“Times have changed, and whenever they change for the worse, as they have, in-group boundaries tend to tighten. The stranger is suspect and excluded.” (Allport 1954)

To say that we are in a time of change in the USA would be an understatement. In regard to immigration, there has been a major change to the typical pattern of settlement—namely, an increasing number of immigrants are eschewing the traditional urban centers for more suburban and rural pastures. This, in turn, has spurred on a change in attitudes from suburban and rural residents—specifically, switching from a focus on White-Black relations to a native-foreign perspective. This then has incurred a change in how researchers view intergroup relations and attitudes concerning this demographic and contextual shift—from the typical sociodemographic antecedents (e.g., political orientation, race, education, and income) on these attitudes to more socio-contextual factors (e.g., political ideology, views on multiculturalism, rate of immigration growth, and frequency of outgroup contact). Through the incorporation of three well-documented strands of research on attitudes toward unauthorized immigrants and immigration—specifically, political ideology, social labels, and social context—the aim of this experimental study was to investigate the effect of system-justifying motives and immigration-related social label priming (illegal and undocumented) on urban and suburban residents’ attitudes on unauthorized immigrants and immigration.

Political Orientation, Political Ideology, and Immigration

A key facet of investigating attitudes toward unauthorized immigrants in the USA has typically been to focus on political orientation. As a topic of deep contention, political partisanship divides support and opposition to immigration-related issues (Karoly and Perez-Arce 2016), with restrictive policies found in traditionally Republican states and unrestrictive policies found in Democratic states (p.18). Past research presents a clear demarcation between Democratic and Republican voters on policies concerning guest worker visas, a path to citizenship, and deportation (Chavez and Provine 2009; Hajnal and Rivera 2014; Knoll et al. 2010; Walker and Leitner 2011). Politically liberal environments have tended to pass more supportive measures for undocumented immigrants such as driver’s licenses and in-state tuition (e.g., New York City and San Francisco), whereas politically conservative environments have typically passed policies either preventing or prohibiting certain immigrant rights and privileges such as employment and education (e.g., Alabama, Arizona, and Virginia). However, other work seems to argue that demographic factors such as education and income, as well as contextual factors such as rate of Hispanic county population growth and intergroup contact with Hispanics, work in conjunction with political preferences (Hood and Morris 1997, 1998; O’Neil and Tienda 2010).

Previous scholarship also highlights the bidirectionality of partisanship and immigration policy, in that pro- and anti-immigration legislation can be viewed as stemming from individual voters and individual US states separately. Hajnal and Rivera (2014) find that the more anti-immigrant non-Hispanic Whites were, the more strongly they identified as Republican voters. As the authors note, negative views of illegal immigrants are strongly associated with being a Republican. Conversely, voters with more positive views of immigrants tend to be Democratic voters. Consequently, exclusionary policies at both the county- and state-level tend to be found in Republican-leaning areas, thereby showcasing the tendency of municipalities to implement policies consistent with resident attitudes toward immigrants and immigration (Walker and Leitner 2011). The attitudes of the voting public, and by extension their ideology, is a key determinant in immigration legislation (Chavez and Provine 2009).

Although political orientation appears as a strong determinant of attitudes concerning immigration in the USA, others have noted that ideology can trump partisanship such that liberals perform as conservatives under certain conditions, such as issue importance and security threat (Hainmuller and Hopkins 2014; Knoll et al. 2010; Lahav and Courtemanche 2012). Therefore, a psychological focus on political ideology, rather than a sole focus on political orientation, may be better equipped to handle the deeper complexities of sociopolitical attitudes. In fact, liberal-conservative ideology has typically been used to explain anti-immigrant sentiment and hostility in both US and European contexts (Kiehne and Ayón 2016; Saxton and Benson 2003). Other social psychological theories such as Right Wing Authoritarianism (RWA) and Social Dominance Orientation (SDO) have helped explain the bases of these attitudes on an individual level, by focusing on orientations toward coercive control, obedience, and respect (in the case of RWA), or a preference toward hierarchical and unequal intergroup relations (in the case of SDO).

