Divided Loyalties? Understanding Variation in Latino Attitudes Toward Immigration*

DIVIDED LOYALTIES? UNDERSTANDING VARIATION IN LATINOS’ ATTITUDES TOWARD IMMIGRATION

Stella M. Rouse University of Maryland

Betina Cutaia Wilkinson Wake Forest University

James C. Garand Louisiana State University

**Direct all correspondence to: Professor James C. Garand (email: [email protected]), Emogine Pliner Distinguished Professor, Department of Political Science, Louisiana State University, Baton Rouge, Louisiana 70803-5433. The corresponding author will provide all data and coding information to those wishing to replicate this study. An appendix reporting descriptive statistics and principle components analysis results discussed in the paper is available for interested readers by writing the corresponding author. We are grateful to the Pew Hispanic Center for making their data available, and particularly appreciate the efforts of Dulce C. Benavides for her assistance with the data used in this study. The Pew Hispanic Center and the Kaiser Family Foundation bear no responsibility for the interpretations offered or conclusions made based on analysis of the Pew Hispanic Center/Kaiser Family Foundation 2004 National Survey of Latinos, Politics and Civic Engagement data.

DIVIDED LOYALTIES? UNDERSTANDING VARIATION IN LATINOS’ ATTITUDES TOWARD IMMIGRATION

Abstract Objective: In this article, we develop and test a model of competing theoretical explanations of Latinos’

attitudes toward immigration; specifically examining their policy preferences on legal immigration, illegal

immigration, and a proposed policy for dealing with illegal immigrants. We also consider whether Latino

attitudes toward legal and illegal immigration are related and comprise a single coherent structure.

Method: Using data from a 2004 national survey of Latinos, we perform regression, logit, and ordered

logit analyses to examine the determinants of Latinos’ attitudes toward immigration. Results: We highlight

three important findings. First, our results demonstrate “within group” differences in immigration attitudes

among Latinos, based on both national origin and generational status; we find that Mexicans are more

pro-immigration than Latinos from other countries, and, foreign-born Latinos have much more positive

attitudes about immigration than second-generation and third-generation Latinos. Second, we find that

Latino support for various aspects of immigration is primarily a function of ethnic and linguistic identity and

attachment to the American culture, with self interest, contextual variables, and political and demographic

attributes playing a smaller, more specialized role. Finally, we demonstrate that Latino attitudes toward

legal and illegal immigration are highly interrelated. Conclusion: There is a coherent structure underlying

Latinos’ attitudes toward legal immigration, illegal immigration, and a policy option for dealing with illegal

immigrants. Our tests of competing theoretical approaches reveal the importance of national origin and

ethnic attachment and acculturation in explaining differences among Latinos on their attitudes toward

immigration.

The issue of immigration has been at the forefront of American politics during the past several years.

There have been numerous periods throughout American history in which immigration was a salient

issue, and there is little doubt that immigration is a major part of contemporary political debate in the

United States. The rapid growth in the Latino population since the early 1990s, coupled with the

Minuteman movement and the highly-visible pro-immigration marches during 2006, have reinforced the

importance of immigration in the minds of the mass public, the news media, and elected officials. Indeed,

the 2004 and 2006 elections were fought with immigration as a major issue in the battle for control of the

federal and state governments.

Although immigration has regained a prominent place in U.S. political and policy discourse, concerns

about its impact have not been applied equally to all immigrant groups. Rather, immigrants from Mexico

and other Latin American countries have become the major focus of the immigration debate in the United

States. This focus is, in large part, due to the growth of the Latino population. Latinos number almost 42

million, accounting for 14.5% of the total U.S. population (U.S. Census 2005; Pew Hispanic Center

2006b). Fully two-thirds (66%) of the Latino population is a direct result of immigration (Pew Hispanic

Center, 2006a). Further, the Pew Hispanic Center (2006c) estimates that illegal immigrants from Mexico

and other Latin American countries account for 78% of the approximately 12 million illegal immigrants in

the United States. The changes in the U.S. population associated with these patterns have generated

considerable controversy over immigration policy, and whether fair or unfair, Latinos have become the

prominent face of the immigration debate. Given the large-scale increases in the size of the Latino

population in recent years and the connection of Latinos to the immigration issue, it is important to

consider what Latinos think about immigration, whether variations in immigration attitudes exist among

Latino subgroups, and the determinants of these attitudes. The salience of the immigration issue provides

a rich opportunity for scholars to probe the preferences of Latinos and to test various theories that may

explain such preferences. In particular, are Latinos’ attitudes toward immigration responsive to the same

forces that drive the attitudes of non-Latinos on immigration and other issues?

The purpose of this paper is to develop and test a series of models of Latino attitudes toward legal

and illegal immigration. We use data from the 2004 National Survey of Latinos, a Pew Hispanic Center

survey that measures Latino opinions about a wide range of issues, including immigration. These survey

2

data are supplemented by contextual data obtained from the U.S. Census. We have two primary

objectives. First, we explore the determinants of four measures of immigration attitudes, including

attitudes toward legal immigration, perceptions of the benefits and costs of illegal immigration, support for

a plan to permit illegal Latino immigrants to remain in the U.S. permanently with legal status, and a global

measure of immigration attitudes that is based on a factor analysis of these three items. Our models

include several theoretically-relevant clusters of independent variables. Our goal is to test empirically and

compare these theoretical orientations to ascertain which explanations of Latino immigration attitudes

draw the most support. Second, we are interested in the structure of Latinos’ attitudes toward

immigration. Many political commentators and scholars distinguish between attitudes toward legal

immigration and attitudes toward illegal immigration (cf. Goldberg 2007), but whether these two sets of

attitudes represent distinct dimensions or are part of a coherent, structured view toward immigration has

not been resolved. Hence, we consider explicitly the degree to which Latinos differentiate their attitudes

toward legal and illegal immigration. We recognize that interest in the issue of immigration is not limited

to Latinos, but we contend that the attitudes of the largest minority group in the country and the group

most tied to the immigration debate warrants scholarly attention.

LATINO ATTITUDES TOWARD IMMIGRATION: A REVIEW OF COMPETING EXPLANATIONS

What explains variation in Latinos’ attitudes toward immigration? Some theoretical approaches

isolate the effects of Latinos’ attributes that can be differentiated from the attributes of other groups in

American society (e.g. subgroup differences based on country of origin and language); these explanatory

factors may have an effect on Latinos’ immigration attitudes that differ in systematic ways from their

effects on immigration attitudes among other Americans. Further, there are disparities in levels of

acculturation among Latinos, especially involving generational differences; first generation Latinos (i.e.,

those born outside the U.S.) are less likely to be integrated into American society than second and third

generation (native born) Latinos (Leal 2002; Branton 2007; Abrajano and Singh 2009). It is also possible

that the variables driving immigration attitudes in non-Latino populations may have important effects on

immigration attitudes among more incorporated Latinos (e.g. feelings of patriotism and “Americanism”).

Our approach here is to develop a model of Latinos’ attitudes toward immigration that permits us to

explicitly compare empirical support for competing theories that can inform our understanding of

3

immigration attitudes. We consider the effects of several theoretical clusters of independent variables that

are expected to influence Latinos’ immigration attitudes: (1) economic self-interest and labor competition;

(2) subgroup differences based on national origin; (3) measures of ethnic attachment and acculturation

(e.g., generation status, Latino and language identity, attachment to the American culture); (4) political

and demographic attributes; and (5) contextual variables. By including variables representing each of

these approaches in the same model, it is possible to identify those theoretical approaches that are

successful in explaining variation in Latino immigration attitudes, as well as those that are less useful.

