Abstract
This paper invokes a membership-exclusion theoretical model of immigrant integration to investigate political incorporation. Specifically, we examine the extent to which unauthorized migration status is associated with general and particular political knowledge and with other kinds of structural incorporation. In the analyses, we use data from the initial wave of the 2012 Latino Immigrant National Election Study (LINES) targeting adult immigrants from Spanish-speaking countries in Latin America. Consistent with theoretical expectations, we find that unauthorized Latino immigrants have significantly lower levels of general political knowledge than green card holders, those with other government IDs, or naturalized citizens, and that the difference between the unauthorized and the legal groups holds up when controls are introduced for exposure (quantity and quality of time in the country) and various kinds of structural incorporation, although differences among the legal groups do not. Thus, forms of structural integration mediate the effects of exposure on acquisition of general political knowledge by legal immigrants, but they do not for unauthorized immigrants, providing evidence that membership exclusion severely restricts political incorporation. At the same time, unauthorized immigrants show more awareness about changes in the unemployment rate than legal immigrants do, a result consistent both with their main reason for migration (to work) and with their having recourse only to collective action as a form of political expression.
Many discussions of immigrant political incorporation begin by focusing on race and ethnicity (Hochschild et al. 2013). In the U.S. case, this seems appropriate, given that most of the migration to the United States in the last half-century has been non-European (Bean, Lee, and Bachmeier 2013). Such discussions tend to veer toward either the optimistic or pessimistic, depending on whether they take as their model the successful integration of the descendants of European immigrants of the early twentieth century or the lingering socioeconomic disadvantage of African Americans in the long wake of slavery and in the subsequent injustices of Jim Crow laws (Warner and Srole 1945; Gordon 1964; Jaynes and Williams 1989; Bean and Bell-Rose 1999; Alba and Nee 2003; Holzer 2009). Yet neither narrative seems fully to apply to new nonwhite immigrants, especially Latino newcomers, who tend not to see themselves in monoracial terms (Lee and Bean 2010; Telles 2014). In part to move beyond such conceptualizations, Michael Jones-Correa (1998, 2007) recommends that students of Latino politics focus less on ethnic or transnational dynamics and more on immigrant politics, defined as “a liminal state of disengagement from the politics of both sending or receiving countries, even if not from politics entirely” (2007, 53–54).
Because so many Latino immigrants to the United States come as unauthorized entrants (Bean et al. 2014), their migration status needs to be taken into account explicitly in models of political incorporation. Political scientists have widely noted that the nature of initial societal membership among immigrants may matter for political incorporation (Hochschild and Mollenkopf 2009; Mollenkopf 2013). Similarly, sociologists have argued that the initial lack of legitimate and official societal membership has unleashed forces that are highly exclusionary but potentially impermanent (Bean and Brown 2015). Thus, in addition to studies of political integration as a function of individual background and experience on factors affecting voting and civic participation, we also need to assess the effects of local organizations and institutions, such as those with implications for settlement or citizenship, as well as the effects of more distal bureaucracies that establish policies and priorities and carry out enforcement (Bloemraad 2013; Ramakrishnan 2013; Jones-Correa 2013).
This paper looks at the latter kind of factors. The lack of political integration, in particular, starts with the absence of initial societal membership, together with the organizational and institutional factors that define and sustain this lack, in a process called membership exclusion (Bean, Brown, and Bachmeier 2015). This approach argues that the official denial of societal membership through organizational and institutional means curtails individual-level structural integration. That is, initial societal membership constitutes a necessary, if not sufficient, condition for immigrant integration, including individual-level political integration. This paper assesses this idea by investigating the extent to which unauthorized status among immigrants is associated with both general and particular forms of political knowledge, as well as with other kinds of structural integration, the dimensions of which tend to reinforce both one another and political knowledge. Thus, we hypothesize that gradations of legalization and citizenship will relate positively both to general political knowledge and to other kinds of integration. We further hypothesize that being unauthorized will be associated with low general political knowledge and the absence of mutually reinforcing forms of structural integration, although we also suggest why unauthorized status may positively be associated with particular knowledge. Empirically, we examine how unauthorized status among Latino immigrants on the one hand and different kinds of legal and citizenship status on the other separately affect general and particular political knowledge and relationships among aspects of structural integration.