Duckitt and Sibley (2010) provide a conceptual framework for how right-wing politics, nationalism, ethnocentrism, and prejudice are the result of the confluence of RWA and SDO working together, but separately. In essence, the worldview of RWA is that the world is a dangerous and threatening place, while the worldview of SDO is that the world is a competitive environment over group dominance. In regard to sociopolitical attitudes and beliefs, while both are correlated with prejudice, discrimination, and political conservatism, authoritarianism is linked to beliefs in traditionalism and purity, while SDO is associated with power and destructiveness. As such, Duckitt and Sibley (2010) highlight research where: (a) RWA was higher in European countries where immigrants were perceived as increasing the crime rate, while SDO was higher where there was a higher relative immigrant unemployment rate; and (b) RWA predicted aggression toward immigrants who would not assimilate, while SDO predicted aggression toward immigrants who were assimilating. It appears clear that the relationship between attitudes and sociopolitical ideology should be an analytical focal point if we are to better understand the bases of these attitudes (Hainmuller and Hopkins 2014).

System Justification Theory (SJT; Jost and Banaji 1994; Jost et al. 2004), on the other hand, focuses on the propensity of individuals to justify the status quo and view it as legitimate and fair, often on an implicit level. Although it has been shown to reflect the motivations of dominant group members who would wish for the maintenance of social, political, and economic policies that support their interests (as RWA and SDO would predict), system-justifying motives are also prevalent among the disadvantaged as well—those most negatively affected by existing social arrangements (Jost 2019). Hennes et al. (2012) provide evidence that ideological outcomes on policy issues such as global climate change, healthcare law, and immigration reflect epistemic, existential, and relational needs to endorse and support the social, economic, and political status quo. According to SJT, political conservatism is a form of system justification, in that it provides moral and intellectual support for the status quo by resisting change and rationalizing the existence of inequality (Jost et al. 2004). In regard to immigration, a need to defend and justify current US policy enforcement such as raids and detentions can coincide with the legitimization of racial profiling by police and familial separations at the southern border. As such, SJT would predict that these attitudes will be seen as strongest among both political conservatives and those who support the sociopolitical status quo the most.

While the endorsement of Arizona immigration law SB1070 in 2012 was the measure used by Hennes et al. (2012) to gauge immigration policy attitudes, there is missing evidence for a more substantial account of how SJT might impact attitudes toward unauthorized immigration and/or immigrants. For example, meritocracy (as an example of a particular system-justifying ideology) is the notion that one’s actions are the sole explanation for one’s destiny and life path, including success and failure (Patel 2013). McCoy and Major (2007) found that priming meritocracy increases system-justifying attitudes regardless of group status, and decreases perceptions of discrimination among disadvantaged group members. Could system justification help explain other attitudinal items concerning unauthorized immigrants, such as their work ethic, their “criminality,” and their deserving of social welfare benefits? There is plenty of research evidence to suggest that system justification is a particularly strong predictor of attitudes toward immigrants in the context of the USA and its notions of the “American Dream,” but could there be other system-justifying primes, such as language itself, that might also exert an influence on attitudes?

Social Labels and Attitudes Toward Immigrants

Research on the psychological effects of social label manipulation has demonstrated replicated results on a number of different topics including weight bias (Brochu and Esses 2011), sexual orientation (Carnaghi and Maass 2007; Çirakoğlu 2006), race and ethnicity (Donakowski and Esses 1996; Eberhardt et al. 2003; Sibley et al. 2011; Verkuyten and Thijs 2010), and immigration (Knoll et al. 2010; Ommundsen et al. 2014). Although results are not always statistically robust, they are consistent in that a marked difference in attitudes is shown between two or more labels accompanying a topic in the USA, Canada, Europe, and New Zealand. It is worthwhile to mention that the negative attitudes that stem from exposure or usage of a certain label should reflect the derogatory nature of that label—whether the individual acknowledges this or not. Therefore, social labeling need not be a deliberate and motivated selective effort, but rather a cognitive and symbolic representation of “good” vs. “bad.” These cognitive representations, when simplistic and negative (in the case of nouns), have historically been tied to the emergence of ethnophaulisms and the exclusion of ethnic immigrant outgroups in the USA (Mullen 2001).

Could social forces (in the form of language and lexicon) influence policymakers and the general public toward or against certain attitudes thereby impacting legislation? In political and voting matters, the importance of labels is amplified since gender and racial labels and phrases have proven to sensitize and influence voters’ perceptions of candidates, although seemingly innocuous (Zilber and Niven 1995). This translates into the power of symbolic labels and phrases to convey political information and elicit emotional reaction. Domestic legislation concerning racial and ethnic minorities has historically been connected with a linguistic element inherent in criminalizing and punitive terminology, as the history of immigration to the USA has also been accompanied by the “naming” of exclusion (e.g., Operation Wetback).

