Skip to main content
NCBI home page
Public Opinion Quarterly logoLink to Public Opinion Quarterly
. 2016 Jan 13;80(1):1–25. doi: 10.1093/poq/nfv054

Party Identification, Contact, Contexts, and Public Attitudes toward Illegal Immigration

Timothy B Gravelle 1,*
PMCID: PMC4884820  PMID: 27257305

Abstract

Illegal immigration is a contentious issue on the American policy agenda. To understand the sources of public attitudes toward immigration, social scientists have focused attention on political factors such as party identification; they have also drawn on theories of intergroup contact to argue that contact with immigrants shapes immigration attitudes. Absent direct measures, contextual measures such as respondents’ ethnic milieu or proximity to salient geographic features (such as borders) have been used as proxies of contact. Such a research strategy still leaves the question unanswered – is it contact or context that really matters? Further, which context, and for whom? This article evaluates the effects of party identification, personal contact with undocumented immigrants, and contextual measures (county Hispanic population and proximity to the US–Mexico border) on American attitudes toward illegal immigration. It finds that contextual factors moderate the effects of political party identification on attitudes toward illegal immigration; personal contact has no effect. These findings challenge the assumption that contextual measures act as proxies for interpersonal contact.

Current estimates now place the number of undocumented immigrants in the United States at over 11 million (Baker and Rytina 2013; Passel et al. 2014), making the challenge of illegal immigration one of the most pressing issues on the American policy agenda. In recent years, deadlock between the House of Representatives and the Senate on reforms to immigration and naturalization policy, state-level initiatives targeting undocumented immigrants, President Barack Obama’s executive action staying the enforcement of deportation for millions of undocumented immigrants, and provocative statements by candidates for the 2016 Republican presidential nomination have all contributed to raising the political temperature on immigration issues in the United States. The manner in which the debate around illegal immigration plays out highlights the diversity of opinion among the American public (Segovia and Defever 2010; Muste 2013), and public opinion will remain an important consideration in any new policy proposals.

What, then, are the factors that account for Americans’ policy preferences regarding illegal immigration? In endeavoring to explain mass public opinion toward illegal immigration (and immigration attitudes more generally), social scientists point to political predispositions, perceptions of competition for jobs between native-born and immigrant workers, and perceptions of cultural threat. In addition to these individual-level factors, numerous studies borrow from the long tradition of research on racial attitudes in the United States in developing explanations of immigration attitudes grounded in intergroup contact (McLaren 2003; Harrison 2006; Berg 2009; Newman 2013). A different approach is pursued by those seeking a contextual explanation of immigration attitudes in the demographic composition of individuals’ local areas, though results are mixed (Hood and Morris 1997, 2000; Campbell, Wong, and Citrin 2006). A similarly contextually and spatially oriented approach analyzes immigration attitudes as a function of proximity to the US–Mexico border (Branton et al. 2007). Work on the effects of racial/ethnic context and spatial proximity often assumes that such contextual measures serve as a proxy for personal contact (e.g., Fetzer 2000; Branton and Jones 2005; Dunaway, Branton, and Abrajano 2010). This assumption often stems from the fact that measures of personal contact with (unauthorized) immigrants are rare in social surveys while contextual data are often available.

The current state of research on public opinion toward immigration, then, is one of disparate results relating to the effects of both intergroup contact and ethnic and spatial context, which typically are not evaluated in tandem (but see McLaren 2003; Harrison 2006). This leaves unanswered the question: Does context serve as a proxy for personal contact, or does it capture other psychological or social processes? Put more plainly, is it contact or context that ultimately matters for immigration policy attitudes? Further, which context (or contexts) matter, and for whom?

This article takes up these questions. It contends that the contexts individuals occupy are characterized by a variety of concurrently operating dynamics, and so it makes sense to measure and model the effects of context in multiple ways. Further, this article demonstrates that how individuals experience a given context is contingent upon their preexisting political characteristics. To advance these arguments, nationally representative survey data are used to evaluate the effects of party identification, personal contact with undocumented immigrants, and contextual measures on attitudes toward illegal immigration. Findings show that larger Hispanic populations at the local level, larger increases in the local Hispanic population over time, and increasing proximity to the US–Mexico border all amplify the partisan cleavage between Democrats and Republicans in attitudes toward illegal immigration. That is, contextual and spatial factors moderate the effect of party identification on attitudes toward illegal immigration. At the same time, reported personal contact with undocumented immigrants has no significant effect. These findings challenge the assumption that contextual measures are proxies for interpersonal contact.

The article first reviews the current immigration policy attitudes literature, paying particular attention to research dealing with party identification, contact, and contextual influences. It then advances hypotheses relating to party identification, contact with undocumented immigrants, and the interaction of party identification with contextual measures. The following section describes the survey data and methods used in the study, and the results of the analyses. The final section discusses the findings’ implications for explaining attitudes toward illegal immigration, and outlines directions for future inquiry.

The Sources of (Illegal) Immigration Attitudes

In seeking to explain why individuals are alternatively open and accommodating or strict and restrictionist in their views toward illegal immigration (or immigration more broadly), the social science literature offers a range of hypotheses (see Hainmueller and Hopkins [2014] for a recent review). Research in a political economy tradition focuses on economic self-interest in the form of labor-market competition between native-born and foreign-born workers, with native-born workers more likely to experience competition for jobs expressing more restrictionist attitudes (Espenshade and Calhoun 1993; Citrin et al. 1997; Scheve and Slaughter 2001; Malhotra, Margalit, and Mo 2013). Research testing the sociotropic effects of national economic conditions on immigration attitudes generally finds that area unemployment levels exert little effect on immigration attitudes (Campbell, Wong, and Citrin 2006), though changes in unemployment over time or a downturn in the national economy increase restrictionist sentiment (Hopkins 2010; Goldstein and Peters 2014). Other explanations for restrictionist attitudes emphasize the “symbolic” threat posed by immigrants to core national values or national culture (Schildkraut 2011; Newman, Hartman, and Taber 2012; Hainmueller and Hopkins 2014) or the liberalizing effect of higher education (Haubert and Fussell 2006; Hainmueller and Hiscox 2007).

In the American context, policy responses to the challenge of illegal immigration—ranging from Propositions 187 and 227 in California in the mid-1990s, to SB-1070 in Arizona in 2011—take on a partisan cast among political elites and the mass public. Research testing the effect of party identification on immigration attitudes finds that Republicans tend to favor restrictionist immigration policies more frequently than Democrats (Hawley 2011; Schildkraut 2011; Hainmueller and Hopkins 2014).

Some authors draw on strands of inquiry first developed in the study of racial attitudes in the United States to understand immigration policy attitudes. One such strand is intergroup contact theory (Allport 1954; Pettigrew 1998; Pettigrew and Tropp 2006; Pettigrew et al. 2011). Its core argument is that contact between members of different groups holds the potential to decrease outgroup prejudice. For this to occur, contact requires close acquaintance, equal status, and common goals (Allport 1954; Amir 1969). Conversely, in situations involving competition for scarce resources, contact serves to increase outgroup prejudice (Blalock 1966). In transferring the concepts and hypotheses of intergroup contact theory from racial attitudes to immigration attitudes, research finds that some forms of close contact decrease restrictionist immigration attitudes (McLaren 2003; Harrison 2006; Berg 2009). On the other hand, incidental contact increases restrictionist attitudes through increased perceptions of cultural threat (Enos 2014; Newman 2015).