Economic Self-Interest

Perhaps the most widely-discussed theoretical approach for framing attitudes toward immigration is

economic self-interest. The central tenet of economic self-interest models is that individuals develop

political attitudes to coincide with their perceptions of the relative costs and benefits associated with the

subject of those attitudes. Self-interest theories suggest that individuals develop policy attitudes based on

whether their material well-being is improved or diminished by the adoption of one policy over another.

Applied to immigration, self-interest theories would predict that Latinos’ immigration preferences will be

based on the degree to which they perceive greater personal benefits than costs associated with

increased immigration.

Scholars have debated whether immigrants take jobs away from American citizens and provide low-

wage labor that suppresses wages (Borjas, 2003, 2004; Altonji and Card, 1991; Butcher and Card, 1991;

Orrenius and Zavodny, 2003). Given this, Latinos who are economically vulnerable (i.e., those who are

unemployed or working in low-wage jobs) and/or who perceive negative personal consequences of

immigration will be more likely to develop negative attitudes toward increased immigration. On the other

hand, Latinos who perceive that they will be made better off economically by immigration will develop

more favorable attitudes toward immigration. There is some support for economic self-interest theories in

research on Latinos’ attitudes toward immigration, though the evidence is mixed (Hood, Morris, and

Shirkey, 1997; Binder et al., 1997; Suro 2005).

National Origin

Latinos are not a homogeneous group. Rather, they are often aggregated into a single classification

due to their common Spanish language and heritage. Although Latinos share a linguistic tradition, there

4

are significant intra-group differences based on cultural and political backgrounds depending on country

of origin. These differences are highlighted in many studies that have analyzed various aspects of Latino

political behavior. Leal (2002) examines the extent of political participation among Latino non-citizens in

the U.S. He finds differences in participation between Mexicans and Cubans, two of the three Latino

groups delineated in the survey. However, Leal notes that comprehensive Latino surveys, which include

respondents of many nationalities, are needed in order to offer more than a baseline comparison of Latino

sub-group differences. Kelly and Kelly (2005) examine the relationship between religion and political

participation among Latinos, paying particular attention to differences in religious tradition among Latino

sub-groups. Branton (2007) also considers country of origin differences among Latinos with respect to

various immigration policies, and Abrajano and Singh (2009) contrast the political awareness and

immigration attitudes of Latino sub-groups. Based on these and other studies, we expect that national

origin will provide some explanatory power for Latinos’ attitudes toward immigration.

Ethnic Attachment and Acculturation

Latino Identity and Attachment to Spanish Language. Individuals differ considerably in terms of their

attachments to groups. Group identification, or feelings of psychological attachment to a particular social

stratum in society, plays a significant role in attitude formation (Conover and Feldman 1984). Scholars

have found that Latinos’ political behavior and attitudes toward issues such as immigration may be

affected by their affinity for and attachment to Latino culture and the Spanish language (Graves and Lee,

2000; Newton, 2000; Branton, 2007). We suggest that Latinos’ attitudes toward immigration are

influenced both by their Latino Identity (cultural affinity) and their attachment to the Spanish language.

The general consensus in the scholarly literature is that Latinos who perceive themselves as strongly

attached to Latino culture and/or the Spanish language are more likely to support expanded immigration

than other Latinos. Scholars suggest that Latinos living in the U.S. who are attached to Latino culture and

the Spanish language are likely to have a cultural affinity for citizens of Mexico and Central and South

American countries, the areas from which most U.S. immigration originates. On the other hand, Latinos

who have weaker attachments to Latino culture and the Spanish language are less likely to be motivated

by cultural affinity and less inclined to support expanded immigration (de la Garza, Garcia, Garcia, and

Falcon, 1993; de la Garza, Polinard, Wrinkle, and Longoria, 1991; Hood et al., 1997; Branton, 2007).

5

Acculturation and American Identity. The degree of acculturation or accommodation (i.e., attachment

to or incorporation into American society) is also thought to affect Latino attitudes.1 One of the primary

concerns raised by many Americans is that the influx of immigrants has a potentially negative effect on

American culture, particularly if these immigrants are not willing to conform to the American way of life

(Huntington 2004). Levels of social and political integration among immigrants are influenced by many

factors. For one, Latino politics scholars often point to a generational effect of acculturation. Generally

speaking, 2nd generation immigrants (i.e., those who have immigrant parents but are born in the U.S.) are

more likely to incorporate than 1st generation immigrants (i.e., those born outside the U.S.). Previous

literature has shown that immigration attitudes are, in part, a function of integration levels; 2nd and 3rd

generation Latinos who are typically more integrated into American society are generally less supportive

of liberal migration policies (Hood et al. 1997; de la Garza 1992; Branton 2007; Abrajano and Singh

2009). Hence we expect that generational status will help explain Latinos’ attitudes toward immigration.

There is some variation about what constitutes acculturation or incorporation, but the literature

generally references several common elements—e.g., voting, support for the English language, support

for American symbols, attachment to being an American. Many Americans express concern about the

perceived diminution of English as the national language, as well as other cultural changes presumed to

occur with the influx of immigrants. These views will characterize some Latinos, and we expect that

greater incorporation into American culture among Latinos will depress support for expanded immigration.

Most of the research that explores levels of acculturation focuses on individual characteristics thought

to be related to strong Americanism. Hood et al. (1997) use the “Cultural Divide” as a metaphor for the

theoretical cluster of variables that focuses on the critical roles that race, ethnicity, and affinity for

American culture play in forming individuals’ opinions about who should or should not be allowed to

immigrate to the U.S. Although they find that Latino identity does not stand out as a predictor of

immigration attitudes, Hood et al. conclude that acculturation—the measurement of which is based in part

on whether respondents or their parents were born in the United States, as well as in part on perceived

importance of learning English—has a highly significant effect on Latinos’ attitudes toward immigration.

As Latinos develop a greater attachment to American society and American cultural traits and practices,

and as they lose some of their Latino identity (i.e. they move from ethnic attachment to incorporation),

6

they are more likely to support restrictive immigration policies. This finding is reinforced by Branton 2007,

who finds a similar relationship between acculturation—i.e., a scale based on generation status and use

of English—and hostility to immigration.

Contact and Context

Individuals do not live in a vacuum; rather, they live in environments in which there is variation in the

degree to which they have regular interaction with Latinos and other immigrant-dominated groups. Most

of the research on the role of contact and context is found in the literature on racial politics, where there is

considerable disagreement about whether and how different racial contexts affect individuals’ political

attitudes and behavior (Amir, 1976; Pettigrew, 1986, 1998; Powers and Ellison, 1995; Giles and Buchner,

1993; Glazer, 1994; Voss, 1996). Among theories relating to Latinos and context and social interactions,

scholars differentiate two sets of processes that can connect individuals and their environments: (1)

contact, which relates to the direct interactions that Latinos have with others in their environments; and

(2) context, which involves more diverse forms of connection between Latinos and their surrounding

culture, such as the observation of the ethnic composition of one’s community, headlines of local news

stories, and context-relevant information learned through communication with friends, neighbors, and

coworkers in one’s social network. To take one example, Abrajano and Singh (2009) find that Latinos who

access their news from Spanish-language sources are more likely to hold pro-immigration sentiments.