When discussing immigration policy in general and unauthorized immigration/immigrants in particular, the word illegal appears in media, and in popular and policy discourse—as expressed by news organizations and political parties (Finch 2014). There is immense political polarization between liberals and conservatives around the label attached to unauthorized immigration, with liberals preferring to use the term “undocumented” while conservatives advocating the use of “illegal” to describe immigrants (Merolla et al. 2013). If there is an attitudinal difference in the social labels used to refer to unauthorized immigrants (as illegal or undocumented), then could social labels interact with system justification to help explain these same attitudes? Previous work on authoritarianism, group labels, and attitudes indicates that an interaction exists between the three, since the labeling of social groups and their movements and organizations are the production of politics, and particular wording will resonate with particular ideologies, thereby affecting attitudinal and policy preferences (Donakowski and Esses 1996; Smith et al., 2018). The attitudinal difference in the social labels used to refer to unauthorized immigrants (as illegal or undocumented) may also derive from system justification, in that illegal (with its legal undertone) might resonate more with those who view the social, political, and economic policies affecting unauthorized immigrants as just and fair, while those who do not share this belief may agree more with undocumented (with its circumstantial connotation).

Language may very well influence downstream effects on perceptions and attitudes toward immigrants. As social scientists, however, we are grounded firmly in the belief that attitudes and ideologies are shaped by the social context in which individuals and groups exist and interact. As language is a mediational process between individuals, then ideologies, norms, and attitudes about undocumented immigration may be shared through the socialization of the labels themselves among societal members. This is perhaps dependent on specific environments. Indeed, as linguistic differences in media coverage seem to enhance group biases at the cognitive level (Dunn et al. 2005), while experimental evidence suggests that the framing of immigration policy, rather than of immigrants, has a greater effect on policy attitude change (Merolla et al. 2013), a special focus is directed at the ideological differences between socially and politically different communities that might give rise to these opposing attitudes. Therefore, what role does social context—namely, urban and cosmopolitan vs. suburban and provincial—play in shaping attitudes toward immigrants?

Context Matters

At the heart of the social psychology of sociopolitical ideology and labeling lies the social contexts in which these two elements reside. As previously noted, these two factors are positively correlated. A follow-up question should then be, “Where, when, and how does ideology and labeling converge on immigration?” The answer lies in the social context—more specifically, the communities—where individuals interact and engage with sociopolitical actors and beliefs.

If language is mediational, then social labels, as components of language, reflect tools that individuals use to plan and execute actions in the world (Vygotsky 1986). These tools are cultural in the sense that they are formed, shaped, and/or rejected by the interactions that individuals have in the social world. Therefore, language is seen as the result of a dialectical relationship between inner consciousness and exposure to the outside world (Vygotsky and Luria 1994). This exposure entails a level of participation into pre-existing discourses—broader systems of communication and understanding developed from other, and previous social practices and interpersonal relationships. Legal mandates and policy issues are components of our social worlds and as such, are debated and contested by multiple actors through discourse. Inquiry into such a process could begin with how different communities connect terminologies with ideologies and worldviews.

Political and sociological scholarship on the ideological and intergroup differences between US urban, suburban, and rural residents provide a rationale for the liberal-conservative divide between cities and their outlying environs (Ebert and Ovink 2014; Lichter 2012). Williamson (2008) makes note of the numerous studies confirming the relationship strength between urban residence and Democratic voters, and suburban residence and conservative political orientation. As political conservatism is usually synonymous with anti-immigrant attitudes, suburban residents are more likely to support restrictive immigration policies compared with urban residents (Fennelly and Federico 2008; Marrow 2005).

According to this work, there is a greater frequency of diverse and heterogeneous intergroup contact in cities, whereas there is an emphasis on privacy, homogeneity, and a preference for the familiar in suburban and rural communities. In essence, the socio-ideological differences between cities and their surrounding communities—namely, where conservative political orientation is found in the suburbs, while liberalism tends to be higher in metropolitan and urban locales—are mostly explained by the open and closed socio-spatial structures and space inherent in each. Social interaction in open structures are unavoidable and frequent, and therefore individuals interact with others outside of the typical family and peer groups, while closed structures encourage tight and isolated social interactions, such as family and workplace—thereby emphasizing similarity in cultural norms and beliefs (Thompson 2012). Immigration may be seen by rural and suburban residents as a threat to small-town cohesion, tradition, and the status quo (Garcia and Davidson 2013). Williamson (2008) argues that the private enjoyment of space (e.g., the automobile, the stand-alone home) and the limited interactions between strangers by means of scripted activity through the lack of public space are the hallmarks of the sprawling suburb. This emphasis on material privacy contributes to a private and conservative social and political worldview, one that is typically uncontested by smaller peer groups (Thompson 2012). It also contributes to who we view as “citizens,” what we consider to be “citizenship,” whose perspectives are considered in political decision-making, how power should be distributed, which political institutions are seen as legitimate, and which social policies should be enacted, since this “bounded space” determines who “see” regularly, or not. As a result of the interconnection between social interaction and spatial arrangement, worldviews on diversity and citizenship take shape. A “natural” laboratory to observe this interconnection may very well be the college campus, where the beginning stages of ideological acquisition take place (van Dijk 2002).