Another strand of inquiry originating in the racial attitudes literature focuses on the effects of local demographic context. In Key’s (1949) and Blalock’s (1956, 1957) classic work on the southern United States, the proportion of African Americans is shown to be positively associated with more racially conservative attitudes among whites—findings that have since been reconfirmed and extended (Glaser 1994). Similar contextual approaches are applied to analyses of immigration attitudes, but with mixed results. Some studies find that increases in the proportion of foreign-born residents or Hispanics in the local area increase the likelihood of expressing restrictionist immigration attitudes (Hero and Tolbert 1996; Tolbert and Hero 1996; Campbell, Wong, and Citrin 2006), while others find the opposite—larger proportions of foreign-born residents or Hispanics decrease restrictionist attitudes toward immigration (Hood and Morris 1997, 1998). Still others report inconsistent or null results (Citrin, Reingold, and Green 1990; Fetzer 2000; Hood and Morris 2000). Others argue that it is not the ethnic context per se that influences attitudes toward immigration policy; rather, it is change in the context over time—for example, local areas experiencing rapid growth in the Hispanic population exhibit higher levels of restrictionist immigration attitudes (Hopkins 2010; Newman and Velez 2014).

A final line of research emphasizes proximity to (or, conversely, distance from) the US–Mexico border, as Mexico serves as the country providing the largest number of immigrants to the United States. This work finds that individuals residing closer to the US–Mexico border are more likely to support nativist policy initiatives such as restricting undocumented immigrants’ access to social services (Branton et al. 2007).

Theory and Hypotheses

Given the multiplicity of both theoretical explanations for individual-level immigration policy attitudes and empirical findings, this study focuses on the hypothesized effects of party identification, intergroup contact, ethnic and spatial contextual factors, and their interactions. I acknowledge other competing explanations (as discussed above), and account for them by including controls in this analysis.

Recent attempts at immigration reform at the federal level, and state-level political developments, such as Propositions 187 and 227 in California and SB-1070 in Arizona, produced a partisan and ideological sort among political elites. Partisan polarization translates into greater cue-taking among the mass public and thus sharpens partisan cleavages in mass opinion on immigration issues (cf. Zaller 1992). Research on different conceptions of American national identity finds that the ethno-cultural aspect of American identity resonates most strongly with conservatives and Republicans, where belonging is defined in terms of particular ascribed characteristics: being white, Protestant, of northern European descent, and speaking English as a native language (Schildkraut 2005, 2011). Democrats and liberals are more likely to endorse “incorporationism,” or, a set of beliefs about American identity that acknowledges the role of immigration in shaping the country, and that values pluralism and tolerance of difference (Schildkraut 2007; Hajnal and Rivera 2014). Thus, this article proposes the following hypothesis linking party identification to attitudes toward undocumented immigrants:

H1: Democrats are more likely to favor allowing undocumented immigrants to remain in the United States; Republicans are more likely to favor requiring undocumented immigrants to leave.

As for the expected effects of contact and contextual factors on attitudes toward illegal immigration, it is useful to distinguish between the expected effects of interaction (direct, personal contact), proximity (to a given outgroup), and proportions (of an outgroup within an area), since each implies different processes (Forbes 1997). With respect to the hypothesized effects of contact, expectations are straightforward. In line with the literature on intergroup contact and immigration attitudes, this article posits:

H2: Individuals having personal contact with undocumented immigrants are more likely to favor allowing them to stay in the United States; those not having experienced such contact are more likely to favor requiring undocumented immigrants to leave.

As for contextual factors, the milieu inhabited by individuals can be characterized in multiple ways. Both the proportions of and proximity to particular outgroups are potentially relevant (Forbes 1997). It is therefore reasonable to examine the effects of ethnic (Hispanic) context, change in ethnic context, and proximity to the US–Mexico border on attitudes toward illegal immigration. All of these measures may bear on attitudes toward illegal immigration, as “people develop subjective understandings of the places they live in based on objective local characteristics, particularly the social composition” (Cutler 2007, 579). The focus on Hispanics and proximity to Mexico in the context of American immigration issues is warranted not only by Mexico’s position as the first-ranked source of immigrants to the United States, but also because the issue of illegal immigration is repeatedly tied to Mexicans in public debate (Campbell, Wong, and Citrin 2006). Non-Hispanic whites’ attitudes toward legal immigration, specifically Mexican immigration, and immigration reform (granting amnesty to undocumented immigrants) are highly correlated (Ayers et al. 2009).

Further, it is plausible that these contextual factors will not exert the same effects on all segments of the American public (cf. Johnston, Newman, and Velez 2015). Indeed, recent research on contextual effects suggests that demographic and spatial contexts are more relevant for their moderating effects on individual-level partisanship than as predictors of policy attitudes in their own right. In this formulation, party identification is the “focal” independent variable, and contextual factors are the moderator variables (Jaccard and Dodge 2004). Recent work shows that Democrats and Republicans are influenced by context (variously measured) in different ways. Generally, those on the political right are more likely to react defensively to perceived local threats (Fischhoff et al. 2003). For example, Hopkins (2014) finds that high-density Republican-registration Census block groups exhibit stronger support for Proposition 227 (intended to end bilingual education in California) when English–Spanish bilingual ballots are employed in their county than comparable block groups with English-only ballots. Branton et al. (2007) find that support for Proposition 187 (intended to deny access to social and health services to undocumented immigrants) and Proposition 227 among California Democrats increases as the US–Mexico border becomes more proximate while remaining consistently high among Republicans. Drawing on national data, Hawley (2011) finds a somewhat different pattern: Republicans are consistent in their support of more restrictionist immigration policies regardless of spatial or demographic context, but Democrats become less (not more) restrictionist as the county-level proportion of foreign born increases. Research finds that providing an explicit ethnic (Mexican) cue in survey item wording elicits a stronger restrictionist response among Iowa Republicans (Knoll, Redlawsk, and Sanborn 2010). Taken in sum, these spatially and contextually contingent results point toward the following hypotheses:

H3: With increasing Hispanic concentration in the local area, partisans’ attitudes will further diverge, with Republicans becoming less likely to favor allowing undocumented immigrants to stay in the United States, and Democrats becoming comparatively more likely to favor such a policy.

H4: With increasingly large changes in the Hispanic population in the local area over time, partisans’ attitudes will similarly further diverge, with Republicans becoming less likely to favor allowing undocumented immigrants to stay in the United States, and Democrats becoming comparatively more likely to do so.

H5: With increasing proximity to the US–Mexico border, partisans’ attitudes will again further diverge, with Republicans becoming less likely to favor allowing undocumented immigrants to stay in the United States and Democrats becoming more likely to favor such a policy.

Does it make sense, though, to advance hypotheses relating to (and to model simultaneously) reported contact and contextual factors? One view might characterize this as a form of double-counting at the level of theory, and an invitation to collinearity at the level of empirical analysis. Empirical measures such as Hispanic density, Hispanic change, and US–Mexico border proximity are, in this view, simply proxies of personal, face-to-face contact, and serve as a second-best solution when more direct measures of contact are not available (Hopkins, Tran, and Williamson 2014; Newman 2015). What matters at the level of theory, then, is still contact, not context. Some research makes this proxy admission explicit, arguing that this is reasonable given that “geographic proximity increases the likelihood of contact” (Ayers et al. 2009, 596). Other work exhibits ambiguity on this point, for example when intergroup contact theory is described as “point[ing] to an explanation based on ethnic context” (Dunaway, Branton, and Abrajano 2010, 363), or the percentage of foreign-born residents or Hispanics at the county level is described as a measure of “interaction” or “contact” with immigrants (Hood and Morris 1997; Fetzer 2000; Berg 2009). Still, if it follows as a matter of theory that contextual measures serve only as proxies for direct intergroup contact, then as an empirical matter one should not find effects of contextual measures on immigration attitudes when controlling for contact.