For Latinos, living in a community with a large population of fellow Latinos is expected to reinforce

support for the predominant views of the Latino community. Applied to immigration attitudes, we

hypothesize that Latinos who live in local communities with a large Latino population will be more

favorable in their immigration attitudes, while Latinos living in communities with few Latinos and/or

immigrants will, on balance, be less supportive of immigration.

Political Attitudes and Demographic Attributes

Core political attitudes and demographic attributes may also affect attitude formation on the issue of

immigration. Scholars have found that several political and demographic characteristics are related to

Latinos’ attitudes toward immigration. First, party affiliation and ideology are good indicators of policy

preferences. Since Democrats and Republicans tend to hold distinct positions on immigration issues, we

expect that party affiliation, coupled with ideological beliefs, will influence immigration preferences.

7

Second, life-cycle effects play “a significant role in the formation of immigration opinions among

Hispanic-Americans” (Hood et al. 1997: 640). The question is whether age increases or depresses

support for immigration. Hood et al. conclude that, regardless of whether one is a U.S. native or

immigrant, older Latinos are more likely to favor restrictive immigration policies. This may be due to the

idea that older citizens define being a “true American” in narrower terms than younger citizens (Citrin,

Reingold, and Green 1990). On the other hand, Sum, Harrington, and Khatiwada (2005) suggest that

increased immigration has the strongest negative effect on the earnings of younger Americans. From a

psychological/values standpoint, one might expect age to be negatively related to support for increased

immigration; from a self-interest standpoint, one might expect a positive relationship between age and

pro-immigration attitudes.

Third, gender has been found to have a positive effect on support for restrictive immigration policies.

In their work on U.S. public opinion toward immigration policy, Burns and Gimpel (2000) find that women

are more likely to have negative views toward immigration than men. On the other hand, some scholars

conclude that gender has no significant effect on the opinion of Latinos toward immigration (Hood et al,

1997; Binder et al., 1997; Branton, 2007). Women generally tend to be more liberal than men (Welch and

Hibbing 1992; Thomas 1994; Swers 2002), and this might prompt one to speculate that women are more

supportive of immigration than men. Given the conflicting findings and theoretical expectations in the

literature, the role of gender on attitudes toward immigration remains somewhat ambiguous.

Finally, education seems to influence views toward U.S. immigration. Hood et al. (1997: 640) find that

Latinos with higher education levels are less likely to favor legal immigration. They associate education

with economic competition by suggesting that Latinos who are highly educated perceive greater threat

from increasing levels of immigration, since they are likely to link increases in legal immigration with

increases in tax burdens. On the other hand, Borjas (2004) demonstrates that both native and naturalized

Americans with low levels of education are the most vulnerable to income and job displacement

associated with immigration; since many Latino immigrants are at the lower end of the educational scale,

this would suggest that education will be positively related to support for immigration.

8

DATA AND VARIABLES: MODELLING LATINOS’ ATTITUDES TOWARD IMMIGRATION ISSUES

In this section we delineate the contours of our comprehensive model of attitudes toward illegal

immigration, legal immigration, and a specific policy proposal designed to address the issue of

undocumented immigrants. Following the prevailing literature, we consider the effects of economic self-

interest, sub-group differences based on national origin, ethnic attachment and acculturation, political and

demographic control variables, and contextual variables. The data used in this study come from the 2004

National Survey of Latinos: Politics and Civic Participation (NSL), a data set collected by the Pew

Hispanic Center and the Kaiser Family Foundation. This survey of a nationally-representative random

sample of 2,288 Latino adults was conducted by telephone between April 21 and June 9, 2004. The

sample design is based on a highly stratified disproportionate random digit dialing (RDD) sample of the

48 contiguous states, with the results weighted to represent the actual distribution of adults throughout

the United States. The data set includes a wide range of variables relating to immigration, as well as

variables related specifically to Latinos’ attitudes toward immigration.

Dependent Variables

At the outset, we note that there are three dependent variables used in this study, as well as an

overall immigration attitudes scale that combines these three variables. First, we consider Latinos’

attitudes toward illegal immigration. Some studies include items in which respondents are asked whether

they support increases or decreases in illegal immigration, but such an item is not included in the NSL.

Instead, respondents were asked about their perceptions of the effects of illegal immigration on the

economy. This variable is coded +1 for respondents who perceive that illegal immigrants help the

economy by providing low-cost labor, and 0 for respondents who perceive that illegal immigrants hurt the

economy by driving wages downward. Of course there are other, non-economic reasons why an

individual may support or oppose illegal immigration, but in focusing on the economy this item captures

one of the major points of contention relating to illegal immigration.

Second, we use a variable that reflects Latinos’ attitudes toward legal immigration. This variable is

coded +1 for respondents who support increases in legal immigration, 0 for those who would keep legal

immigration at current levels, and -1 for those who support decreases in legal immigration.

9

We also consider Latinos’ preferences that are related to a plan to deal with the legal status of illegal

immigrants. The proposal calls for undocumented Latino immigrants to remain in the U.S. permanently,

with legal status, and then to have the opportunity to become citizens of the United States. This variable

is coded 1 if respondents support the plan, and 0 otherwise.

Economic Self-Interest and Labor Competition

We rely on two sets of variables from the Pew Survey to measure the role played by self-interest in

shaping Latinos’ attitudes toward immigration. For our first set of variables we include measures

representing individuals’ economic vulnerability and security. First, the NSL measures household income

on a scale ranging from 1 (less than $5000 per year) to 26 (greater than $200,000 per year). We suggest

that high-income Latinos are less economically vulnerable and may be less likely to perceive a personal

economic threat from immigration, and they may have a strong self-perception of empowerment that

supports a stronger pro-immigrant stance. On the other hand, low-income Latinos are likely to perceive

themselves as economically vulnerable, and they may consider themselves as being in competition with

other immigrants for available jobs and wages. Hence, we posit that income is positively related to

support for immigration. Second, unemployment is measured as a binary variable, coded 1 for

unemployed respondents, and 0 otherwise. We expect the coefficient for this variable to be negative,

suggesting that unemployed respondents are more likely to perceive economic threat from immigrants

and stake out a position against legal and illegal immigration.

In addition, we include three variables that represent issues that some individuals may deem as

important to the immigration debate: (1) economy and jobs; (2) crime; and (3) immigration. Each of these

variables is coded on a four-point scale, ranging from 0 (not too important) to 3 (extremely important). We

suggest that individuals who perceive the economy/jobs and/or crime as important will be likely to link

immigration to these issues and perceive a broader social vulnerability to increased immigration, and

hence they will be less inclined to support legal immigration, perceive that illegal immigration benefits the

economy, and support policies that relax the immigration status of illegal immigrants. On the other hand,

the expectations for Latinos who perceive immigration to be an important issue are less clear. Latinos can

consider immigration an important issue because they perceive that there is more immigration than

desirable, or they can consider immigration a problem because they support reforms to permit more

10

immigration into the country. Hence the direction of the relationship between the perceived importance of

immigration and support for immigration is ambiguous.