IMMIGRANT-GROUP POLITICAL INCORPORATION
Thinking about the dimensionality of integration helps us understand the implications of the membership-exclusion approach for political integration. The classic assimilation perspective and variants that emphasize ethnoracial status (see Brown and Bean 2006; Bean and Brown 2015) all share the idea that different aspects of integration tend to progress together over time, even if somewhat haltingly. They thus tend to be one-dimensional (Alba and Nee 2003). More pluralist frameworks, as exemplified by European multicultural approaches, emphasize that the sociocultural aspects of integration (compared with the economic, spatial, and political aspects) may occur at different times or not at all (Fokkema and De Haas 2011; Kymlicka 1995; Montserrat and Rex 2010). Thus, the retention of specific ethnic-group values, customs, and behavioral practices is not foreclosed. Moreover, sociocultural distinctiveness is envisioned as not incompatible with other kinds of integration. Such perspectives are more multidimensional in conceptualization because they allow immigrant groups to maintain religious and family practices and other ethnic values and behaviors that are distinct from those of native mainstream groups and from those of other ethnic immigrant groups (Modood 2007; Reitz et al. 2009; Wright and Bloemraad 2012), thus fostering social diversity. Because these sociocultural practices remain distinct from other forms of structural integration, overall incorporation embodies at least two dimensions.
Exclusionary dynamics are another example of multidimensional incorporation. Political scientists often distinguish between entry dynamics (meaning kind of initial membership) and subsequent forms of political incorporation. Jennifer Hochschild and John Mollenkopf (2009) and Hochschild and her colleagues (2013) present conceptual schema that underscore how different kinds of political integration take place at different points in the overall process, especially the difference between entry and later attitudes and behaviors. They note that U.S. immigrants cannot vote until they naturalize, meaning that voting behavior cannot become a relevant aspect of political incorporation until immigrants have lived in the country long enough to become eligible to apply for citizenship. This sequence alone implies that prior membership is important for other integration outcomes. Just as one cannot vote without becoming a citizen, neither can newcomers apply for citizenship without having become legal permanent residents. The lack of social citizenship for unauthorized immigrants reaches its extreme form when such migrants are officially excluded from many arenas of social and economic participation (by dint of policy and law), so that other forms of incorporation are effectively shut off. This foundational principle undergirds the theoretical perspective of membership exclusion (Bean, Brown, and Bachmeier 2015).
Legal status provides access to opportunity (and thus social mobility), as well as partial access to some degree of a social safety net. In short, the lack of initial societal membership implies an important precondition for other forms of integration. Official boundaries that demarcate ineligibility for social citizenship in a given society not only cut off access to opportunity, they also foster stigmatization and exploitation that further hamper structural advancement (Joppke 2010; Hochschild et al. 2013; Hochschild and Mollenkopf 2009; Zolberg 1999). The nature of such exclusionary boundaries can vary from country to country, involving variously, for example, religion, colonial origins, race, residency restrictions, or citizenship (Freeman 1979; Fredrickson 1988; Pickus 1998; Solinger 1999; Alba and Silberman 2002; Bauer et al. 2005; Papademetrios 2006; Sokatch and Myers 2014). Whatever the particular basis for a strong exclusionary boundary, the presence and status of migrants falling outside the boundary results in their being seen and treated as illegitimate. This is even more strongly the case when such boundaries are institutionalized in formal law.
In the United States historically, previous instances of notable denials of social citizenship include the enslavement of African Americans in the South, Mexican Americans in Texas through the first half of the twentieth century, the forced relocation of Native Americans in the Trail of Tears, and Chinese exclusion (Lieberson 1980; Montejano 1987; Foley 1997; Lee 2002; Perdue and Green 2007; Rumbaut 2006). All of these groups were either dislocated or subjected to apartheid-like exclusions that were not officially eliminated until the civil rights legislation of the mid-1960s (Nee and Holbrow 2013). Immigrants incurring exclusions today, like these ethnoracial minorities earlier, experience social and psychological distress (Gonzales 2011; Nee and Holbrow 2013; Yoshikawa 2011), strain and tension stemming from nonmembership dampening forms of immigrant structural integration, such as educational attainment (Bean et al. 2015).