There is some doubt, though, that ethnic context proxies intergroup contact: “local intergroup contact may be limited by language barriers and may be overwhelmed by the real or perceived threat that immigrants pose” (Hainmueller and Hopkins 2014, 236). Indeed, those few studies that parse out the effects of contact and context “highlight the importance of distinguishing context from contact in theory and measurement and suggest strong caution in relying upon the former as an indicator of the latter” (Newman, Hartman, and Taber 2012, 641). In light of this, if contextual measures are not mere proxies for contact, then what do they capture? Recent literature points to two alternative mechanisms of influence: passive exposure to the Spanish language and media exposure.

In a field experiment involving the random assignment of native Spanish speakers to commuter trains, Enos (2014) finds that passive exposure to Spanish shifts attitudes in the direction of tighter restrictions on immigration: subjects in the treatment group were significantly more likely to express preferences for reduced immigration from Mexico, and less likely to favor allowing undocumented immigrants to remain in the United States. More restrictionist immigration attitudes are also observed in survey-based experiments when test subjects are exposed to Spanish-language text (Newman, Hartman, and Taber 2012; Hopkins, Tran, and Williamson 2014). It is a reasonable expectation that such passive exposure to Spanish will occur more frequently in areas with larger proportions of Hispanics or with rapidly growing Hispanic populations. The reason for observing such effects is that “the Spanish language operates as a potent cue on immigration-related issues,” increasing the salience of immigration (Hopkins, Tran, and Williamson 2014, 37). Exposure to Spanish within one’s local milieu may also engender a sense of cultural threat, which in turn leads to more restrictionist immigration attitudes (Newman, Hartman, and Taber 2012; Enos 2014).

While proximity to the US–Mexico border provides cues through border checkpoints, fences, and warning signs, and thus heightens perceptions of threat, border proximity also influences the media environment in which individuals reside. Media content analysis has shown that newspapers closer to the US–Mexico border publish more negative stories about immigration (both legal and illegal immigration) more frequently than those further away (Branton and Dunaway 2009a, 2009b). Dunaway, Branton, and Abrajano (2010) find that media coverage of immigration in turn both raises the salience of the issue among the public and engenders more restrictionist immigration attitudes, though Lawlor (2015) finds that the effects of contextual factors (such as the proportion of foreign-born residents) on media framing of immigration are inconsistent.

Data and Methods

The data for this study come from a representative survey of the American public (18 years of age and above) conducted by the Pew Research Center on June 12–16, 2013, using a dual-frame sample design comprising landline and cellular telephone samples (n = 1,512). The response rates (AAPOR RR3 definition) for the survey were 8.7 and 5.8 percent for the landline and cell phone samples, respectively. In keeping with both the substantive focus and methodological practice of much of the literature on American immigration attitudes, I limit my analysis to the subsample of non-Hispanic white respondents. I further limit my analysis to data from the lower 48 continental states (since Alaska and Hawaii are not part of the contiguous land mass sharing a border with Mexico), bringing my final sample to 1,086.

The key measure of Americans’ attitudes toward illegal immigration is a dichotomous survey item that asks: “Which comes closer to your view about how to handle undocumented immigrants who are now living in the US? They should not be allowed to stay in the country legally, [or] there should be a way for them to stay in the country legally, if certain requirements are met.” The first response category expresses a restrictionist sentiment in that it proposes to continue to exclude undocumented immigrants from legal residence and citizenship in the United States, while the second response category expresses greater openness to immigration and greater willingness to extend the rights of citizenship. The first response is a form of moral parochialism, in that the right to reside in the United States ought to be restricted largely to those having acquired that right by birth, or those having become naturalized citizens or having other legal status (cf. Wong 2010). The second response speaks to the liberal (or Tocquevillian) tradition in American national identity that stresses openness, freedom, and egalitarianism (Smith 1993) as well as incorporationist notions of pluralism and tolerance (Schildkraut 2007). In the statistical analyses that follow, the second response category (“there should be a way for them to stay in the country legally”) serves as the modeled outcome in a set of binary logit models.

The Pew data indicate that 32 percent of non-Hispanic white Americans would exclude undocumented immigrants from residence and citizenship, while 66 percent are prepared to allow them to regularize their status in the United States given a set of requirements. These results are consistent with those of other research organizations from the mid-2000s to the present, with opposition to allowing undocumented immigrants to stay typically in the range of 26 to 32 percent, and support for allowing them to stay (whether as temporary workers or as citizens) in the range of 64 to 70 percent (Muste 2013).

The Pew survey data contain the individual attitudinal and behavioral variables of interest, including measures of party identification, contact with undocumented immigrants, as well as relevant demographic and attitudinal controls. (Full details of the data coding appear in the appendix.) To measure respondents’ ethnic context, I use the percentage of Hispanics at the county level taken from the 2010 US Census of Population (www.census.gov/popest/). I also calculate the percentage-point change in Hispanic population between the 2000 and 2010 decennial censuses. While counties are not the most granular measure of local context, the only indicators available in the Pew data are county FIPS codes and ZIP codes. Still, it is important to stress that counties are politically consequential “containers” in the American context: elections and the provision of certain public goods take place at the county level (Glaser 1994; Branton and Jones 2005), and counties also approximate media markets (cf. Branton and Dunaway 2009a, 2009b).1

Obtaining a measure of spatial proximity to the US–Mexico border is more involved. To calculate this, I first geocode (append latitude–longitude coordinates to) individual survey respondents based on reported ZIP codes and ZIP codes retained from the telephone sample files. Nearly all respondents (1,076, or 99 percent) were geocoded using ZIP codes, with the remaining 10 respondents geocoded based on county FIPS codes. Next, I perform a Cartesian join between the survey data and a digitized map (shapefile) of the US–Mexico border to calculate the distance between each respondent and the border.2

Since missing data can introduce bias into the model parameter estimates and also reduce statistical power, missing survey data were imputed 10 times in order to retain all cases in the regression analysis (Allison 2001; Little and Rubin 2002). The binary logit models of the multiply imputed data were then fit using procedures that account for the complex sample design used in collecting the survey data. These results were then combined according to procedures detailed by Rubin (1987) to produce the final reported results.3

Results

The models reported in table 1 yield several findings of note. As shown in model 1, Republican Party identification has only a marginally significant negative effect (p = 0.07) after controlling for demographics, contact, and contextual factors on allowing authorized immigrants to stay. This finding suggests that the partisan cleavage between Democrats and Republicans is neither large nor robust. The data thus fail to confirm H1, and contrast with previous findings of a partisan cleavage in immigration policy attitudes (e.g., Hajnal and Rivera 2014).

Table 1.