National Origin

As previously discussed, we hypothesize that differences in nationality among Latinos will influence

attitudes about immigration. Unlike other surveys, one of the main benefits of the NSL is that it surveys

Latino respondents from many nationalities, including Mexicans, Puerto Ricans, Cubans, Dominicans

Salvadorans, Central Americans, South Americans, and Spaniards. We include in our model dichotomous

variables for each of these nationalities, with Mexicans as the omitted category.

Ethnic Attachment and Acculturation

Latino Identity and Attachment to Spanish Language. We speculate that individuals who exhibit a

Latino political identity and who are oriented toward the Spanish language will have attitudes that are

more favorable toward both legal and illegal immigration. Thus we include in our models several

measures of Latino and language identity. First, we consider the effects of respondents’ identification with

the Spanish language. The NSL places respondents into one of three categories, based on their

responses to a series of questions about the use of Spanish in written and oral communication: (1)

Spanish dominant; (2) bilingual; and (3) English dominant. We expect Spanish-dominant and (to a lesser

extent) bilingual Latinos to be more favorably oriented toward legal and illegal immigration. In addition,

attachment to the Spanish language is measured using an item asking respondents about how important

it is for “future generations of Latinos living in the United States to speak Spanish;” this variable is coded

on a four point scale, ranging from 0 (not important at all) to 3 (very important). We expect the coefficient

for this variable to be positive, suggesting that individuals who are committed to the Spanish language will

be more likely to favor legal and illegal immigration.

The NSL also includes two items relating to perceptions of discrimination against Latinos, which we

suggest indicate respondents’ level of identification with other Latinos. One item is “sociotropic,” insofar

as it applies to Latinos in general rather than to respondents themselves; this variable is coded 2 if

respondents perceive discrimination against Latinos to be a major problem, 1 if respondents perceive it to

be a minor problem, and 0 if they perceive it not to be a problem. We also include a “personal” measure

of discrimination, indicating whether respondents or a member of their families had experienced

11

discrimination in the past five years. This variable is coded 1 for respondents who report experience with

discrimination, and 0 otherwise. We hypothesize that both variables will be positively related to support

for legal and illegal immigration.

Acculturation and American Identity. We also consider the effects of several measures that represent

varying levels of acculturation into American culture. First, the NSL includes several variables to capture

individuals’ generational and citizenship status. Using these variables, we divide our Latino sample into

three groups denoting each respondents’ generational status: (1) first-generation Latino, coded 1 for

individuals who were born outside of (and immigrated to) the United States; (2) second-generation Latino,

coded 1 for individuals who were native born in the United States but who have at least one parent who

was born in another country; and (3) third-generation Latino, coded 1 for individuals who were native born

in the United States and who had both parents born in the United States. We also include a separate

binary variable to differentiate first-generation Latinos who are naturalized U.S. citizens (coded 1) from

those who are not (coded 0). We posit that Latinos who are second generation or beyond are more likely

to have been incorporated into American culture and are less likely to have as strong an attachment to

Latino immigrants. We include in our model variables for second- and third-generation, as well as the

naturalized citizen variable, with first-generation Latinos the excluded category. We hypothesize that the

coefficients for the three variables representing generational and citizenship status will be negative,

indicating that these three groups are less supportive of immigration than first-generation Latinos.

The NSL also includes a limited number of attitudinal measures that reflect individuals’ attachment to

American culture and values. The NSL includes four dichotomous items that reflect respondents’

perceptions of the responsibilities that immigrants must meet to be a part of American society: (1) speak

English; (2) believe in the U.S. Constitution; (3) be a U.S. citizen; and (4) vote in U.S. elections. We use a

principal components factor analysis (based on the matrix of tetrachoric correlations) to estimate the

degree to which these four items represent a single dimension. All four variables load on this dimension,

so we use factor scores as a measure of this concept, denoted the “American society scale” (eigenvalue

= 2.389; variance explained = 0.597) These standards for incorporation should affect respondents’

perceptions about legal and illegal immigration; specifically, Latinos with strict standards are expected to

be more opposed to both legal and illegal immigration.

12

Third, we include a measure of pro-acculturation attitudes among Latinos. NSL respondents are

asked about their perceptions of how important it is “for Latinos to change so that they blend into the

larger society, as in the idea of a melting pot of cultures.” This seems to capture the concept of

acculturation/incorporation very closely. Responses are coded on a four point scale, ranging from 0 (“not

important at all) to 3 (“very important”). We suggest that pro-acculturation Latinos will be less supportive

of expanded immigration, and hence we hypothesize that the coefficient for this variable will be negative.

County Contextual Variables

Our model contains three contextual variables. First, we include the county Latino population from

2004 (measured as the proportion of total population). Latinos residing in counties with a larger Latino

population are likely to have greater contact with fellow Latinos, and we suggest that this may promote

pro-immigration attitudes. Hence, we hypothesize that Latino population will be positively related to our

immigration variables. Second, we measure change in the proportion of county Latino population from

2000 to 2004. On one hand, this variable could capture the disruption that might occur as a result of rapid

increases in the Latino population. Hence, the coefficient for this variable could be negative, suggesting

that rapid increases in the Latino population is perceived by non-Latinos as threatening to the economic

livelihood and cultural continuity of local communities. On the other, this coefficient could be positive,

suggesting that the rapid increase in the Latino population reinforces positive views toward immigration

among Latinos. Third, we consider the effects of the unemployment rate for respondents’ county of

residence, measured for 2004, the year of the survey. We suggest that Latinos living in communities with

high unemployment are concerned about the negative effects of immigration on their job prospects and

thus will be less supportive of immigration.2

Political Attitudes and Demographic Attributes

We include several political attitudes and demographic attributes as control variables in our model.

First, we measure partisan identification with two dichotomous variables, one for Republican respondents,

the second for Independent respondents. Each variable is coded 1 for respondents in the appropriate

partisan category, and 0 otherwise. Since Democrats are represented in the excluded category, we

expect that the coefficients for these variables will be negative in each model. Second, since the NSL

does not include a global measure of political ideology, we rely on a simple proxy measure. Respondents

13

are asked about their support for expanded government; this variable is coded 1 for respondents who

would “rather pay higher taxes and support larger government with more services,” and 0 for respondents

who would “rather pay lower taxes and support smaller government with fewer services.” We expect

respondents on the more liberal side of the ideological spectrum to support legal immigration and be less

hostile toward illegal immigration, so here again we expect the coefficient for this variable to be positive in

each model. Third, we measure education using an eight-point scale, ranging from 1 (0-8 grades

completed) to 8 (post-graduate training), and given the noted “liberalizing” influence of education (Hood et

al. 1997), we expect this variable to be positively related to our dependent variables.

We also consider the effects of age. As previously discussed, age has been shown to have a varying

impact on attitudes toward immigration depending on whether it is examined through a psychological or

self-interest perspective. Thus, we are left with competing hypotheses about the effect of age on

immigration attitudes among Latinos.

Next, we include a variable for gender. The effect of gender is interesting because scholars often

speak of a gender gap in which women stake out policy positions to the left of men (Thomas, 1994;

Swers, 2002), yet on the immigration issue some research has found that women are less favorable

toward immigration and more supportive of restrictive immigration policies (Burns and Gimpel, 2000). We

measure gender as a binary variable, coded 1 for women and 0 for men; because of ambiguity about the

effects of gender, we do not have a directional expectation for this coefficient.