NYCC and NJCC: a Tale of Two Colleges

The research sites for this study consisted of two neighboring, but unrelated, community colleges in the USA: New York Community College (NYCC) in New York City and New Jersey Community College (NJCC) in New Jersey. Whereas NYCC is located in an urban social context, NJCC is located in a more suburban location. These two sites also symbolize somewhat contrasting poles of the political spectrum. New York City is, and has been, a socio-politically liberal environment, with 69% of voters registered as Democrats (New York State Board of Elections 2014). By comparison, the corresponding figure in the county where NJCC is located is 21% (New Jersey Department of State 2014). Demographically, as reflected in the student profiles retrieved from the respective colleges’ institutional data for the Fall 2014 semester (the time of data collection), this is a tale of 2 colleges. The student population at NYCC is slightly older, more ethnically and racially diverse, and female—compared with the NJCC population which tends to be slightly younger, White/Caucasian, and evenly split between males and females (Table 1).

Table 1 Fall 2014 student demographic data
Full size table

As public higher education receives its funding primarily from state and local government coffers, the sociopolitical environment of the community college merits consideration. While NYCC and NJCC share the similarity that they are both publicly funded colleges (i.e., taxpayer and voter-supported), they are not situated equally. NYCC is the largest college (26,000 students) within the City University of New York (CUNY) system, which is itself the largest urban public university system in the nation, comprising of 24 campuses (274,000 students). A significant proportion of CUNY first-time freshmen are also first-generation immigrants (Foner 2007). By comparison, the county in which NJCC is located is generally homogeneous. In 2010, US Census records indicate that roughly 4 out of 5 residents were White (82%), while in 2000 the corresponding figure was 87%. As a result, questions arise regarding the influence of cultural heterogeneity and homogeneity.

Whereas there is no national or state prohibition against the admission of undocumented students to US colleges, the individual colleges and universities may have their own policies on admitting undocumented students (Gonzales 2009). Therefore, a closer inspection of the welcoming context of the public community college, as influenced by the political landscape, is merited—particularly in the case of NJCC. Immediately following the September 11, 2001, attacks on the World Trade Center and Pentagon in the USA, the NJCC Board of Trustees authorized an institutional policy preventing the registration of undocumented students. Almost a decade later, the Board rescinded this policy and added the provision that undocumented students could now pay the in-county tuition rate ($115 per credit), which at the time was twice the out-of-county rate ($326 per credit) that undocumented students were subject to, as international students (Caicedo 2014). Yet only after two months, the college repealed this policy after a contentious public hearing where community members voiced their disapproval (Caicedo 2019).

These two disparate colleges were chosen in order to draw clear demographic and geographic differences. It would not have been wise to select another two-year college in the New York State region due to possible contamination effects. Relying on the first author’s residence in the county where NJCC is located, it was decided that a comparison between two-year colleges in neighboring US states, separated by approximately 30 miles (50 km), would form a sufficient geographic contrast. It is certainly possible to draw comparisons between other neighboring social contexts in the geographic sense, but these two sites fit the additional criteria of being socio-politically different as well.

Accordingly, the context of reception as influenced by the political climate of the college community warrants attention as it may be involved in the resources available for undocumented students. Politically and economically conservative communities, compared with liberal ones, may be less willing to support these students—either fiscally (in the form of tuition rates and scholarships) or socially (in the form of international and minority student-oriented events and fairs, and student groups) (Caicedo 2019). It is argued, then, that a contrast between politically distinct communities justifies an analysis regarding social and psychological processes involving immigration.