Unauthorized migrant status has constituted an ever stronger exclusionary boundary in the United States over the past three decades, since the 1986 Immigration and Reform Act criminalized the hiring of unauthorized workers (Bean, Vernez, and Keely 1989). More recent laws deemed that migrants themselves were committing illegal acts by working, thus deepening the marginalization of the Latinos who make up the bulk of the U.S. unauthorized population (Massey 2013; Tienda and Sánchez 2013; Motomura 2014). Because so many Latinos remain in marginal statuses for long periods, such legal exclusions, including the inability to get a driver's license and ineligibility for many kinds of jobs, limit individual economic mobility. Burdens also extend to the children of such immigrants, undercutting their access to higher education, heightening stress, undermining motivations to achieve, and slowing cognitive and emotional development (Bean et al. 2011; Nee and Holbrow 2013; Yoshikawa 2011). A membership-exclusion hypothesis explicitly specifies that other forms of integration, including later political incorporation, are unlikely to any appreciable degree without this earliest form of political incorporation. In this sense, membership exclusion explicitly involves a multidimensional conceptualization of integration because it posits a sharp integration divide between those with membership, for whom aspects of integration proceed in a relatively unhindered fashion, and those without such status, whose integration does not.
Unauthorized Status and Political Knowledge
Here we study political integration by focusing on both general and particular forms of political knowledge. By general political knowledge, we mean information regarding how much Latino immigrants know about certain key civic features of American national political institutions. By particular political knowledge, we mean their knowledge of a particular feature of the American economy that we would expect to be especially salient to labor migrants, who come to the United States to work, often expecting not to stay. This category applies almost universally to unauthorized Latino migrants (Bean et al. 2015). The two specific forms of particular knowledge we examine are how accurately they follow the national unemployment rate and how accurately they perceive changes in unemployment over the past year.
The latter measures are political economy variables because awareness of unemployment conditions reflects knowledge about the socioeconomic context within which political integration occurs but does not tap general civics knowledge per se. Rather, it shows awareness of the strength of the labor market and its changes resulting from government policies. Although this is different from civic knowledge, it may be particularly useful given that research shows voters often cast their ballots based on broad economic conditions, especially job availability (Delli Carpini and Keeter 1996). For unauthorized migrants, it may indicate their potential for collective political action, given that other forms of political involvement are unavailable. We thus use data on three indicators—a measure of general political knowledge and two measures of particular awareness of unemployment. These come from a recent survey of Latino immigrants, one carried out both just before and after the 2012 national U.S. elections. Specifically, we examine data on Latino immigrants’ knowledge of the maximum number of terms a U.S. president can serve and of the length of a single term for those elected to the U.S. Senate, as well as the accuracy of perceptions about the national unemployment rate and its recent change. General political knowledge reflects broad familiarity with the American political system and relates to wider awareness of and participation in U.S. politics more generally (Galston 2001), thus providing a basis for gauging Latino immigrant political integration.
We do not assess other general political behaviors or perceptions and feelings about American politics. Thus we do not try to examine phenomena like voting because so many Latino immigrants, because they are not citizens, are ineligible to vote. Nor do we examine subjective feelings about politics because these may be appreciably affected by many Latinos’ not being citizens and thus unable to vote. Our examination of knowledge about changes in unemployment represents an effort to tap into a secondary indicator of more concrete politically relevant knowledge that may not relate to eligibility to vote per se but reflects knowledge of economic conditions related to the reasons for migration. Although unauthorized labor migrants may be excluded from structural opportunities that would provide incentives to acquire general political civics knowledge, they may be quite sensitive to employment opportunities. In fact, research has consistently shown that the size of the flow of unauthorized migrants to the United States closely tracks the U.S. unemployment rate (Council on Foreign Relations 2009; Hanson 2006). Unauthorized migrants’ politically relevant knowledge about this aspect of the economy may thus actually exceed that of legal migrants, whose positions are more secure and motives more varied. With this in mind, we thus carry out assessments of Latino immigrant political incorporation by examining how both general knowledge about American civics and awareness about important particular features of the national economic context vary with migration and citizenship status, length of time spent in the country, and key aspects of structural integration.
Why is this likely to be fruitful, given that native-born citizens have been found to have unusually low levels of political and economic knowledge? For example, only a third of the U.S. population in 2014 could approximately identify the unemployment rate (Pew Research Center 2014). Moreover, such knowledge has remained low for decades, even as levels of education have risen (Galston 2001). Although political scientists continue to discuss the implications of low levels of civic knowledge for politics in the United States, research indicates both the quality and quantity of political participation is associated with greater civic knowledge. In part because greater civic knowledge may result from more participation, it is often argued that low levels of particular civic knowledge do not preclude active and constructive political participation based on general and diffuse knowledge (Galston 2001; Popkin and Dimock 1999). Moreover, among the respondents in the 2012 Latino Immigrant National Election Study (LINES), who were foreign born, even though only 10.5 percent could correctly report the length of term served by U.S. senators, this is not much different from the responses of Latino voters overall (23.4 percent) or of African American voters (23.6 percent).1 It thus seems reasonable to examine both how and how much general and particular political knowledge varies with migration status and indicators of structural integration.