Finally, we consider the effects of urban residence on attitudes toward immigration. A high

percentage of immigrants reside in urban areas, and this enhances the contact that individual immigrants

will have with other immigrants. Moreover, the potential for creating Latino organizations designed to

bring together and empower Latinos and to reinforce Latino identity is greater in urbanized areas. We

measure urban residence as a five-point scale, ranging from 0 (respondent resides in rural area) to 4

(respondent resides in urban, central city area). We expect the coefficient for this variable to be positive.

EMPIRICAL RESULTS

Support for Illegal Immigration

In surveys of the general population from 2004 to 2006, over 60% of respondents express the belief

that illegal immigrants hurt the U.S. economy (Polling Report, 2007). The views of Latinos on the

14

economic benefits of illegal immigration stand in sharp contrast to those of the general population. Latinos

generally perceive that illegal immigrants have a positive effect on the U.S. economy. Fully 71.9% of NSL

respondents believe that illegal immigrants help the economy, compared to only 28.1% who have a

negative perception of the economic effect of illegal immigrants. What explains variation in Latinos’ beliefs

in these two positions?

In the first model in Table 1 we report logit estimates of our model of respondents’ perceptions that

undocumented immigrants improve the economy. At the outset we note that this model does a

satisfactory job of explaining perceptions regarding the degree to which illegal immigrants improve the

economy. The pseudo-R2 for the model is 0.162, and the goodness-of-fit for the model easily achieves

statistical significance (Wald χ2 =317.52, prob. < 0.0001). Moreover, measures from all five categories of

independent variables are found to have significant effects on the dependent variable.

First, we point out the significant effect of national origin on opinions about the benefits of illegal

immigrants. The coefficients for most of the national origin variables are negative and significant,

suggesting that these groups perceive less favorable effects of illegal immigration than Mexicans, the

excluded group. It appears that Mexicans, more than any other Latino group, drive perceptions that illegal

immigrants help the economy.

Turning to the economic self-interest and labor competition measures, we are surprised to find that

neither unemployment nor income has a significant effect on respondents’ perceptions that illegal

immigrants help the economy. Unemployed individuals or employed individuals in low-wage jobs have the

most to lose from the increased labor force competition and depressed wages associated with illegal

immigration, and we hypothesize that Latinos who are unemployed or have low incomes would be less

likely to perceive a positive benefit to the economy for illegal immigrants. However, the observed

coefficients suggest that there is no link between unemployment and income, on one hand, and

perceptions of economic benefits and costs of illegal immigration, on the other.3 Thus, it seems that the

effects of self- interest on opinions about illegal immigration are limited among Latinos. However, we do

find that Latinos who mention the economy and crime as important issues are significantly less likely to

believe that illegal immigrants help the economy. This suggests more of a social (rather than individual)

concern about the impact of illegal immigrants.

15

In considering the other clusters of independent variables in our model, we find that ethnic attachment

and acculturation measures collectively have the strongest effects on perceptions that illegal immigration

helps the economy. Among the ethnic attachment variables, having Spanish as one’s dominant language

(b = 0.672, t = 3.24), considering it critical for future generations to speak Spanish (b = 0.257, t= 3.49),

and perceiving that discrimination against Latinos is a problem (b = 0.261, t = 3.25) are strongly related to

the dependent variable. Clearly, identifying with Latinos and the Spanish language prompts Latinos to

have more favorable perceptions of the economic benefits provided by illegal immigrants. Among the

acculturation variables, the American Society scale is negatively and significantly related to the

dependent variable (b = -0.274, t = -4.22); Latinos who support higher standards in defining individuals as

“American” are less likely to see positive benefits from illegal immigration than other Latinos. Moreover,

we find strong effects of the generational status variables; both second-generation Latinos (b = -0.631, t =

-2.71) and third-generation Latinos (b = -0.780, t = -2.77) are significantly less favorably oriented toward

illegal immigration than first-generation Latinos. On the other hand, pro-acculturation attitudes and being

a naturalized U.S. citizen are unrelated to opinions about the economic benefits of illegal immigration.

Among our political attitudes and demographic attributes variables, we find that Latinos who support

expanded government perceive that undocumented individuals improve the economy; simply, pro-

government Latinos are more likely to hold favorable attitudes toward illegal immigrants than those who

support a retrenchment of government spending and tax levels. Support for the size of government is a

proxy measure for liberalism, and the coefficient for this variable suggests that more liberal Latinos are

more tolerant of illegal immigration and are favorably oriented toward the perceived benefits that illegal

immigrants bring to the economy. On the other hand, partisanship surprisingly has little effect on

perceptions that illegal immigrants affect the economy. Our findings for the remaining demographic

attribute variables generally point to a weak or null effect. The one exception is the coefficient for gender

(b = -0.394, t = -3.32), which indicates that Latina women are substantially less likely to perceive that

undocumented immigrants help the economy.

Finally, of our three county contextual variables, only Latino population holds any explanatory value.

Contrary to our expectations, Latinos residing in counties with larger Latino populations are less likely to

16

find that illegal immigrants are beneficial to the economy. It appears that greater contact with fellow

Latinos actually leads to lower support for the assertion that illegal immigration helps the economy.

Support for More Legal Immigration

Latinos are mixed in their support for legal immigration, though on balance they favor expanding

(rather than decreasing) levels of legal immigration. In the 2004 NLS, one-half of Latinos support keeping

legal immigration at current levels, while 35.7% support increased levels and only 14.3% support

decreased levels. In the second set of columns in Table 1, we estimate an ordered logit model of

individuals’ support for increased legal immigration. The pseudo-R2 for the model is only 0.055, but the

model does significantly better than chance in predicting support for legal immigration (Wald χ2 = 280.16,

prob. < 0.001). Similar to the results in the model for illegal immigrants, there are telling differences found

in our national origin variables; again, Mexicans appear to drive favorable sentiments about increases in

legal immigration, although only Puerto Ricans and South Americans are significantly less supportive of

higher levels of legal immigration. Further, we find that ethnic attachment and acculturation measures

have strong effects on support for increases in legal immigration. Latinos who believe that future Latino

generations should speak Spanish and those who feel that discrimination against Latinos is a problem are

significantly more supportive than others of increased legal immigration. On the other hand, Latinos who

have strict beliefs about the responsibilities that immigrants must meet to be part of American society

(American society scale) and those who hold pro-acculturation attitudes are significantly less likely to

support increases in illegal immigration. Presumably, Latinos who have such strict beliefs see

immigrants—even legal immigrants—as less likely to live up to the expectations associated with being

part of American society, and this translates into significantly lower levels of support for legal immigration.

Moreover, we find strong generational effects in attitudes about legal immigration. Second-generation

Latinos are moderately (but significantly) less likely than first-generation Latinos to support increases in

legal immigration, while third-generation Latinos are strongly and significantly less supportive of

expanded legal immigration. These findings are consistent with other studies that show a generational

divergence in opinions about support for immigration (Hood et al. 1997; Branton 2007; Abrajano and Sing

2009). Overall, these results indicate that Latinos who have become well integrated into American culture

and society are less supportive of increases in legal immigration than other Latinos.