Explaining American Attitudes toward Illegal Immigration (Binary Logit)

Model 1 Model 2 Model 3 Model 4
b (SE) b (SE) b (SE) b (SE)
Intercept 0.63 (0.33)† 0.76 (0.33)* 0.71 (0.33)* 0.83 (0.35)*
Male –0.03 (0.17) –0.03 (0.17) –0.02 (0.17) –0.08 (0.17)
Age (ln years) –0.23 (0.34) –0.23 (0.35) –0.22 (0.35) –0.27 (0.34)
Education: College 0.60 (0.17)*** 0.57 (0.17)** 0.59 (0.17)*** 0.61 (0.17)***
Region (ref = Northeast)
 Midwest 0.15 (0.26) 0.09 (0.26) 0.12 (0.26) 0.06 (0.26)
 South 0.40 (0.27) 0.34 (0.27) 0.38 (0.27) 0.29 (0.28)
 West 0.77 (0.34)* 0.71 (0.35)* 0.75 (0.34)* 0.66 (0.35)†
Economy (ref = About the same)
 Getting worse –0.73 (0.19)*** –0.75 (0.19)*** –0.74 (0.19)*** –0.73 (0.19)***
 Getting better 0.75 (0.23)*** 0.73 (0.23)** 0.74 (0.23)** 0.76 (0.23)***
Party identification (ref = Democrat)
 Independent –0.35 (0.31) –0.46 (0.31) –0.42 (0.32) –0.44 (0.31)
 Republican –0.39 (0.22)† –0.47 (0.22)* –0.45 (0.22)* –0.48 (0.22)*
Ideology (ref = Liberal)
 Moderate 0.04 (0.27) 0.05 (0.27) 0.06 (0.27) 0.04 (0.27)
 Conservative –0.39 (0.30) –0.38 (0.30) –0.39 (0.30) –0.38 (0.30)
Contact with undocumented immigrants 0.25 (0.19) 0.26 (0.19) 0.25 (0.19) 0.25 (0.19)
ln county Hispanic % 0.13 (0.16) 0.41 (0.21)* 0.13 (0.16) 0.17 (0.16)
ln county Hispanic % point change (+ 5) –0.12 (0.51) –0.08 (0.52) 0.94 (0.63) –0.27 (0.51)
ln distance to US–Mexico border (km) 0.21 (0.16) 0.14 (0.16) 0.18 (0.16) –0.41 (0.34)
Independent × ln county Hispanic % — –0.53 (0.30)† — —
Republican × ln county Hispanic % — –0.45 (0.18)* — —
Independent ×
ln county Hispanic % point change		 — — –0.94 (1.07) —
Republican ×
ln county Hispanic % point change		 — — –1.71 (0.64)*** —
Independent ×
ln distance to US–Mexico border		 — — — 1.23 (0.46)**
Republican ×
ln distance to US–Mexico border		 — — — 0.67 (0.33)*
N 1,086 1,086 1,086 1,086
Model χ2 135.82*** 146.72*** 145.26*** 147.79***
Likelihood ratio χ2 — 10.90** 9.44** 11.98**
Nagelkerke pseudo-R2 0.16 0.18 0.17 0.18
Open in a new tab

†p ≤ 0.10; *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001

Also, while the coefficient for contact with undocumented immigrants is positively signed (and thus in the theoretically expected direction), it does not approach conventional levels of statistical significance. This finding of no significant effect of contact with undocumented immigrants is in line with the null results for different forms of “impersonal contact” (such as contact with recent immigrants in service establishment settings) reported by Harrison (2006). It contrasts, however, with the results reported by Ellison, Shin, and Leal (2011) and McLaren (2003) where contact reduces restrictionist sentiment. The critical distinction appears to be in the quality of contact. The Pew questionnaire prompts respondents to think about contact with known or suspected undocumented immigrants in one’s daily life, while the data relied upon by Ellison, Shin, and Leal and McLaren ask specifically about Latino or immigrant friends. This arguably speaks to the distinction between “casual contact” and more meaningful “acquaintance” in Allport’s (1954) original articulation of contact theory. The measure of contact provided by the Pew data thus may not tap the kinds of contact necessary to influence immigration policy attitudes.4

Regarding the effects of the contextual variables (county Hispanic percentage, Hispanic percentage-point change, and proximity to the US–Mexico border) in model 1, no significant results emerged, though it is important to recall that no specific expectations about their simple main effects were advanced. Still, it appears that the significant effects of being college educated and holding a positive view of the direction of the national economy point to the merit of the liberalizing education and labor-market competition hypotheses.

It may still be the case, however, that the effect of partisanship is conditional upon ethnic or spatial context. Models 2 to 4 test the expectation that different contexts influence Democrats and Republicans differently; namely, they test separate interactions between party identification and county-level Hispanic population percentage, Hispanic population percentage-point change, and proximity to the US–Mexico border (corresponding to H3, H4, and H5, respectively). The results confirm that an interaction is present in each instance, with each interaction effect statistically significant (using a likelihood ratio chi-square test comparing each of models 2 to 4 to the main-effects-only model 1). Further, the statistically significant coefficients for the product terms between Republican Party identification and Hispanic percentage (model 2), Hispanic percentage-point change (model 3), and distance to the US–Mexico border (model 4) imply that the attitudes of Republican Party identifiers toward illegal immigration are influenced by ethnic and spatial context differently than Democrats. The results therefore confirm H3, H4, and H5. They also reconfirm the finding of heterogeneous effects of context on Democrats’ and Republicans’ immigration policy attitudes (Hawley 2011; Hopkins 2014).

To gain a better substantive understanding of these interactions, predicted probabilities for the “stay in the country legally” response can be plotted for the separate categories of party identification (the focal independent variable) while varying the values of the moderator variables, county Hispanic percentage, county Hispanic percentage-point change, and distance to the US–Mexico border. Given a categorical focal independent variable and a continuous moderator variable, this is the strategy recommended by Jaccard and Dodge (2004). How these predicted probabilities vary when values of the moderator variables range from low values to high values gives us insight into the interaction effects. I thus generate predicted probabilities for Democrats and Republicans separately, setting all other variables at their means (or reference categories), setting the census region to the West, and then varying the values of the contextual and spatial factors in models 2 to 4. These plots provide visual confirmation that the effects of ethnic (Hispanic) context on non-Hispanic white Democrats and Republicans differ markedly. As figure 1 (based on model 2) indicates, a Democrat in a county with a five-percent Hispanic population has a predicted probability of supporting undocumented immigrants staying in the United States of 0.79 while a Republican has a predicted probability of 0.73. In a county with a ten-percent Hispanic population, the predicted probabilities for Democrats and Republicans are 0.83 and 0.73, respectively. Further increasing the county Hispanic population to 20 or even 50 percent puts the predicted probabilities for Democrats and Republicans at 0.87 and 0.72, and 0.91 and 0.72, respectively. Thus, increasing prevalence of Hispanics at the county level creates more polarized opinions about illegal immigration among non-Hispanic white partisans.

Figure 1.

Figure 1.

Open in a new tab

Effect Plot, Party Identification, and County Hispanic Percentage.

Looking at the interaction between party identification and Hispanic population percentage-point change in model 3 (and plotted in figure 2), it is clear that a similar dynamic exists in that progressively larger changes in the county-level ethnic context produce ever greater partisan cleavages. A Democrat in a county that experienced a two-percentage-point increase in the Hispanic population between 2000 and 2010 has a predicted probability of supporting undocumented immigrants staying in the United States of 0.79, while a Republican has a predicted probability of 0.75. For a county that experienced a four-percentage-point increase in the Hispanic population, the respective predicted probabilities for Democrats and Republicans are 0.83 and 0.71. Considering counties that experienced either a six- or eight-percentage-point increase, the predicted probabilities for Democrats and Republicans are 0.86 and 0.69, and 0.87 and 0.65, respectively. These results suggest that change in ethnic context—and the fear associated with change—has a greater moderating effect than ethnic context as such.

Figure 2.

Figure 2.

Open in a new tab

Effect Plot, Party Identification, and County Hispanic Percentage Point Change (2000–2010).

In the case of spatial context in model 4 (plotted in figure 3), proximity to the US–Mexico border similarly acts to amplify partisan differences (or, conversely, distance acts to mute partisan differences). To illustrate, a Democrat 50 kilometers from the US-Mexico border has a predicted probability of supporting undocumented immigrants staying in the United States of 0.95 while a Republican has a predicted probability of 0.54. Moving to 200 kilometers, the predicted probabilities for Democrats and Republicans are 0.91 and 0.63, respectively. Moving further still to 1,000 or even 2,000 kilometers from the border, the predicted probabilities for Democrats and Republicans are 0.83 and 0.72, and 0.79 and 0.75, respectively.

Figure 3.