Current Study

In the current study, we aimed to explore the role that political orientation, system justification (as political ideology), social labels of immigrants, and social context play in shaping attitudes toward immigrants. The study hypotheses were generated from previous work indicating that political conservatism is correlated with anti-immigrant attitudes, as well as providing moral and intellectual support by resisting change and rationalizing the existence of inequality (Jost et al. 2004). Specifically, we hypothesize that, adjusting for demographic variables (age, sex, US citizenship status):

  • H1a: Self-reported political conservatism will predict more negative attitudes toward immigrants;

  • H1b: Greater system justification motivation will predict more negative attitudes toward immigrants;

  • H2: The illegal social label attached to immigrants (as compared with undocumented) will predict more negative attitudes toward immigrants;

  • H3: Social context (namely, NJCC as compared with NYCC) will predict negative attitudes toward immigrants; and

  • H4: Social context will moderate the relationship between political ideology (both political orientation and system justification) and attitudes toward immigrants.

Method

Participants

Participants were community college students at either NYCC or NJCC and, as such, the majority of the NYCC participants were New York City residents, whereas the majority of the NJCC participants resided in the county where NJCC is located. The total sample size was N = 744 and N = 467 from NYCC, and N = 277 from NJCC. Males and females were 42% and 57%, respectively, with ages ranging from 18 to 60. Immigration status was not investigated. Because the research objective was to explore the college samples’ attitudes on unauthorized immigration through an implicit factor in social labeling, inquiry into immigration status might have posed a risk in the shifting of results, or possibly the recruitment of future participants.

The composition of the sample reflects the traditional undergraduate student population, but under two distinct student profiles. The NYCC student tended to be female, older, first-generation immigrant, and multilingual. The NJCC student, on the other hand, tended to be younger, US-born, and monolingual. These differences in the diversity of the student population are representative of the respective colleges’ student profiles.

Materials

Attitudes toward unauthorized immigrants

All participants were presented with a 7-item scale tapping into attitudes toward immigrants, with questions such as “____ immigrants are criminals”; “_____ immigrants are hard-working people”; and “_____ immigrants are deserving of social welfare benefits in the U.S.” (reverse-scored). Response options were in Likert-scale format, ranging from − 3 (strongly disagree) to + 3 (strongly agree). Due to the experimental nature of the study, participants were randomly assigned to one of two conditions: “illegal” and “undocumented.” As referenced above, students were asked to agree or disagree with certain statements, but if they were in the “illegal” condition, they answered statements reflecting “illegal immigrants,” as opposed to those in the “undocumented” condition. All scale items were kept uniform, except for the label manipulation. As such, the manipulation was embedded within the attitude scale. The aggregate attitudes scale had good reliability, Cronbach’s α = 0.778, 95% CI = [0.753, 0.802].

General System Justification (Kay and Jost 2003)

Participants responded to an 8-item measure of General System Justification (Kay and Jost 2003). The GSJ scale reflects endorsement of sociopolitical ideology, such as “In general, you find society to be fair”; “Everyone has a fair shot at wealth and happiness,” and “The United States is the best country in the world to live in.” Two items from this scale were adapted to reflect attitudes toward the status quo of immigration policy, namely “The state’s immigration policies serve the greater good” and “U.S. immigration policy needs to be restructured” (reverse-scored). Items were rated on a scale from − 3 (strongly disagree) to + 3 (strongly agree). The scale displayed good internal consistency, Cronbach’s α = 0.712, 95% CI = [0.679, 0.743].

Political orientation

Political orientation was assessed using one self-report measure, namely, “How liberal or conservative would you consider yourself to be?”, rated on a scale from 0 (completely liberal) to 1 (completely conservative).

Demographic information

The last two pages of the survey consisted of demographic items. This section consisted of residence zip code, sex, country of birth, country of citizenship, marital status, religious group membership, years lived in the USA, and years lived in the State (New York or New Jersey).

Procedure

Recruitment of participants occurred during the Fall 2014 academic semester, using convenience sampling from the two college campuses, NYCC and NJCC.

Random assignment of the questionnaire was completed prior to the visitation of the investigator to the classroom. Utilizing a random number generator, a list of numbers was compiled (0–1000). In sequential order, if a number was “even,” then the corresponding survey was the “undocumented” version. If a number was “odd,” then the corresponding survey was “illegal.” The surveys were then ordered in sequence to the numbers on the list.

Several classrooms were visited across the two colleges to recruit participants. After a brief introduction of the project, a consent form was distributed to the entire class. Once a signed and completed consent form was returned, a randomized survey was given to that individual. If a student did not wish to sign the consent form and/or not participate, then no survey was given to that individual. Upon collection of all the surveys, the class was debriefed and thanked for their participation.

