visits, and a variety of other sociodemographic factors. Hence, the data are consistent with the first hypothesis. These findings are consistent with other research that shows citizens report better self-rated health and have greater access to health care than noncitizens (Bustamante & Chen, 2012; Gubernskaya et al., 2013; Yu et al., 2004). The data are also consistent with broader studies of structural inequities, which show that groups with higher social status enjoy lower morbidity (Adler & Stewart, 2010; Link & Phelan, 1995). Our second hypothesis, that naturalized and native-born citizens would differ from one another with respect to psychological distress, was not supported by the findings. Thus, the greatest contrasts with regard to distress were between citizens and noncitizens.

Although psychological distress was similar across all citizenship groups in unadjusted analyses, disparities emerged when we controlled for potential confounders such as social desirability. Indeed, social desirability was related to less psychological distress, a finding that may be related to other literature that shows that many Asian Americans try to “save face” and avoid embarrassment by underreporting their problems (Gong, Gage, & Tacata, 2003; Zane & Yeh, 2002). Hence, it was important to account for such potential reporting biases in the present study.

Our analysis provides some initial clues regarding the disparity between naturalized citizens and noncitizens. First, it is notable that differences by citizenship status remained after controlling for discrimination and socioeconomic characteristics such as education and employment status. If citizenship is purely an aspect of structural inequality, we would have expected smaller differences in psychological distress by citizenship category after including socioeconomic characteristics in our model, but the data did not support this interpretation. Thus, our third hypothesis was not supported. However, it would be premature to conclude that these more structural indicators do not matter. One important factor related to socioeconomic circumstances is related to visa and documentation status. Many Asian noncitizens are in the United States on work and student visas, and indeed, immigrants are preferentially selected for entry into the United States based on their occupational skills (National Academies of Sciences, Engineering and Medicine, 2015). Additionally, some Asians are also undocumented immigrants and others are refugees. We did not have the data to study this complexity between socioeconomic circumstances, visa and documentation status, but these issues are important considerations for future research.

In contrast, aspects of citizenship as social identity seem to play a role in explaining differences in psychological distress, as per our fourth hypothesis. In multivariable models, reports of subjective social status were lowest among noncitizens, and it was the only factor that explained differences in psychological distress between noncitizens and naturalized citizens. It is also interesting to note that despite the higher educational status of noncitizens, they reported the lowest social position. Tests of mediation provided tentative evidence that subjective social status partially mediated the disparity between noncitizens and naturalized citizens. We caution that mediation was only marginally statistically significant and the findings deserve verification in future research. If our findings are replicable and durable, they suggest that perceiving oneself to be of higher social status may explain some of the disparities in psychological distress between noncitizens and naturalized citizens.

Taken together, these findings broadly speak to the possible role of citizenship as a form of aspirational identity (Frye, 2012; Thornborrow & Brown, 2009). Although this study was not a full test of the propositions of identity control theory (Burke, 1991; Stets & Burke, 2005), the findings are consistent with the general claims. Plausibly, immigrants attach meanings to their identity as a citizen or noncitizen, and these meanings are then used as a standard of comparison. Immigrants whose identities compare favorably to these standards would experience positive emotions, whereas immigrants whose identities do not compare to these standards would experience negative emotions. In the present case, it is possible that the identity of being a naturalized citizen facilitates positive comparisons to a standard of climbing higher on the social ladder, which then reduces negative emotions such as psychological distress. Identity control theory generally focuses on the discrepancy between one’s perceived identity and their idealized standards for this identity. Our findings extend these ideas further by suggest that one’s aspirations for a new social identity, when unrealized, may be distressing. We did not have direct measures of aspirations, however, and it would be important to measure aspirations directly in future research.

Subjective social status appeared to account for differences between naturalized citizens and noncitizens, but none of our measures explained differences between native-born citizens and noncitizens. Speculatively, the latter differences might be related to social networks and residence in ethnic enclaves. For example, a recent study noted that among recently arrived immigrants, living in neighborhoods characterized by linguistic isolation and low collective efficacy were risk factors for depressive symptoms (Vega, Ang, Rodriguez, & Finch, 2011). This suggests that neighborhood conditions may play a role in explaining disparities by citizenship and may be a fruitful area of investigation.

Several additional caveats should be noted. The data are cross-sectional, and therefore, the findings should not be viewed as causal. Although our theory suggests that citizenship contributes to positive mental health, it is nonetheless plausible that psychological symptoms make some immigrants less likely to become naturalized. Future research should verify these findings longitudinally, including data before and after an immigrant obtained citizenship (Gee, de Castro, Wang, Crespi, & Morey, 2015). Such a time-series design would allow for a firmer assessment regarding the temporal ordering of our focal variables. Furthermore, we did not have sufficient samples of ethnic subgroups to conduct disaggregated analyses, although it would be useful to do so in future research because of within-group heterogeneity (Srinivasan & Guillermo, 2000). Also, as noted earlier, our analyses exclude Latinos because there are special considerations that make them different from Asians (e.g., Puerto Rican citizenship). We plan to study these issues among Latinos in more detail in future research. Finally, the data come from 2002 to 2003 and it is unclear to what extent these findings generalize to the present time. The NLAAS is the only nationally representative study with large samples of Asian Americans that also includes our key covariates (e.g., citizenship, subjective social status, discrimination, distress). Nonetheless, it would be important to replicate and extend our analysis with more recent data.

Many other research questions arise from the present findings. For example, it would be useful to consider whether the relationship between citizenship and health outcomes depends on immigration cohort. Similarly, this relationship might further vary by the age at which someone obtains citizenship (Gubernskaya et al., 2013). For instance, someone obtaining citizenship at a younger age would have a longer period from which to reap the social and economic advantages of citizenship, suggesting a stronger effect of citizenship among those who became citizens at a younger age compared with those who gain citizenship at older ages. Thus, although socioeconomic factors did not explain the citizenship disparities presently, such factors might be particularly important among persons who became citizens at younger ages.

In closing, the present study provides preliminary evidence of disparities in psychological distress by citizenship among a nationally representative sample of Asian Americans. Some of these disparities may be related to immigrants’ perceptions of climbing the social ladder and achieving, or not achieving, their “American dream.”

Author Biographies

Gilbert C. Gee, Ph.D. is a Professor in the Department of Community Health Sciences at the University of California, Los Angeles. His research focuses on the social determinants of health inequities among immigrant and racial/ethnic minority populations.

Brittany N. Morey is a Ph.D. candidate at the UCLA Fielding School of Public Health. Her research examines health disparities and environmental justice, especially among Asian Americans, Native Hawaiians, and Pacific Islanders.

Katrina M. Walsemann is an associate professor of health promotion, behavior, and education at the University of South Carolina. Her research examines how social inequities influence health and racial health disparities across the life course.

Annie Ro is an Assistant Professor in the UC Irvine Program in Public Health. She is a health demographer who studies the social determinants of immigrant health.

David T. Takeuchi is the Dorothy Book Scholar and Associate Dean for Research at Boston College. His research focuses on the social, structural, and cultural contexts that are associated with different health outcomes, especial among racial and ethnic minorities.