Figure 3.

Open in a new tab

Effect Plot, Party Identification, and Distance to the US–Mexico Border.

Further insight into these interactive relationships can be gained by repeating the regression analysis and re-centering the contextual variables at different values to see how the coefficients for party identification (and their confidence intervals) change. These methods reveal the range of values of the moderator variables over which the effect of party identification holds (Jaccard 2001; Braumoeller 2004; Brambor, Clark, and Golder 2006). Probing the party identification × county Hispanic percentage interaction in this way reveals that the effect of Republican Party identification is significant only in counties with a Hispanic population of 6.3 percent or greater (see figure 4 in the supplemental online appendix). That is, in counties where the Hispanic population is a relatively small proportion—roughly one in 16 people, or fewer—party identification plays no role in attitudes toward permitting undocumented immigrants to remain in the United States. It is only in areas with proportionately larger Hispanic populations that party identification becomes consequential, producing an effect on policy preference, where Republicans are less likely to favor allowing undocumented immigrants to stay in the country.

Probing the party identification × county Hispanic percentage-point change interaction reveals that the effect of Republican identification is significant only in counties that have seen a 2.8 percent or greater increase in the Hispanic population (see figure 6 in the online appendix). Increases of this magnitude were observed in 801 of 3,143 (or 25.5 percent) of the counties in the United States between the 2000 and 2010 decennial censuses. Thus, it is only in counties with a marked growth in the proportion of Hispanics that Republicans exhibit more restrictionist attitudes than Democrats. In those areas where the relative size of the Hispanic population is unchanged over time (or is growing only slowly), all else equal, Republicans and Democrats are statistically indistinguishable from each other.5

As for the party identification × US–Mexico border proximity interaction, probing reveals that the partisan difference between Democrats and Republicans is significant for distances less than or equal to 1,481 kilometers – somewhat less than the weighted sample mean of 1,660 kilometers, and roughly the distance from northern California to Tijuana, Mexico, or Missouri to Nuevo Laredo, Mexico (see figure 8 in the online appendix). What this implies, then, is that proximity to the US–Mexico border widens the differences in opinions between non-Hispanic white partisans over a large expanse of the United States, but not among those in the densely populated and geographically distant Northeast. Nevertheless, ethnic context (and change in ethnic context) may still be pertinent in these parts of the United States.

Though little theoretical basis exists for expecting self-reported contact to exert different effects on Democrats and Republicans, one may still test an interaction between reported contact with undocumented immigrants and party identification as an empirical exercise. Such an interaction, however, turns out to be non-significant. Consequently, neither a simple main effect of contact nor a more nuanced moderated effect involving party identification exists. (These results are reported as model 5 in table A1 in the online appendix.)

Given the focus on contact, demographic and spatial contexts, and partisan identification in these analyses, a comment is warranted on self-selection bias. It may be the case that individuals select into specific demographic contexts or situations where particular types of contact are expected based on their political (or other) characteristics (Pettigrew and Tropp 2006). A scenario in line with this argument is that Republicans select out of ethnoracially diverse areas, and conversely, Democrats select into such areas. Similarly, Republicans may be more likely than Democrats to view any Hispanic they encounter as a possible undocumented immigrant. While outwardly plausible, tests of such residential selection processes in previous research on racial and immigration attitudes find little support. For example, Ha (2008) finds that racial prejudice fails to predict whites’ proximity to minority groups. Oliver and Wong (2003) similarly find that the effect of demographic context on anti-Latino affect remains robust to the inclusion of a variable capturing preferences for living in a majority-white neighborhood in the model. Prejudice thus fails to explain the relationship between area demographic composition and anti-Latino affect. Most convincingly, Hopkins (2010) uses panel data to address the limitations of cross-sectional data, and similarly concludes that the direction of the relationship is from local context to individual immigration attitudes, and not the reverse.

As further confirmation of this point, supplementary analyses regress contact with undocumented immigrants and the contextual variables on party identification, ideology, and demographic controls (see tables A2 and A3 in the online appendix). For the residential self-selection argument to have support, party identification (and possibly also ideology) ought to have significant effects on these measures of contact and context. The results fail to demonstrate any consistent relationships between the political variables and measures of contact and context. The effect of party identification is not significant in any of these models, and only modest effects for conservative ideological self-placement occur in the models for county-level Hispanic percent and distance to the US–Mexico border. Underwhelming results such as these cast further doubt on the residential self-selection argument.

Conclusion

This article began by asking whether contact or context mattered more in shaping Americans’ attitudes toward illegal immigration. The results presented here point in favor of context over contact. Having contact with known or suspected undocumented immigrants did not have a significant effect on attitudes toward illegal immigration. One should not be too quick, though, to take this as a finding contrary to intergroup contact theory, as the measure of contact employed here may not adequately capture the kinds of close personal contact and friendship that the proponents of intergroup contact theory stipulate are required to reduce outgroup prejudice (Allport 1954; Amir 1969; Pettigrew 1998). Further research should endeavor to test the effects of more refined measures of intergroup contact on immigration policy attitudes while simultaneously testing the effects of racial/ethnic and spatial context.

The findings of this article do make a compelling case for the relevance of context—or rather, contexts—and also for conceptualizing context as more than a mere proxy or substitute measure of contact. Individual-level preferences over whether undocumented immigrants should be permitted to remain in the United States or compelled to leave bear the imprint of local ethnic context and the change in this context that occurs over time. They also bear the imprint of spatial context, namely proximity to Mexico, the primary source of immigrants to the United States. Further, these findings show that different contexts have important roles in amplifying (or muting) the effect of party identification. The varied results of previous research in testing the effects of different contextual measures may be due to the failure to consider that not all individuals experience a given context in the same way. A Democrat and a Republican in a high-density Hispanic area, an area experiencing rapid demographic change with an influx of Hispanics, or a locale close to the US–Mexico border will perceive their milieu differently, and their attitudes about how the United States ought to deal with the challenges posed by undocumented immigrants will differ as a consequence.

While this article presents evidence that a number of spatial and contextual factors moderate the effect of party identification on attitudes toward illegal immigration, further research is needed to clarify the mechanisms through which these factors operate. Local ethnic context, change in ethnic context, and proximity to the US–Mexico border may be linked to attitudes toward illegal immigration through other factors, such as passive language exposure (Enos 2014; Hopkins, Tran, and Williamson 2014), exposure to negative media content on illegal immigration (Dunaway, Branton, and Abrajano 2010), and feelings of cultural threat (Newman, Hartman, and Taber 2012; Johnston, Newman, and Velez 2015). Other possible paths through which spatial and contextual variables might influence immigration attitudes are local discussion networks (Huckfeldt and Sprague 1995) and perceptions of local group interests (Cutler 2007). Designing survey instruments to test such possible mediated relationships will further advance our understanding of Americans’ attitudes toward immigration issues.

Finally, it is important to note that my analyses do not exhaust the ways in which context might be conceptualized and measured. Specifically, in testing the effects of demographic context at the county level, it is important to remain open to the possibility of other, perhaps countervailing dynamics operating at an even more local level (Baybeck 2008). Further, static measures of demographic context (at the county level, census tract level, or otherwise) may not adequately capture how individuals move through space (e.g., by commuting from their homes to their places of work or study) and the contexts they experience as a result. Individualized and “dynamic” measures of local context that account for (and measure) such movements are another methodological advance that ought to be pursued (see Moore and Reeves [forthcoming]). In short, numerous avenues of inquiry relating to the politics of immigration in the United States and the interactive dynamics of individual political characteristics, spatial context, and proximity remain to be explored.

Supplementary Data

Supplementary data are freely available online at http://poq.oxfordjournals.org/.