Results

In order to test our hypotheses, we conducted a three-stage hierarchical multiple regression with attitudes toward immigrants as the outcome variable. Demographic variables, namely age, sex, and citizenship status, were entered in the first step. Our main predictive variables, namely (a) college (dummy coded: 0: NYCC, 1: NJCC), (b) label (dummy coded: 0: illegal, 1: undocumented), (c) political orientations, and (d) system justification were entered in the second step. In the third and final step, we entered the college by political orientation and college by system justification interaction terms to test the moderation hypothesis.

Results are displayed in Table 2 below. We mean-centered all the continuous predictors as well as the outcome variable. All assumptions for hierarchical multiple regression were satisfied. Finally, we removed two multivariate outliers/influential from the analysis so that our results are not influenced (Tabachnick et al. 2007).

Table 2 Demographic characteristics of sample
Full size table

Results of the analysis are displayed in Table 2 below. The hierarchical multiple regression revealed that at stage one, demographic variables contributed significantly to the regression model, F (3, 672) = 9.540, p < 0.001, and accounted for 4% of the variation in attitudes toward unauthorized immigrants. Introducing the predictors, namely label (illegal vs. undocumented), college (NYCC vs. NJCC), general system justification, and political orientation explained an additional 11.2% of variation in attitudes and this change in R2 was significant, F (7, 668) = 17.200, p < 0.001. Finally, adding the interaction terms of college x system justification and college x political orientation to the regression model explained only an additional 0.8% of the variation in attitudes toward immigrants, but this change in R2 was also significant, F (9, 666) = 14.137, p < 0.001. When all the variables and their interactions were included in stage three of the regression model, college (context), system justification, and the interaction between college and political orientation were significant predictors, and the interaction between college and system justification was marginally significant (p = 0.093), alongside sex (p = 0.079). These results support some of our hypotheses (namely, H1b, H3, and H4) and reject others (namely H1a). We speculate on why the label manipulation in the discussion—and we mainly attribute it to the fact that the manipulation was embedded within the attitudinal measure, which was our outcome variable. Interestingly, including the interaction terms in the model removed the formerly significant effect of political orientation on attitudes toward immigrants, which indicates that the effect of political orientation on attitudes was dependent on context—which does not completely disqualify H2, as in Step 2 of the regression model below, political orientation stood as a significant predictor (β = 0.111, p = 0.003). The more politically conservative the participants self-reported to be, the more likely they were to hold negative attitudes toward immigrants. The most important predictor of attitudes toward immigrants was college context (β = − 0.298), where participants from NJCC were more likely than participants from NYCC to hold negative attitudes toward unauthorized immigrants, followed by system justification (β = 0.148), where individuals who justified the status quo more were more likely to harbor more negative attitudes toward unauthorized immigrants, in line with our hypotheses (Table 3).

Table 3 Summary of hierarchical regression analysis predicting attitudes toward immigrants
Full size table

In order to probe our fourth and final hypothesis, which suggested that social context will moderate the relationship between political ideology (both political orientation and system justification) and attitudes toward immigrants, we conducted moderation analyses using Hayes’ (2017) PROCESS Macro v3.3 for SPSS. The interaction between political orientation and college (social context) was statistically significant, F (1, 712) = 4.111, p = 0.043, and the interaction graph is displayed in Fig. 1 below. Analyses of simple slopes revealed that for participants from NJCC, political conservatism predicted an increase in negative attitudes toward immigrants to a greater extent (B = 0.109, SE = 0.031, z = 3.515, p < 0.001, 95% CI = [0.048, 0.170]) than for participants at NYCC (B = 0.054, SE = 0.015, z = 3.60, p < 0.001, 95% CI = [0.024, 0.085]). This indicates that political orientation was a stronger predictor of negative attitudes in NJCC as compared with NYCC. The interaction between system justification and context on attitudes toward immigrants was not significant, F (1, 736) = 0.854, p = 0.356.

Fig. 1
figure 1

Interaction graph showing the moderation effect of college on the relationship between self-reported political conservatism and negative attitudes toward immigrants