Research Hypotheses
A membership-exclusion conceptualization of integration implies that not only will unauthorized status be associated with distinctly low levels of political knowledge, but also other kinds of integration will tend to be low and not related to political knowledge. Moreover, with structural integration suppressed, these other kinds of integration will often not be strongly connected with each other either. Membership thus constitutes a necessary condition for other integration processes to occur. Given this, we focus first on explaining variations in general political knowledge. Because the simple passage of time brings some increase in familiarity with the new society, political knowledge may vary positively with the length of time Latino immigrants (authorized and unauthorized) have lived in the United States. However, exposure without membership is less likely to open structural doors to opportunity.
We expect unauthorized migrants to show less general political knowledge than legal migrants (who have been green card holders for varying lengths of time) and naturalized citizens, but higher levels of the particular knowledge of the unemployment rate. This is likely if for no other reason than unauthorized migrants are more cut off from social and economic opportunities and cannot advance as rapidly in ways that might themselves increase political knowledge, but at the same time are more sensitive to labor market fluctuations. In short, we expect unauthorized migrants to show exceptionally low levels of general political knowledge, even after we control for differences between them and legal migrants in other factors that might be expected to boost political knowledge (such as exposure and life-course measures and indicators of economic and linguistic integration). Among legal migrants, however, we expect to find that both gradations of legalization and citizenship and higher levels of other kinds of integration will be positively related to political knowledge. Stated differently, aspects of structural integration among legal but not unauthorized migrants will be more likely to mediate relationships between exposure and life-course dynamics in the new society on the one hand and the acquisition of political knowledge on the other.
DATA, MEASURES, AND STRATEGY OF ANALYSIS
To assess these ideas, we use data from the LINES. This survey was conducted in 2012 in two waves, one in the month leading up to the election in November, and one in the six weeks after the election. Through landline and cell phone interviews, the survey targeted adult immigrants from Spanish-speaking countries in Latin America, resulting in a sample size of 855 in the preelection wave. The postelection wave involved interviews with 435 respondents from the first wave and 451 new ones, for a total of 886. Most of the interviews were conducted in Spanish, and much of the survey instrument was identical to that of the 2012 American National Election Study. The bulk of our analysis comes from the preelection wave, but the question on knowledge of the actual unemployment rate comes from the postelection wave.
We measure specific political knowledge by answers to two civics questions and two broad general economic and political knowledge questions. The latter asked about the national level of unemployment and how much it had recently changed. The former asked about how many terms a president of the United States can serve and about the length of terms U.S senators serve. In all, 56.6 percent of the respondents provided a correct answer to the presidential term question, but only 10.5 percent did to the Senate term question. Only 4.9 percent answered both questions correctly, so we coded civics knowledge as a dichotomous variable based on answering at least one question correctly.2 In regard to knowledge about change in the unemployment rate, the respondents were queried as to whether they thought the national unemployment rate had gotten better, gotten worse, or stayed the same over the previous year. Because the unemployment rate fell by 1 percentage point between October 2011 and October 2012, from 8.8 percent to 7.8 percent (Bureau of Labor Statistics 2014), we again constructed a dichotomous variable, coding as accurate those who said that unemployment had improved versus all other answers. In this sample, 35.2 percent of respondents correctly indicated that unemployment had gotten better. On the question about exact knowledge of the unemployment rate, we accepted answers from 5 to 10 percent as knowledgeable. In all, 9.5 percent of the sample knew the approximate unemployment rate.
We measure migration status based on information obtained in the LINES survey in response to several questions. These include a question about naturalization, which was asked of everyone; a question asked of those who said no to the naturalization question about whether they “came with papers”; a question asked of those who said they did come with papers about whether “they had a green card now”; a question of those who did not come with papers about whether they “had a valid driver's license” (this question was also asked of those who did not answer the naturalization or the “came with papers” questions); and a question of those who did not have a valid driver's license about whether they had some other form of valid government picture ID. Answers to the questions in this “decision tree” enabled all immigrants to be classified as either naturalized citizens, legal permanent residents (LPRs, often called green card holders), driver's license or government picture ID immigrants, or unauthorized immigrants. We assume the driver's license immigrants have valid visas because in 2012 every state except New Mexico and Washington required verification to obtain such a license; Utah offered a driving privilege card (Wang 2013). This category also includes immigrants who say they possess a valid government-issued picture identification (see figure 1).