17

Among the other clusters of independent variables, there are several coefficients that are both

statistically significant and in the expected direction. For the economic self interest and labor competition

variables, perceptions that immigration is an important issue have discernable effects on attitudes toward

legal immigration; Latinos who view immigration as an important issue are more likely to support

increases in legal immigration. Interestingly, the importance of the economy, unemployment, crime, and

household income are unrelated or weakly related to increased support for legal immigration. Overall, it

appears that Latinos’ preferences relating to legal immigration levels are not influenced by economic

perceptions. Among the political attitudes and demographic attribute variables, only gender stands out as

a significant predictor; Latinas are more likely than Latino men to oppose increases in legal immigration,

though the gender coefficient is just outside conventional levels of significance. Finally, none of the

county contextual variables have significant effects on attitudes about legal immigration.

In sum, support for increases in legal immigration among Latinos is largely a function of national

origin, generational status, and variables that tap into levels of ethnic attachment and acculturation.

Mexicans seem to drive support for increases in legal immigration, but overall, Latinos who are second- or

third-generation and/or who are well incorporated into American society are much less willing to support

higher levels of immigration.

Support for Illegal Immigrant Worker Status Program

NSL respondents were asked about whether they support a plan relating to the work status of illegal

immigrants. This plan would allow undocumented Latino immigrants to remain in the U.S. permanently,

with legal status, and then to have the opportunity to become citizens of the United States. This plan

received broad support (90.5%) from NSL respondents. In the third set of columns in Table 1, we report

the logit estimates for our model of support for this proposed plan, labeled illegal immigrant work status

program, applying the same set of independent variables that are found in the other two models. From

the onset, we note that the model is statistically significant (pseudo-R2 = 0.172; Wald χ2 = 324.91, prob. <

0.001), with several clusters of independent variables influencing support for the plan.

Once again we emphasize the consistent effect of national origin. Mexicans appear to drive support

for a plan that would grant illegal immigrants the opportunity to stay in the U.S. permanently. Except for

Central Americans and Salvadorians, all other nationalities/regions are significantly less supportive of

18

such a plan than Mexican respondents. Many of the ethnic attachment and acculturation variables have

significant effects on support for the illegal immigrant work status program. As predicted, Latinos who

think it is critical for future generations to speak Spanish (b = 0.521, t = 4.34), whose dominant language

is Spanish (b = 1.294, t = 3.80) or who are bilingual (b = 0.533, t = 2.43), and who perceive that

discrimination against Latinos is a problem (b = 0.247, t = 1.83) are more supportive of the plan granting

illegal immigrants permanent status in the U.S. Also as expected, Latinos who hold pro-acculturation

attitudes (b = -0.211, t = -1.95) and who believe that immigrants must meet certain conditions to be part of

American society (b = -0.274, t = -2.59) are significantly less likely to support this plan. Of the

generational variables, only the coefficient for third-generation Latinos achieves statistical significance,

indicating that Latinos born in the U.S. and with parents born in the U.S. are less likely to support a plan

granting illegal immigrants permanent U.S. status. The effects of the variables in this category showcase

an interesting divide between affinity for one’s traditional culture and affinity for American culture. In line

with the findings of other scholars (Branton 2007; Abrajano and Singh 2009), Latinos with high levels of

cultural affinity are more likely to favor less restrictive immigration policies, but those who have a strong

attachment to American culture are much less receptive to more liberal immigration policies. There are

only a few variables in the remaining categories of independent measures that significantly influence

Latinos’ attitudes about the proposed plan to grant illegal immigrants permanent status. Respondents

who view the economy as an important issue, who are in favor of an expanded role for government, or

whose residential county has experienced a rapid change in the Latino population are significantly more

likely to support the permanent status plan.

Are Attitudes Toward Legal and Illegal Immigration Linked?

One important question raised by our analyses is whether or not Latinos’ attitudes about legal and

illegal immigration are linked. A significant point of contention in debates about immigration in the United

States is whether unfavorable attitudes toward illegal immigration really reflect, at least for some

Americans, a more general hostility toward foreigners and immigrants. Most Americans have more

favorable views toward legal immigration than illegal immigration, a pattern that may well indicate

concerns about legal status, the rule of law, and control of the U.S. border. As National Review columnist

Jonah Goldberg (2007) writes:

19

The reason immigration restrictions are aimed at Mexicans and Latinos is quite simple: That’s

where the problem is. Whether you agree with them or not, most “anti-immigration” conservatives

actually think that there is an important distinction between legal and illegal immigration. Want a

hint as to why? One is legal and the other isn’t.

Under this view, attitudes toward legal and illegal immigration should be relatively detached from each

other, with support for one unrelated or only weakly related to support for the other. However, it is also

possible that there are more general concerns about the overall influx of immigrants without a consistent

distinction made between legal and illegal status. Do Latinos’ attitudes about legal and illegal immigration

adhere more to the former or the latter view?

To examine whether legal and illegal immigration are linked, first we conduct simple bivariate

analyses of the relationship between (1) respondents’ opinions regarding legal immigration from Latin

America and whether they perceive illegal immigration as helping or hurting the economy, and (2)

respondents’ opinions about legal immigration and their support for a permanent worker status program;

and (3) respondents’ perceptions of the economic effects of illegal immigration and support for a

permanent worker status program. We find that all relationships are positive and highly significant.4

It would appear that these three sets of attitudes are highly related to one another. But do these

separate items coalesce to form a single dimension of attitudes toward immigration? In order to consider

the possibility of a single dimension of immigration support, we conduct a principal components factor

analysis on these immigration items. A single dominant factor emerges with an eigenvalue of 2.11 and

70% of the variance explained.5 All three variables have similar factor loadings and scoring coefficients,

and based on these results we can create a single standardized factor score, which we denote the

general immigration scale. The distribution of observations on this scale is summarized in Figure 1.

The scale is skewed heavily to the left, with a vast majority of Latino respondents taking a pro-immigration

position and smaller proportions staking out less favorable positions toward immigration. These results

suggest that Latinos see legal and illegal immigration in similar terms. Individuals who are most strongly

supportive of increased legal immigration have more positive perceptions of the effect of illegal

immigrants on the economy and favor a more lenient policy of permanent legal status for illegal

immigrants currently in the country.

20

One caveat is in order here. Due to the nature of our data, it is not possible to say that Latinos who

are supportive of increases in legal immigration are necessarily supportive of increases in illegal

immigration. It is quite possible for individual Latinos to perceive that illegal immigration has a positive

effect on the economy but oppose increases in illegal immigration on other grounds, such as national

security, border control, support for the rule of law, or personal self interest. What we can say with

confidence is that attitudes toward legal immigrants, the economic benefits of illegal immigrants, and

support for permanent legal status for immigrants are correlated highly enough to form a coherent

“general immigration” dimension.

In the fourth set of columns in Table 1, we reports OLS estimates for a model with the general

immigration scale serving as the dependent variable. As one would expect, the significance of national

origin observed in the individual immigration models remains strong in the immigration index model.