Supplementary Data
supp_80_1_1__index.html (775B, html)

Appendix. Descriptive Statistics

Mean Min Max SD N Missing
Should be way for undocumented immigrants to stay 0.67 0 1 0.47 1064 22
Male 0.48 0 1 0.50 1086 0
Age (ln years) 3.82 2.89 4.57 0.41 1074 12
Education: College 0.30 0 1 0.46 1083 3
Economy: Getting better 0.25 0 1 0.43 1067 19
Economy: About the same 0.51 0 1 0.50 1067 19
Economy: Getting worse 0.24 0 1 0.43 1067 19
Party: Democrat 0.39 0 1 0.49 1081 5
Party: Republican 0.51 0 1 0.50 1081 5
Party: Independent 0.10 0 1 0.31 1081 5
Ideology: Conservative 0.39 0 1 0.49 1062 24
Ideology: Moderate 0.41 0 1 0.49 1062 24
Ideology: Liberal 0.20 0 1 0.40 1062 24
Contact with undocumented immigrants 0.30 0 1 0.46 1067 19
County Hispanic % (2010) 11.81 0.44 79.58 12.57 1086 0
ln county Hispanic % (2010) 1.93 –0.83 4.38 1.09 1086 0
County Hispanic % point change (2000–2010) 3.18 –3.52 16.06 2.43 1086 0
ln county Hispanic % point change (2000–2010) (+5) 2.06 0.39 3.05 0.28 1086 0
Distance to US–Mexico border (km) 1659.78 23.45 3375.41 763.95 1086 0
ln distance to US–Mexico border (km) 7.22 3.15 8.12 0.76 1086 0
Open in a new tab

Appendix. Data Coding

Stay: “Which comes closer to your view about how to handle undocumented immigrants who are now living in the US? They should not be allowed to stay in the country legally (0), [or] there should be a way for them to stay in the country legally, if certain requirements are met (1).”

Sex: male (1), female (0).

Age: years logged (loge), mean centered.

Education: College (1), less than college (0).

Region: dummy-coded with indicators for the Midwest, South, and West Census Regions; Northeast is the reference category.

Economy: “A year from now, do you expect that economic conditions in the country as a whole will be better than they are at present, or worse, or just about the same as now?” (“about the same as now” is the reference category).

Party: “In politics TODAY, do you consider yourself a Republican, Democrat, or independent?” [If Independent/No preference/Other party/Don’t know/Refused] “As of today do you lean more to the Republican Party or more to the Democratic Party?”; dummy-coded with indicators for Independents and Republicans (including Republican leaners); Democrats (and Democratic leaners) are the reference category.

Ideology: “In general, would you describe your political views as...Very conservative, Conservative, Moderate, Liberal, or Very liberal?”; dummy-coded with indicators for moderates and conservatives; liberals are the reference category.

Contact with Undocumented Immigrants: “Thinking about your daily life, do you have personal contact with any recent immigrants who you know for a fact (1), or who you suspect (1), are in the United States illegally, or not? (0).”

ln County Hispanic %: County-level percentage of the population that is Hispanic (2010 Census) logged (loge), mean centered.

ln County Hispanic % point change: County-level percentage point change in the Hispanic population (2010 Census – 2000 Census) logged (loge), +5 and mean centered.

ln Distance to US–Mexico Border: kilometers logged (loge), mean centered.

Note on geocoding and distance calculations:

ZIP code geocoding was conducted using the GEOCODE procedure in the SAS statistical package and the July 2013 ZIP code lookup data set provided by the SAS Institute. County FIPS code-based geocoding was performed by matching county codes from the telephone sample files to county centroids created from the 2013 US Census Bureau TIGER/Line county shapefile.

For the US–Mexico frontier line, I use the line segment of the US–Mexico border subsetted from the Large Scale International Boundaries (LSIB), Africa and Americas shapefile obtained from the US Department of State, Humanitarian Information Unit (hiu.state.gov/data/). I then perform a Cartesian join between the survey data and the border line segment, and then select the minimum distance for each respondent.

Footnotes

1

A comment is warranted on how contextual variables ought to be scaled. Most research simply uses the percentage of a racial or ethnic group within a particular area (see, e.g., Citrin, Reingold, and Green 1990; Hood and Morris 1997; Branton et al. 2007; Schildkraut 2011), or the percentage-point change in area ethnic composition over time (Hopkins 2010, 2011; Newman 2013; Newman and Velez 2014). This reflects the assumption that the relationship between the percentage (or percentage-point change) of a group in a given area and individual attitudes is linear. By contrast, classic work in the psychology of sensory perception (psychophysics) points to a logarithmic relationship. “Fechner’s law” states that the perceived magnitude of a stimulus is a logarithmic function of its physical magnitude (Laming 2011; Leshner and Pfaff 2011). Modeling a logarithmic relationship has the further intuitive appeal that an increase in the local Hispanic population from, say, 5 to 10 percent registers as a larger increase than a nominally identical increase from 40 to 45 percent.

2

As with ethnic context, a similar argument can be made in favor of a logarithmic transformation of geodetic distances (such as distance to the US–Mexico border). The idea of a “distance decay function” has a long lineage in geography (Taylor 1971). Further, research on policy attitudes in several substantive domains finds that the effect of distance to salient geographic features follows a logarithmic (and not a linear) trend (Berezin and Díez Medrano 2008; Gravelle 2014a, 2014b; Gravelle and Lachapelle 2015).

3

The multiple imputations were conducted using the IVEware 0.1 add-in program for SAS (Raghunathan et al. 2001). The logit models were fit using SAS PROC SURVEYLOGISTIC, with results then submitted to SAS PROC MIANALYZE to combine the imputations.

4

It is worth noting that these results are not due to collinearity between self-reported contact and the contextual measures. The bivariate (Pearson) correlations between contact and county-level Hispanic percentage (r = 0.17), county-level Hispanic percentage-point change (r = 0.13), and logged distance to the US–Mexico border (r = –0.16) are weak.

5

The results also indicate that Republicans are more (not less) likely to favor a policy of allowing undocumented immigrants to remain in the United States when the percentage-point change is –2.3 or less. This result, however, is very likely a statistical artefact, as such a decrease describes only 9 of 3,143 (or 0.3 percent) of counties in the United States, and only a single county contained in the Pew data—Arlington County, Virginia.