Categorization of Migration or Citizenship Status
As expected, the civics knowledge score (percentage knowing the answer to at least one question) for naturalized citizens (71.5 percent) exceeds those for all of the other migration status groups (table 1), green card (60.5 percent) and driver's license holders (63.3 percent) showing similar percentages. Most significantly, only about half as many of the unauthorized migrants (35.5 percent) could answer at least one of the questions correctly. On accuracy of knowledge of the change in the unemployment rate over the past year, however, unauthorized immigrants show a higher percentage (40.2) than any of the legal groups (33.8 percent combined). This difference for an indicator of politically relevant labor market knowledge is consistent with the idea that, although immigrants may be subject to membership-exclusion forces that cut off their access to many sources of structural opportunity and general knowledge about civics, their awareness of those features of their political or economic context most salient to their reasons for migrating (such as the change in the unemployment rate) can exceed that of legal immigrants. This is consistent with other research on specific awareness among unauthorized Mexicans of political issues directly relevant to their lives. For example, Carole Uhlaner (1996) finds that in 1994, 98 percent of Mexican immigrants lacking citizenship, versus 89 percent of Mexican American citizens, knew about Proposition 187, the ballot initiative in California to bar unauthorized migrants from access to social services.
Political Knowledge, U.S. Latino Immigrants, 2012
We also include three other kinds of variables in the statistical models predicting political knowledge. The first set consists of indicators of socio-structural incorporation—education (as measured by years of schooling completed), household income (as measured annually in thousands of dollars), and preference for speaking English at home (scored on a 6-point scale). These variables indicate levels of structural incorporation, specifically economic incorporation. Based on the membership-exclusion theoretical considerations introduced earlier, we expect these to covary among LPR and citizen immigrants, both with each other and with political incorporation, in the cases of general and particular indicators of political knowledge (that is, by civics knowledge and by the accuracy of knowledge of change in the unemployment rate).
The second set of variables consists of two temporal exposure measures. One is the length of time the immigrant has been in the country. According to classic assimilation theory, the greater the exposure, the greater the political knowledge, both specific and general (Alba and Nee 2003). The second measure is youthful age of arrival. Viewed through a life-course lens, the younger a person is at arrival, the stronger the effect of exposure and the greater the integration (Gubernskaya, Bean, and Van Hook 2013). To tap this aspect of integration, we looked at seventy-five minus age of arrival.3 Again, membership-exclusion theoretical expectations would expect these variables to vary positively with political knowledge and with structural indicators of incorporation among LPR and naturalized citizen immigrants but not unauthorized migrants. The third indicator is gender, which we simply include as a control variable. Although women show similar or lower levels of political knowledge than men, these do not vary by migration status.
The core of the membership exclusion hypothesis is that unauthorized migration status precludes (or at least severely limits) other kinds of incorporation except sociocultural facets. Accordingly, we expect unauthorized immigrants to have much less political civics knowledge than legal and naturalized immigrants, and for relationships among indicators of integration for the unauthorized to be weak and not explain much variation in political knowledge. In other words, with the possible exception of simple duration, which constitutes a strictly temporal measure of exposure, we expect other indicators—education, income, English, and life-course accentuation—not to matter much for political knowledge among unauthorized migrants. Thus, in the analyses, we first examine the zero-order and adjusted effects of migration status on the political knowledge indicators for all migrants. We then run a series of regression models separately for the group we call legal immigrants (ID holders, LPRs, and naturalized citizens) on the one hand and for unauthorized migrants on the other to ascertain whether structural and exposure factors explain knowledge variation more so among those with legal standing than among those without it. The differences in the patterns implied by the hypotheses are shown in figure 2.
Hypothesized Relationships Between Exposure and Political Knowledge
EMPIRICAL FINDINGS
We have already seen that civics political knowledge for the unauthorized is disproportionately below that for legal migrants (see table 1). Only 35.5 percent of the unauthorized correctly answer either of the two civics knowledge questions (the maximum number of terms a president can serve and the length of Senate term). Also, unauthorized immigrants fall notably below the other groups in how long they have been in the country, years of schooling, income, and English preference (see table 2). For example, unauthorized migrants report only 9.3 years of schooling on average compared with 11.1 years for the naturalized, and annual household incomes below $20,000 compared with the naturalized, whose household incomes are almost 60 percent higher ($31,400). However, in regard to knowledge of change in unemployment, the unauthorized are more aware of current unemployment trends than the other groups (40.2 percent knew the rate had dropped versus 33.8 percent for the legal and naturalized immigrants combined). As we noted earlier, knowledge of changes in the unemployment rate, though plausibly a particularly politically relevant indicator that taps into political incorporation, seems especially salient for unauthorized Mexican migrants, who are quintessential labor migrants (Bean and Stevens 2003; Bachmeier and Bean 2011). Thus, better knowledge of unemployment in their case may mostly reflect heightened sensitivity to labor market conditions, a not surprising tendency. Our conjecture that knowledge of unemployment fluctuations might vary in a direction opposite to their knowledge of civics is thus borne out. If such knowledge reflects unauthorized status and perhaps also potential for collective political action, we would not expect this indicator to necessarily show much relationship with other variables.