Full size image

Auxiliary Analyses

What is it about the two social contexts (i.e., the two campuses, NJCC, and NYCC) that drove such differences in results? We conducted some post hoc analyses to uncover some answers to this question. In the survey, we had asked participants to estimate the percentage of the time their friends and family use the terms illegal(s), undocumented, or alien(s) when talking about immigration topics. Chi-square tests revealed that friends of participants from NJCC were more likely than friends of participants from NYCC to use the term illegal(s) (Χ2 (33) = 46.581, p = 0.0294, 1-sided) and alien(s) (Χ2 (30) = 40.310, p = 0.0495, 1-sided), and less likely to use the term undocumented (Χ2 (33) = 39.791, p = 0.0345, 1-sided). Similar results emerged when looking at the family context, where family members of participants from NJCC were more likely than friends of participants from NYCC to use the term illegal(s) (Χ2 (28) = 51.279, p = 0.002, 1-sided) (p = 0.0495, 1-sided), and less likely to use the term undocumented (Χ2 (25) = 34.181, p = 0.052, 1-sided, marginally significantly). There were no significant differences in the use of the term alien(s) by families from both contexts (Χ2 (24) = 23.363, p = 0.249). In terms of how often participants encounter the terms illegal(s), alien(s), and undocumented when hearing or reading about immigration topics, similar patterns emerged as well. Participants from NJCC were more likely to hear the terms illegal(s) (Χ2 (29) = 39.273, p = 0.048, 1-sided) and marginally less likely to hear the term undocumented (Χ2 (30) = 38.607, p = 0.0673, 1-sided, marginally significant) compared with those from NYCC. There were no significant differences around hearing the word alien(s) (Χ2 (27) = 29.107, p = 0.177, ns, 1-sided). When reading about immigration issues, participants from NJCC were more likely to encounter the terms illegal(s) (Χ2 (30) = 45.719, p = 0.017, 1-sided) and alien(s) (Χ2 (29) = 39.960, p = 0.042, 1-sided), but not less likely to hear the term undocumented (Χ2 (33) = 30.687, ns, 1-sided, p = 0.291 ) compared with those from NYCC.

Discussion

As an eerie precursor to modern times, Allport (1954) wrote, “During the 124 years for which data are available, approximately 40,000,000 immigrants came to America, as many as 1,000,000 in a single year. Of the total immigration 85 percent came from Europe. Until a generation ago, few objections were heard. But today nearly all applicants are refused admission, and few champions of ‘displaced persons’ are heard” (p.35). The present study adds to the sociological work done on the linguistic bases of public opinion and policy regarding immigration, as well as the political psychological literature on the ideological underpinnings of policy attitudes, by offering a closer look into the communities that house such beliefs. As such, one of the objectives of this study was to merge these overlapping avenues of research in an attempt to tell one coherent story—that the beliefs one has could be matched with the words that one uses, which are themselves found in the place one calls home. To our knowledge, this study is one of a sparse few that attempt to marriage these concepts simultaneously.

Summary of Results

The results of this study indicated that social context—which was broadly operationalized through two college campuses in New Jersey and New York, the former located in a more suburban neighborhood, while the latter in an urban and highly diverse city—influences attitudes toward immigrants, and the effects that certain ideological sets, such as political conservatism, have on attitudes. We found that greater justification of the status quo predicted more negative attitudes toward immigrants. Consistent with prior research, we also found that political conservatism predicted negative attitudes toward immigrants—however, this effect was qualified by a significant context by ideology interaction, where political conservatism predicted negative attitudes more strongly for participants from NJCC than from NYCC. Finally, we did not find an effect of social label (“undocumented” vs. “illegal” immigrants) on attitudes, contrary to one of our main hypotheses. We attribute this null result to methodological constraints and discuss that further in the “Limitations” section below. Finally, we probed as to what about the social context may contribute to the attitudinal differences across both college campuses, and found evidence that participants from NJCC were more likely than their counterparts in NYCC to hear friends and family use the terms illegal(s) and alien(s) and less likely to hear them use undocumented when discussing immigration. NJCC participants were also more likely to hear and read the illegal(s) as opposed to undocumented as they pertain to immigration and immigrants.

Strengths

Among the strengths of this study was its large sample sizes in an experimental format, and its relevance to both the heart of social psychology (i.e., the individual in social context) and to a timely and divisive topic in US political discussion (immigration).

While label priming did not provide conclusive evidence of the cognitive and ideological implications of the language in the immigration debate and policy, social environment did—specifically, that urban and suburban settings provided a point of difference in both unauthorized immigrant attitudes and sociopolitical ideology to a significant degree. Urban students may be, know of, or have had actual interactions with unauthorized immigrants to a higher degree than suburban students, and may therefore have more positive attitudes toward that group. In addition, the “Benetton effect,” where diversity is perceived as an asset, may be higher in urban and cosmopolitan settings compared with suburban and homogeneous settings, where diversity may be viewed as problematic. In a cognitive effort to maintain homogeneity, suburban students may wish to support the status quo, rather than wish to alter it, compared with students in urban settings where diversity is the “norm.”