References

  1. Allison Paul D. 2001. Missing Data. Thousand Oaks, CA: Sage. [Google Scholar]
  2. Allport Gordon W. 1954. The Nature of Prejudice. New York: Doubleday. [Google Scholar]
  3. Amir Yehuda. 1969. “Contact Hypothesis in Ethnic Relations.” Psychological Bulletin 71:319–42. [DOI] [PubMed] [Google Scholar]
  4. Ayers John W., Hofstetter C. Richard, Schnakenberg Keith, Kolody Bohdan. 2009. “Is Immigration a Racial Issue? Anglo Attitudes on Immigration Policies in a Border County.” Social Science Quarterly 90:593–610. [Google Scholar]
  5. Baker Bryan, Rytina Nancy. 2013. Estimates of the Unauthorized Immigrant Population Residing in the United States: January 2012. Washington, DC: Department of Homeland Security, Office of Immigration Statistics. [Google Scholar]
  6. Baybeck Brady. 2008. “Sorting Out the Competing Effects of Racial Context.” Journal of Politics 68:386–96. [Google Scholar]
  7. Berezin Mabel, Medrano Juan Díez. 2008. “Distance Matters: Place, Political Legitimacy, and Popular Support for European Integration.” Comparative European Politics 6:1–32. [Google Scholar]
  8. Berg Justin Allen. 2009. “Core Networks and Whites’ Attitudes Toward Immigrants and Immigration Policy.” Public Opinion Quarterly 73:7–31. [Google Scholar]
  9. Blalock Hubert M. 1956. “Economic Discrimination and Negro Increase.” American Sociological Review 21:584–88. [Google Scholar]
  10. ———. 1957. “Per Cent Non-White and Discrimination in the South.” American Sociological Review 22:677–82. [Google Scholar]
  11. ———. 1966. Toward a Theory of Minority Group Relations. New York: Wiley. [Google Scholar]
  12. Brambor Thomas, Clark William Roberts, Golder Matt. 2006. “Understanding Interaction Models: Improving Empirical Analyses.” Political Analysis 14:63–82. [Google Scholar]
  13. Branton Regina P., Dillingham Gavin, Dunaway Johanna, Miller Beth. 2007. “Anglo Voting on Nativist Ballot Initiatives: The Partisan Impact of Spatial Proximity to the US–Mexico Border.” Social Science Quarterly 88:882–97. [Google Scholar]
  14. Branton Regina P., Dunaway Johanna. 2009a. “Spatial Proximity to the US–Mexico Border and Newspaper Coverage of Immigration Issues.” Political Research Quarterly 62:289–302. [Google Scholar]
  15. ———. 2009b. “Slanted Newspaper Coverage of Immigration: The Importance of Economics and Geography.” Policy Studies Journal 37:257–73. [Google Scholar]
  16. Branton Regina P., Jones Bradford S. 2005. “Reexamining Racial Attitudes: The Conditional Relationship Between Diversity and Socioeconomic Environment.” American Journal of Political Science 49:359–72. [Google Scholar]
  17. Braumoeller Bear F. 2004. “Hypothesis Testing and Multiplicative Interaction Terms.” International Organization 58:807–20. [Google Scholar]
  18. Campbell Andrea Louise, Wong Cara, Citrin Jack. 2006. “‘Racial Threat,’ Partisan Climate, and Direct Democracy: Contextual Effects in Three California Initiatives.” Political Behavior 28:129–50. [Google Scholar]
  19. Citrin Jack, Green Donald P., Muste Christopher P., Wong Cara. 1997. “Public Opinion Toward Immigration Reform: The Role of Economic Motivations.” Journal of Politics 59:858–81. [Google Scholar]
  20. Citrin Jack, Reingold Beth, Green Donald P. 1990. “American Identity and the Politics of Ethnic Change.” Journal of Politics 52:1124–1154. [Google Scholar]
  21. Cutler Fred. 2007. “Context and Attitude Formation: Social Interaction, Default Information, or Local Interests?” Political Geography 26:575–600. [Google Scholar]
  22. Dunaway Johanna, Branton Regina P., Abrajano Marisa A. 2010. “Agenda Setting, Public Opinion, and the Issue of Immigration Reform.” Social Science Quarterly 91:359–78. [Google Scholar]
  23. Ellison Christopher G., Shin Heeju, Leal David L. 2011. “The Contact Hypothesis and Attitudes Toward Latinos in the United States.” Social Science Quarterly 92:938–58. [Google Scholar]
  24. Enos Ryan D. 2014. “Causal Effect of Intergroup Contact on Exclusionary Attitudes.” Proceedings of the National Academy of Sciences of the United States of America 111:3699–3704. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Espenshade Thomas J., Calhoun Charles A. 1993. “An Analysis of Public Opinion toward Undocumented Immigration.” Population Research and Policy Review 12:189–224. [Google Scholar]
  26. Fetzer Joel S. 2000. Public Attitudes toward Immigration in the United States, France, and Germany. Cambridge: Cambridge University Press. [Google Scholar]
  27. Fischhoff Baruch, Gonzalez Roxana M., Small Deborah A., Lerner Jennifer S. 2003. “Judged Terror Risk and Proximity to the World Trade Center.” Journal of Risk and Uncertainty 26:137–51. [Google Scholar]
  28. Forbes H. D. 1997. Ethnic Conflict: Commerce, Culture, and the Contact Hypothesis. New Haven, CT: Yale University Press. [Google Scholar]
  29. Glaser James M. 1994. “Back to the Black Belt: Racial Environment and White Racial Attitudes in the South.” Journal of Politics 56:21–41. [Google Scholar]
  30. Goldstein Judith L., Peters Margaret E. 2014. “Nativism or Economic Threat: Attitudes Toward Immigrants During the Great Recession.” International Interactions 40:376–401. [Google Scholar]
  31. Gravelle Timothy B. 2014a. “Partisanship, Border Proximity, and Canadian Attitudes toward North American Integration.” International Journal of Public Opinion Research 26:453–74. [Google Scholar]
  32. ———. 2014b. “Love Thy Neighbo(u)r? Political Attitudes, Proximity and the Mutual Perceptions of the Canadian and American Publics.” Canadian Journal of Political Science 47:135–57. [Google Scholar]
  33. Gravelle Timothy B., Erick Lachapelle. 2015. “Politics, Proximity and the Pipeline: Mapping Public Attitudes toward Keystone XL.” Energy Policy 83:99–108. [Google Scholar]
  34. Ha Shang E. 2008. “The Consequences of Multiracial Contexts on Public Attitudes toward Immigration.” Political Research Quarterly 63:29–42. [Google Scholar]
  35. Hainmueller Jens, Hiscox Michael J. 2007. “Educated Preferences: Explaining Attitudes Toward Immigration in Europe.” International Organization 61:399–442. [Google Scholar]
  36. Hainmueller Jens, Hopkins Daniel J. 2014. “Public Attitudes Toward Immigration.” Annual Review of Political Science 17:225–49. [Google Scholar]
  37. Hajnal Zoltan, Rivera Michael U. 2014. “Immigration, Latinos, and White Partisan Politics: The New Democratic Defection.” American Journal of Political Science 58:773–89. [Google Scholar]
  38. Harrison Chase H. 2006. “It’s How You Know Who You Know: The Impact of Interpersonal Contact on American Attitudes Toward Immigration.” PhD diss., University of Connecticut, Storrs. [Google Scholar]
  39. Haubert Jeannie, Fussell Elizabeth. 2006. “Explaining Pro-Immigrant Sentiment in the US: Social Class, Cosmopolitanism, and Perceptions of Immigrants.” International Migration Review 40:489–507. [Google Scholar]
  40. Hawley George. 2011. “Political Threat and Immigration: Party Identification, Demographic Context, and Immigration Policy Preference.” Social Science Quarterly 92:404–22. [Google Scholar]
  41. Hero Rodney E., Tolbert Caroline J. 1996. “A Racial Ethnic Diversity Interpretation of Politics and Policy in the States of the US.” American Journal of Political Science 40:851–71. [Google Scholar]
  42. Hood M. V., Morris Irwin L. 1997. “¿Amigo O Enemigo?: Context, Attitudes, and Anglo Public Opinion toward Immigration.” Social Science Quarterly 78:309–23. [Google Scholar]