Means for Exposure, Structural Variables, and Gender
Legal and unauthorized immigrants differ in temporal exposure and other indicators of structural integration. We can see this clearly in table 3. For example, the group we term legals has more years of schooling (10.3 on average) than unauthorized migrants (9.3 on average). They also report greater annual household income, longer durations in the United States, but not younger ages at arrival. In addition, they also score higher on preferences for speaking English at home (or for those in the postelection wave of the survey, greater comfort in speaking English). We also present in table 3 the intercorrelations for legal immigrants compared with those for unauthorized immigrants. When these are examined separately (in the bottom two panels), we see that aspects of structural integration (education, income, and preference for English at home for legal immigrants) show tendencies to relate positively and significantly to one another and to political civics knowledge (middle panel). That is, they tend to reflect positive integration dynamics. They also relate positively to exposure measures, which is also consistent with integration dynamics. For unauthorized immigrants, however, hardly any correlations among indicators are significant, mostly those that involve exposure measures. In short, the exclusion dynamics impinging on unauthorized immigrants limit their progress to such a degree across all values of these variables that indicators of exposure, structural integration, and political knowledge do not interconnect much with one another.
Means, Standard Deviations, and Intercorrelations for Variables
We next report differences in civics knowledge across all of the migration status categories (unauthorized immigrants, ID immigrants, LPRs, and the naturalized) by successively estimating logistic regressions of political knowledge on the categories (with a dummy variable for ID immigrants omitted from the equation, making this group the reference category), followed by models including sets of selected control variables (see table 4). Without adjusting for other factors (column 1), the migration status categories show the expected direction and magnitude of relationships with political civics knowledge. For example, naturalized Latino immigrants are 46 percent more likely than ID immigrants (as indicated by an odds ratio of 1.46) to know either the number or the length of presidential or Senate terms, ID and green card holders not being significantly different from one another. Unauthorized migrants, however, are 68 percent less likely than ID holders to possess such knowledge, a highly statistically significant difference. In short, consistent with the membership-exclusion hypothesis, unauthorized immigrants are disproportionately less likely to show civic awareness than legal immigrants.
Logistic Regressions of General Political Knowledge on Variables
The different migration status categories, however, include persons who have spent varying lengths of time in the country and come at younger ages. When we include the number of years in the United States in the models (column 2), we see as expected that it is positively and significantly related to political civics knowledge. More important, it reduces the size of the naturalization status difference, lowering it by about one-third, from 46 percent above that of ID holders to 20 percent above, a remaining difference that is not statistically significant. In other words, once we take into account that naturalized citizens have lived longer in the country, their level of political civics knowledge and those of the other legal groups are no longer significantly different from each other. By contrast, the deficit for unauthorized immigrants remains quite large (63 percent below that of the ID holders). Including both duration and youthful arrival in the models (column 4) does not change this pattern. Neither does adding indicators of structural integration (columns 5 and 6). Whether we include income by itself, or income, education, and preferring to speak English at home all at once, the deficit in unauthorized political knowledge remains large. However, these factors do explain the differences in civics knowledge among the groups of legal immigrants. As membership-exclusion theoretical considerations would suggest, the major divide in both level and pattern of political knowledge occurs between unauthorized and legal immigrants.
We also estimate models analyzing awareness of unemployment and changes in unemployment. As noted, this indicator represents politically relevant knowledge about the state of the labor market for unauthorized immigrants. In fact, the strongest awareness among Latino immigrants of the direction of recent trends in the strength of the labor market emerges for the group most likely to be labor migrants, namely the unauthorized (table 5, left side). In all of the logistic regression models, such migrants show a tendency to display greater accuracy of unemployment change than the legal groups, though not to a large enough degree to reach statistical significance compared with those holding government IDs, but it is significant when compared with all of the legal groups combined. The two migration-status groups one would expect to enjoy the most secure levels of immigrant integration in the country, the naturalized and the LPRs, are the least likely to know the recent direction of unemployment in the United States, reflecting perhaps that their life situations depend less on such matters. However, this tendency on their part is not statistically significant in either case. Also interesting is that the exclusion of the unauthorized from opportunities for structural advancement in the country appears not to affect their knowledge of the actual strength of the labor market (table 5, right side), though they are sensitive to whether the labor market is changing. This finding is consistent with research showing that unauthorized potential labor migrants in Mexico are well aware of the prospects of finding work in the United States (Massey et al. 1987).