Given the auxiliary results demonstrating that urban students reported “hearing” and “seeing” the term “undocumented” to a greater degree compared with suburban students (who reported “hearing” and “seeing” the term “illegal” more), the relationship between social label exposure and social environment is used as evidence for the interactions that exist between labels and individuals in differing contexts. These results indicate that the role of social interactions, as well as interactions with media, has a powerful influence on sociopolitical attitudes. In other words, these results provide an explanation as to the effect of social environment on sociopolitical attitudes.

Limitations

There are a few methodological limitations that should be considered in the interpretation of these results. First, the NJCC sample was smaller than the NYCC sample by approximately 200 participants. Although still representative of the NJCC student population, sample equity concerns merit caution. Second, the experimental manipulation was embedded in the dependent variable measure. While past research on the linguistic effects of labeling and framing has utilized factual or fictional narratives to prime attitudes, we opted for a “conversational” approach where respondents are asked items without the presence of a narrative prime. However, it is unclear whether the participant attitudinally responded to the label or the statement, and whether there was a disagreement between the two.

Additionally, a multi-item approach to investigating political orientation should be better equipped to differentiate the nuances between areas of liberalism and conservatism (i.e., political, social, and economic), rather than a single-item measure such as the one utilized in this study. Likewise, although various items on the survey inquired into participant demographics, questions relating to race or ethnicity were not included. Therefore, associations between participant demographics and their views on unauthorized immigrants are not able to be made. Likewise, personal connections between participants and the topic of unauthorized immigration were not investigated, thereby limiting the analysis between who the participants were in relation to unauthorized immigration and their attitudes, as reflected in the survey. Finally, while survey anonymity and confidentiality was emphasized to the participants, there are limitations inherent with self-report measures. Nevertheless, given the mixed confirmation of hypotheses, we are confident that the results garnered were not due to any systematic response biases.

Implications

The present study expanded the reach of System Justification Theory to serve as an explanatory theory for the experimental investigation of immigration-related attitudes by including the traditional social psychological factor of social context into the equation. By focusing less on a micro-level analysis on the attitudinal antecedents (e.g., RWA, SDO) and following Pettigrew’s (2006) recommendation that a macro-level emphasis is required for the contextualization of social psychological work, this study used the nexus between political ideology and social context to deepen our understanding of such attitudes at the meso-level. Previous work has demonstrated that regional and temporal differences can affect political culture and ideology (Gaucher et al. 2018; Rentfrow et al. 2009). An investigation into urban, suburban, and rural residents’ tendency to support the status quo on a host of sociopolitical topics might offer a more nuanced insight into how individuals, living under diverse spatial arrangements, view their communities, states, country, and the world. Therefore, future research should draw upon community-level differences in social, cultural, and political ideology to uncover both conceptual and practical intergroup processes.

Further analysis into the antecedents of such community-level differences may also prove fruitful. Probing the sociocultural development of individuals—such as the educational resources and environment present in different communities, or the influence of homogeneous or heterogeneous peer relations—might also demonstrate to be powerful factors in the evolution of sociopolitical attitudes, in distinctive geographic and residential arrangements.

Additionally, as research concerning the effects of social labeling on attitudes, perception, and memory indicates, results are mixed for a variety of conceptual and methodological reasons. While some have demonstrated significant attitudinal effects contingent upon the labels used (Carnaghi and Maass 2007; Çirakoğlu 2006; Donakowski and Esses 1996; Ommundsen et al. 2014; Verkuyten and Thijs 2010), others have argued for the importance of issue framing (rather than label priming), as ideological effects seem more robust (Knoll et al. 2010; Lahav and Courtemanche 2012; Merolla et al. 2013). Latter proponents claim that social labels are often not subjectively equal, and therefore individuals may be emotionally reacting to labels rather than through cognitive appraisal. Indeed, the present study did not find significant effects for differential use of social labels, but future work should focus on how issue framing might serve to increase or decrease support for the system and the status quo—opting to view “the forest” rather than “the trees.”

Heeding the suggestions by Dovidio and Esses (2001) almost two decades ago, this study sought to address three distinct factors involved in the psychological study of immigration—ideology, language, and context. Although data collection occurred during 2014, the results hold relevance today, as seen in the public and legislative responses to immigration debate issues surrounding a path to citizenship, family separation, and national security. The authors confidently argue that this study complements the work of political science and sociology in the field of immigration, but that additional research is required if we are to understand a changing populace for decades to come. As Allport (1954) wrote over half a century ago, “The native American nowadays seldom takes an idealistic view of immigration” (p.34). It is our sincerest hope that this will also change.