  43. ———. 1998. “Give Us Your Tired, Your Poor...But Make Sure They Have a Green Card: The Effects of Documented and Undocumented Migrant Context on Anglo Opinion toward Immigration.” Political Behavior 20:1–15. [Google Scholar]
  44. ———. 2000. “Brother, Can You Spare a Dime? Racial/Ethnic Context and the Anglo Vote on Proposition 187.” Social Science Quarterly 81:194–206. [Google Scholar]
  45. Hopkins Daniel J. 2010. “Politicized Places: Explaining Where and When Immigrants Provoke Local Opposition.” American Political Science Review 104:40–60. [Google Scholar]
  46. ———. 2011. “National Debates, Local Responses: The Origins of Local Concern about Immigration in Britain and the United States.” British Journal of Political Science 41:499–524. [Google Scholar]
  47. ———. 2014. “One Language, Two Meanings: Partisanship and Responses to Spanish.” Political Communication 31:421–45. [Google Scholar]
  48. Hopkins Daniel J., Tran Van C., Williamson Abigail Fisher. 2014. “See No Spanish: Language, Local Context, and Attitudes Toward Immigration.” Politics, Groups, and Identities 2:35–51. [Google Scholar]
  49. Huckfeldt Robert, Sprague John. 1995. Citizens, Politics, and Social Communication: Information and Influence in an Election Campaign. Cambridge: Cambridge University Press. [Google Scholar]
  50. Jaccard James J. 2001. Interaction Effects in Logistic Regression. Thousand Oaks, CA: Sage. [Google Scholar]
  51. Jaccard James J., Dodge Tonya. 2004. “Analyzing Contingent Effects in Regression Models.” In The Handbook of Data Analysis, edited by Hardy Melissa and Bryman Alan, 237–57. Thousand Oaks, CA: Sage. [Google Scholar]
  52. Johnston Christopher D., Newman Benjamin J., Velez Yamil. 2015. “Ethnic Change, Personality, and Polarization over Immigration in the American Public. “ Public Opinion Quarterly 79:662–86. [Google Scholar]
  53. Key V. O. 1949. Southern Politics in State and Nation. New York: Vintage. [Google Scholar]
  54. Knoll Benjamin R., Redlawsk David P., Sanborn Howard. 2010. “Framing Labels and Immigration Policy Attitudes in the Iowa Caucuses: ‘Trying to Out-Tancredo Tancredo.’” Political Behavior 33:433–54. [Google Scholar]
  55. Laming Donald. 2011. “Fechner’s Law: Where Does the Log Transform Come From?” In Fechner’s Legacy in Psychology: 150 Years of Elementary Psychophysics, edited by Solomon Joshua A., 7–23. Leiden: Brill. [Google Scholar]
  56. Lawlor Andrea. 2015. “Local and National Accounts of Immigration Framing in a Cross-National Perspective.” Journal of Ethnic and Migration Studies 41:918–41. [Google Scholar]
  57. Leshner Alan, Pfaff Donald W. 2011. “Quantification of Behavior.” Proceedings of the National Academy of Sciences of the United States of America 108:15537–15541. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Little Roderick J. A., Rubin Donald B. 2002. Statistical Analysis with Missing Data, 2nd ed. New York: Wiley. [Google Scholar]
  59. Malhotra Neil, Margalit Yotam, Hyunjung Mo Cecilia. 2013. “Economic Explanations for Opposition to Immigration: Distinguishing between Prevalence and Conditional Impact.” American Journal of Political Science 57:391–410. [Google Scholar]
  60. McLaren Lauren M. 2003. “Anti-Immigrant Prejudice in Europe: Contact, Threat Perception, and Preferences for the Exclusion of Migrants.” Social Forces 81:909–36. [Google Scholar]
  61. Moore Ryan T., Reeves Andrew. Forthcoming. “Learning from Place in the Era of Geolocation.” In Analytics, Policy, and Governance, edited by Bachner Jennifer, Wagner Hill Kathryn, Ginsberg Benjamin. New Haven, CT: Yale University Press. [Google Scholar]
  62. Muste Christopher P. 2013. “The Dynamics of Immigration Opinion in the United States, 1992–2012.” Public Opinion Quarterly 77:398–416. [Google Scholar]
  63. Newman Benjamin J. 2013. “Acculturating Contexts and Anglo Opposition to Immigration in the United States.” American Journal of Political Science 57:374–90. [Google Scholar]
  64. ———. 2015. “Unfamiliar Others: Contact with Unassimilated Immigrants and Public Support for Restrictive Immigration Policy.” International Journal of Public Opinion Research 27:197–219. [Google Scholar]
  65. Newman Benjamin J., Hartman Todd K., Taber Charles S. 2012. “Foreign Language Exposure, Cultural Threat, and Opposition to Immigration.” Political Psychology 33:635–57. [Google Scholar]
  66. Newman Benjamin J., Velez Yamil. 2014. “Group Size versus Change? Assessing Americans’ Perception of Local Immigration.” Political Research Quarterly 67:293–303. [Google Scholar]
  67. Oliver J. Eric, Wong Janelle. 2003. “Intergroup Prejudice in Multiethnic Settings.” American Journal of Political Science 47:567. [Google Scholar]
  68. Passel Jeffrey S., Cohn D’Vera, Krogstad Jens Manuel, Gonzalez-Barrera Ana. 2014. As Growth Stalls, Unauthorized Immigrant Population Becomes More Settled. Washington, DC: Pew Research Center. [Google Scholar]
  69. Pettigrew Thomas F. 1998. “Intergroup Contact Theory.” Annual Review of Psychology 49:65–85. [DOI] [PubMed] [Google Scholar]
  70. Pettigrew Thomas F., Tropp Linda R. 2006. “A Meta-Analytic Test of Intergroup Contact Theory.” Journal of Personality and Social Psychology 90:751–83. [DOI] [PubMed] [Google Scholar]
  71. Pettigrew Thomas F., Tropp Linda R., Wagner Ulrich, Christ Oliver. 2011. “Recent Advances in Intergroup Contact Theory.” International Journal of Intercultural Relations 35:271–80. [Google Scholar]
  72. Raghunathan Trivellore E., Lepkowski James M., Hoewyk John Van, Solenberger Peter. 2001. “A Multivariate Technique for Multiply Imputing Missing Values Using a Sequence of Regression Models.” Survey Methodology 27:85–95. [Google Scholar]
  73. Rubin Donald B. 1987. Multiple Imputation for Nonresponse in Surveys. New York: Wiley. [Google Scholar]
  74. Scheve Kenneth F., Slaughter Matthew J. 2001. “Labor Market Competition and Individual Preferences over Immigration Policy.” Review of Economics and Statistics 83:133–45. [Google Scholar]
  75. Schildkraut Deborah J. 2005. Press “One” for English: Language Policy, Public Opinion, and American Identity. Princeton, NJ: Princeton University Press. [Google Scholar]
  76. ———. 2007. “Defining American Identity in the Twenty-First Century: How Much ‘There’ Is There?” Journal of Politics 69:597–615. [Google Scholar]
  77. ———. 2011. Americanism in the Twenty-First Century: Public Opinion in the Age of Immigration. Cambridge: Cambridge University Press. [Google Scholar]
  78. Segovia Francine, Defever Renatta. 2010. “American Public Opinion on Immigrants and Immigration Policy.” Public Opinion Quarterly 74:375–94. [Google Scholar]
  79. Smith Rogers M. 1993. “Beyond Tocqueville, Myrdal, and Hartz: The Multiple Traditions in America.” American Political Science Review 87:549–66. [Google Scholar]
  80. Taylor Peter J. 1971. “Distance Transformation and Distance Decay Functions.” Geographical Analysis 3:221–38. [Google Scholar]
  81. Tolbert Caroline J., Hero Rodney E. 1996. “Race/Ethnicity and Direct Democracy: An Analysis of California’s Illegal Immigration Initiative.” Journal of Politics 58:806–18. [Google Scholar]
  82. Wong Cara J. 2010. Boundaries of Obligation in American Politics. Cambridge: Cambridge University Press. [Google Scholar]
  83. Zaller John R. 1992. The Nature and Origins of Mass Opinion. Cambridge: Cambridge University Press. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Data
supp_80_1_1__index.html (775B, html)
supp_nfv054_poq_15_0084_File005.pdf (497.8KB, pdf)

Articles from Public Opinion Quarterly are provided here courtesy of Oxford University Press

RESOURCES