Logistic Regressions of Knowledge of Unemployment Measures on Variables
Theoretical expectations based on consideration of membership exclusion suggest not only that unauthorized Mexican immigrants will show considerably less general political civics knowledge than legal immigrants, but also that they will experience substantially less structural integration. These dynamics imply for unauthorized migrants that measures of structural integration (for example, measures of English usage, education, and income) will show little if any relationship with political civics knowledge, with each other, or with measures of exposure (such as how long people have been in the country or how young they are when they arrive). In fact, when we examine all of these relationships in series of nested regression models like those previously examined, but now for authorized and unauthorized immigrants separately, we see that this is decidedly the case (see table 6, bottom panel). We also pooled these groups and ran models testing for the statistical significance of the effect on general civics knowledge of the interaction between legal status and structural integration (the latter measured by a composite socioeconomic status variable combining education, income, and English proficiency). As expected, the interaction effect was positive and statistically significant, indicating that different patterns of incorporation relationships characterize legal and unauthorized immigrants. Among unauthorized immigrants, scarcely any exposure or structural factors show any relationship with knowledge of political civics, the one exception being temporal exposure, or the longer the immigrants have been here. This is not really surprising, because the simple passage of time inevitably results in acquisition of some familiarity with aspects of the environment. However, because unauthorized status curtails access to structural integration, specific aspects of such integration fail to emerge either singly or in combination with each other, with the result that they show little covariation.
Logistic Regressions of General Political Knowledge on Variables
This contrasts sharply with the patterns among the categories of legal immigrants, where both length of time in the country and facets of structural integration reveal positive relationships with the extent of political civics knowledge (table 6, top panel). The results for authorized immigrants thus in large measure conform with the model of political knowledge that Michael Delli Carpini and Scott Keeter formulated (1996). We see that women are less knowledgeable than men in general about civics matters, this difference coming about both through relationships between gender and other variables (especially education), although gender retains some of its own effect controlling for these (a pattern Delli Carpini and Keeter term a socialization effect). They also observe major effects on political knowledge through the structural positions individuals occupy in society (indicated here by education, English preference, and income). The results for Latino legal immigrants also show these patterns. We also find effects of both exposure and youthful arrival, the latter operating to enhance exposure effects.
Delli Carpini and Keeter also suggest that mediating effects on political knowledge may result from media exposure. Immigrants with more favorable structural positions have more access to media and thus more political knowledge. We would not expect such relationships here for unauthorized immigrants, again because of their minimal structural integration as a result of membership exclusion. But such a pattern might emerge among legal immigrants. We test for this by constructing a simple model involving exposure, structural, and behavioral (media usage) effects on political knowledge. We estimate logistic regressions for these connections consistent with an overall recursive set of relationships among the sets of variables (see table 7). We also use for these the results of the logistic regressions of political knowledge on media usage and other variables shown in model 6 of table 6,4 simplifying the structural effects by showing results for a simple composite structural variable.5 We find that authorized immigrants in fact show some tendency for structural position to influence political knowledge, but this does not operate much through greater access to and reliance on TV news. Rather, the structural factor exerts its own (small) direct effect on political knowledge.
Regressions for Mediating Variables and Antecedent Characteristics
The mediating effect is not large because the effect of TV news on political knowledge, though positive, is not statistically significant. We present pictorial representations of these relationships, for both the groups of legal immigrants and the group of unauthorized immigrants, in figure 3. For the legal immigrants, interlocking associations among exposure, structural, and behavioral mediating variables are extensive, as is clear from the many connections in the diagram and as we would expect among immigrants whose migration status leaves them relatively unfettered in the pursuit and experience of immigrant incorporation dynamics. Among the unauthorized immigrants, by contrast, scarcely any connections among these factors emerge. Again this is what we would expect given the strictures imposed on incorporation opportunities by what we term membership exclusion.