Mexicans drive general support for immigration, while all other Latino groups, save for Salvadorians and

Central Americans, are significantly less likely to hold positive views about immigration. Turning to the

other groups of independent variables, we find general consistency with the results observed in the other

models. Once again we note that ethnic attachment and acculturation variables provide the strongest

explanatory value for general immigration attitudes; all but two of the coefficients in this group are

statistically significant and in the expected direction. Language is a strong predictor of support for

immigration. Latinos whose primary language is Spanish or who are bilingual are more likely to be

favorably oriented toward immigration, as are respondents who think it is important for future generations

of Latinos to speak Spanish. Further, respondents who perceive discrimination against Latinos to be a

problem are more pro-immigration in their attitudes. On the other hand, Latinos who are strongly attached

to and incorporated into American culture are significantly less supportive of immigration. Our findings

also provide strong evidence that support for immigration is divided along generational lines, with first

generation Latinos being strongly supportive of overall immigration and second- and third-generation

Latinos showing significantly more negative attitudes about relaxed immigration.

The overall effects of both economic self interest and labor competition and political attitudes and

demographic attribute variables are not particularly strong. In the former category, only respondents who

view crime as an important issue are less likely to have pro-immigration attitudes. In the latter category,

21

support for government expansion leads to a more pro-immigration stance, but Latina women appear to

stake out a significantly more negative position on immigration than Latino men. Finally, we note the one

significant finding among the county contextual variables. Contrary to expectations, increases in Latino

immigration in respondents’ home counties depresses support for immigration.

DISCUSSION AND CONCLUSIONS

In this paper we use data from a 2004 National Survey of Latinos to examine the determinants of

Latinos’ opinions about immigration. The availability of this dataset, which includes a large sample of

Latino respondents, allows for stronger inferences and greater generalizability—something that has often

been a limiting factor in many previous studies about the immigration attitudes of Latinos. In exploring the

determinants of Latinos’ attitudes toward immigration, we consider attitudes toward illegal immigration,

legal immigration, and a proposed policy designed to address the issue of undocumented workers. We

also consider whether Latino attitudes toward legal and illegal immigration are highly related and

comprise a single coherent structure. We find that there is a coherent structure to Latinos’ attitudes

toward legal immigration, illegal immigration, and policy choices for dealing with illegal immigrants, and

we use factor scores representing that coherent structure as a fourth dependent variable in our analysis.

When we compare the effects of independent variables representing several competing theoretical

frameworks, our empirical findings reveal clear explanatory winners and losers. Taken as a whole, our

test of competing theoretical approaches designed to explain Latinos’ attitudes toward immigration points

to the importance of two sets of variables: (1) national origin; and (2) ethnic attachment and acculturation,

particularly the importance placed on the Spanish language, pro-acculturation attitudes, standards for

immigrant incorporation, and generational differences. The three other sets of independent variables—

economic self interest and labor competition, political attitudes and demographic attributes, and county

contextual variables—do not fare as well in comparison with other explanations and have marginal or

inconsistent effects on Latinos’ immigration attitudes. In particular, we are struck by the null or

inconsistent findings for variables representing individuals’ economic vulnerability. Contrary to

conventional wisdom, opposition to expanded immigration among Latinos is not driven by unemployment

status or low income, but rather, unemployed or low-income Latinos are no different than other Latinos in

their attitudes about immigration. Instead, there is a distinct difference between the immigration attitudes

22

of Mexicans and other Latinos, as well as between first-, second-, and third-generation Latinos. Beyond

country of origin and generational effects, Latinos’ immigration attitudes are dominated by their relative

opinions about other Latinos and the importance of the Spanish language, as well as their incorporation

into American society and their support for the tenets of pro-Americanism.

This study by no means provides an exhaustive understanding of Latinos’ attitudes toward

immigration. It is clear that there may be other determinants of Latino immigration preferences that are

not captured here. We are constrained by the type and number of questions about immigration and

possible independent variables asked across various surveys of Latinos, including (but not limited to) the

NSL. There are other aspects of immigration attitudes that warrant study, and we are unable to explore

the effects of other possible independent variables. This is particularly important given our strong findings

with respect to the influence of country of origin on Latinos’ immigration attitudes. Do differences between

Mexicans and other Latino groups persist when more expansive data are examined? It is important to

note that Mexicans currently make up 65.5% of the U.S. Latino population (U.S. Census Bureau 2006),

but the size of the Latino immigrant population from other countries continues to increase. Despite the

disparity in Latino population makeup between Mexicans and other groups, immigration is often framed

as one of the few issues that unites and mobilizes the heterogeneous Latino community. If differences in

attitudes about immigration run just as deep among Latinos as they do within the general American

population, then there could be significant implications for how the immigration debate is framed in the

future. For this and other reasons, the research agenda on immigration attitudes can benefit greatly from

a comparative approach that explores the applicability of immigration attitude models across different

racial and ethnic groups. In particular, it is important to ask whether the theoretical perspectives that

explain the immigration attitudes of Latinos are equally applicable to the immigration attitudes of whites,

blacks, and Asians.

23

ENDNOTES

1. We note here the difficulty in disentangling the unique meanings of terms such as assimilation,

acculturation, and accommodation found in the literature. We choose to rely on the broader and less

constrained terms that express some sense of the degree to which Latinos fit into American society

without actually trying to delineate all the conditions necessary for full assimilation.

2. Because we include county-level variables in our individual-level model, errors are unlikely to be

independent for individuals residing within the same counties. For three of our primary dependent

variables, the intra-class correlations (ICC) are large enough that we can reject the null hypothesis of

perfect independence of error terms within counties (Cohen, Cohen, West, and Aikin, 2003). Hence in

each of our models we adjust our standard errors to account for the possible effects of clustering by

county (Primo, Jacobsmeier, and Milyo, 2007).

3. We consider the possibility that the effects of income and unemployment on attitudes toward illegal

immigrants are a function of how long Latinos have been in the United States. We do find an

interaction effect for household income and third-generation status (b = 0.152, t = 3.11); this suggests

that third generation, high income Latinos are more favorable toward illegal immigrants, while their

low income counterparts are less favorable in their views toward illegal immigrants. For the sake of

brevity, we do not present these results here, though interested readers may obtain these results by

writing the authors.

4. There is a discernible relationship between attitudes toward legal and illegal immigration (Kendall’s

tau-b = 0.28, Pearson χ2 = 232.45, prob. < 0.0001), between attitudes toward legal immigration and

support for a permanent worker status program (Kendall’s tau-b = 0.30, Pearson χ2 = 277.92, prob. <

0.0001), and between attitudes toward illegal immigration and support for a permanent worker status

program (Kendall’s tau-b = 0.32, Pearson χ2 = 204.78, prob. < 0.0001).

5. The factor loadings and scoring coefficients, respectively, for these variables are as follows: (1) more

illegal immigrants from Latin America: 0.807, 0.382; (2) Illegals help the U.S. economy: 0.817, 0.387;

and (3) illegal immigrants can stay with permanent status: 0.890, 0.422. A full table with principal

component factor results (including the complete wording for the three questions used to create this

scale) is available from the authors upon request.