Relationships Among Variables
CONCLUSIONS AND POLICY IMPLICATIONS
Overall, these findings are consistent with expectations. Unauthorized Latino immigrants show significantly lower levels of political knowledge than LPRs, those with driver's licenses or official picture IDs, and the naturalized. More important, this difference holds up when even controls are introduced for exposure (quantity and quality of time in the country) and various kinds of structural incorporation. In short, those subject to severe membership exclusion distinctively lack political knowledge relative to other kinds of immigrants. Structural controls explain differences in political knowledge within the groups of legal immigrants (the LPRs, ID holders, and the naturalized) but do not account for the lower political knowledge of the unauthorized. Moreover, forms of structural integration often mediate the effects of exposure on political knowledge acquisition for legal immigrants but not for the unauthorized, further illustrating mechanisms by which membership exclusion leads to severe disadvantages in the usual kinds of individual-level political incorporation. However, unauthorized migrants, precisely because their status makes them ineligible for conventional political activity, may find recourse in collective action, as evidenced by the mass protests in 2006 to proposed legislation to criminalize their presence in the country, and by the Justice for Janitors campaigns throughout the country during the 1990s and 2000s and the janitors strike in Los Angeles at around the same time. The heightened awareness of changes in the unemployment rate observed here among the unauthorized is consistent with the possibility that factors that directly affect the lives of unauthorized migrants can spark collective protests, such actions being perhaps a sole outlet for political expression. This would be consistent with the findings later in this issue by David Sears, Felix Zavala, and Vanessa Danbold, who observed that unauthorized migrants are less attached to political parties.
What are the policy options for dealing with the stifling effects of membership exclusion? Here we have demonstrated them only for the case of political knowledge, but they are in reality broader and perhaps even more harmful, and likely to grow larger. What can be done? At present, unauthorized migrants have few pathways to legalization and wait years in line even when a pathway is available. Without immigration reform, a greater proportion of today's unauthorized population will remain excluded longer, and under harsher circumstances, than their predecessors who came in the 1960s and 1970s. If the United States provides no pathways to legalization, or if it institutes difficult or punitive pathways (those that include long wait times, large fines for having been unauthorized, sizable fees for legalizing, or prohibitively high thresholds that discourage applications), the size of the group will grow because of the relatively smaller number of natives available to do less-skilled work (Bean et al. 2014). And growth of the unauthorized population would lead to the persistence of educational and other disadvantages for Latino immigrants and their families.
Acknowledgments
We thank the organizers of the workshop (Michael Jones-Correa and James McCann), the other workshop participants, and Carole Uhlaner for helpful suggestions.
FOOTNOTES
↵1. Of course, the target population of the American National Election Studies (ANES) 2012 Time Series Study is citizens of voting age, most of whom are native-born, and whose knowledge of civics might be expected to exceed that of immigrants.
↵2. Our measure of general political knowledge has three levels: no knowledge, knowledge of one political term requirement, or knowledge of both presidential term limits and the length of Senate terms. Here we collapse the knowledge category into a binary variable of no knowledge versus knowledge of at least one set of term requirements. Less than 5 percent of the sample could answer both questions, and when we estimate ordered logit models to allow effects to operate across all three values, these do not yield improvements in fit to the binary logit model. The ordered logit models predict virtually no cases in the category “knowing both kinds of term limits.” We conclude that we do not lose relevant information by combining the categories.
↵3. This numerical transposition brings into alignment the direction of any observed statistical relationship with the direction expected by theory. In this case, the expectation is that longer durations starting from youthful ages will result in additional positive effects on incorporation beyond that of simply duration, so we make the transposition to show this effect working in a positive direction.
↵4. We determine an effect by simply assessing the statistical significance of the relationship. If it is notable in this sense, we depict its association in figure 3 by an arrow in the diagram running from the independent to the dependent variable.
↵5. The composite variable combines z-scores for preferring English at home and income. We do not include education because its relationship with other variables is distorted because for Mexican immigrants, it captures both schooling effects and birth cohort effects, which often work in opposite directions.
- Copyright © 2016 by Russell Sage Foundation. All rights reserved. Printed in the United States of America. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher. Reproduction by the United States Government in whole or in part is permitted for any purpose. We thank the organizers of the workshop (Michael Jones-Correa and James McCann), the other workshop participants, and Carole Uhlaner for helpful suggestions. Direct correspondence to: Susan K. Brown at skbrown{at}uci.edu, Department of Sociology, University of California, 3151 Social Science Plaza, Irvine, CA 92697; and Frank D. Bean at fbean{at}uci.edu, Department of Sociology, University of California, 3151 Social Science Plaza, Irvine, CA 92697.
Open Access Policy: RSF: The Russell Sage Foundation Journal of the Social Sciences is an open access journal. This article is published under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.