24

Table 1. Estimates for various models of immigration attitudes among Latinos _________________________________________________________________________________________________________________________________

(1) (2) (3) (4) ---------------------------- ---------------------------- ----------------------------- ----------------------------- Illegal Immigrants Support for more Illegal Immigrant General Help Economy Legal Immigration Work Status Program Immigration Scale ---------------------------- ---------------------------- ---------------------------- ----------------------------- b t b t b t b t _________________________________________________________________________________________________________________________________

Economic Self Interest and Labor Competition Important issue: economy [-] -0.216 -1.93** -0.037 -0.45 0.266 1.90* -0.007 -0.17 Important issue: crime [-] -0.162 -2.09** -0.104 -1.53* -0.166 -1.28 -0.075 -2.48*** Important issue: immigration [+/-] 0.067 0.83 0.150 2.03** -0.090 -0.64 0.043 1.11 Unemployed [-] 0.215 0.74 -0.011 -0.05 0.021 0.05 0.074 0.72 Household income [+] 0.006 0.58 0.001 0.17 0.009 0.72 0.003 0.92

Ethnic Attachment and Acculturation Importance of Spanish [+] 0.257 3.49*** 0.322 3.89*** 0.521 4.34*** 0.192 4.87*** Language: Spanish dominant [+] 0.672 3.24*** 0.042 0.20 1.294 3.80*** 0.231 2.20** Language: bilingual [+] 0.296 1.24 0.059 0.34 0.533 2.43*** 0.135 1.34* Discrimination against Hispanics a problem [+] 0.261 3.25*** 0.095 1.59* 0.247 1.83** 0.086 2.94*** Experienced discrimination in past five years [+] 0.067 0.41 -0.042 -0.35 -0.020 -0.11 -0.007 -0.10 Pro-Acculturation (-) 0.017 0.32 -0.115 -2.37*** -0.211 -1.95** -0.044 -2.08** American society scale [-] -0.274 -4.22*** -0.242 -4.49*** -0.274 -2.59*** -0.139 -6.11*** Second-generation Latino [-] -0.631 -2.71*** -0.326 -1.78** -0.294 -0.83 -0.222 -2.86*** Third-generation Latino [-] -0.780 -2.77*** -0.898 -4.84*** -0.940 -3.08*** -0.481 -4.77*** Naturalized U.S. citizen [-] 0.031 0.12 -0.217 -1.55* 0.161 0.51 -0.020 -0.25

Political Attitudes and Demographic Attributes Republican [-] -0.063 -0.41 -0.149 -1.31 0.077 0.39 -0.056 -1.03 Independent [-] -0.413 -1.62 0.104 0.54 -0.0001 -0.00 -0.069 -0.79 Support for expanded government [+] 0.523 3.29*** 0.143 1.07 0.439 2.43*** 0.152 2.34** Education [+] 0.066 1.55* 0.008 0.27 -0.022 -0.41 0.011 0.73 Age -0.0001 -0.02 0.0003 -0.09 0.003 0.42 -0.0004 -0.24 Gender -0.394 -3.32*** -0.173 -1.57 0.052 0.31 -0.102 -2.04** Urban resident [+] -0.036 -0.58 -0.016 -0.36 -0.047 -0.54 -0.019 -1.00

County contextual variables (2004) Latino population [+] -1.047 -2.46** -0.509 -1.62 0.070 0.13 -0.332 -2.17** Change in Latino population, 2000-2004 [+/-] -0.737 -0.16 -6.383 -1.41 9.854 1.71* -2.228 -1.31 Unemployment rate (-) 0.001 0.03 -0.013 -0.38 -0.049 -0.90 -0.002 0.16

25

Table 1 (continued) _________________________________________________________________________________________________________________________________ (1) (2) (3) (4) ---------------------------- ---------------------------- ----------------------------- ----------------------------- Illegal Immigrants Support for more Illegal Immigrant General Help Economy Legal Immigration Work Status Program Immigration Scale ---------------------------- ---------------------------- ---------------------------- ----------------------------- b t b t b t b t _________________________________________________________________________________________________________________________________ National Origin (country or region) Puerto Rico -1.832 -9.42*** -0.417 -2.54** -0.860 -2.73*** -0.558 -6.60*** Cuba -1.021 -6.03*** 0.100 0.63 -0.719 -2.54** -0.147 -2.35** Dominican Republic -1.661 -4.11*** -0.224 -0.72 -1.177 -2.02** -0.446 -2.51** El Salvador -0.490 -0.74 0.032 0.09 0.298 0.30 -0.076 -0.54 Central America -0.160 -0.57 0.193 0.88 -0.421 -0.74 -0.064 -0.60 South America -1.430 -5.70*** -0.538 -2.97*** -0.729 -1.67* -0.438 -4.87*** Spain -1.309 -4.11*** -0.351 -0.98 -0.686 -1.78* -0.473 -2.71*** Constant 1 1.407 2.20** -2.110 -4.33*** 0.955 1.18 -0.105 -0.44 Constant 2 --- --- 0.586 1.16 --- --- --- --- _________________________________________________________________________________________________________________________________ Estimation Procedure Binary Logit Ordered Logit Binary Logit OLS Regression N 1438 1455 1506 1342 R2 / Pseudo R2 0.162 0.055 0.172 0.206 F / Wald χ2 317.52 280.16 324.91 18.86 Prob (F) / Prob Wald χ2 0.000 0.000 0.000 0.000 Note: Symbols in brackets represent the expected direction of the coefficient. T-statistics are based on standard errors estimated with clustering by county. For the country / region of origin variables, Mexico is the excluded (comparison) category. *** prob < .01 ** prob < .05 * prob < .10

26

Figure 1. Histogram for distribution of respondents on pro-immigration scale.

0.2

.4.6

.8D

ensi

ty

-3 -2 -1 0 1Immigration Scale

Note: The pro-immigration scale is based on a factor analysis of three items: (1) support for increases in legal immigration from Latin America; (2) support for the view that illegal immigrants help the economy by providing low-cost labor; and (3) support for plan to permit illegal Hispanic immigrants to remain in U.S. permanently with legal status.

27

REFERENCES

Abrajano, Marisa and Simran Singh. 2009. “Examining the Link Between Issue Attitudes and News

Source: The Case of Latinos and Immigration Reform.” Political Behavior 31 (1): 1-30.

Altonji, Joseph C. and David Card. 1991. “The Effects of Immigation on the Labor Market Outcomes of

Less-skilled Natives,” in John M. Abowd and Richard B. Freeman (eds.), Immigration, Trade, and the

Labor Market. Chicago: University of Chicago Press.

Amir Y. 1976. “The Role of Intergroup Contact in Change of Prejudice and Race Relations. In P.A. Katz

(ed.), Towards the Elimination of Racism. New York: Pergamon.

Binder, Norman E., J.L. Polinard, and Robert D. Wrinkle. 1997. “Mexican American and Anglo Attitudes

toward Immigration Reform: A View from the Border.” Social Science Quarterly 78: 324-337.

Borjas, George J. 2003. “The Labor Demand Curve Is Downward Sloping: Reexamining the Impact of

Immigration on the Labor Market.” Quarterly Journal of Economics November: 1335–74.

Borjas, George J. 2004. “Increasing the Supply of Labor Through Immigration: Measuring the Impact on

Native-born Workers.” Center for Immigration Studies, Washington, D.C.

Branton, Regina. 2007. “Latino Attitudes Toward Various Areas of Public Policy: The Importance of

Acculturation.” Political Research Quarterly 60: 293-303.

Burns, Peter and James G. Gimpel. 2000. “Economic Security, Prejudicial Stereotypes, and

Public Opinion on Immigration Policy.” Political Science Quarterly 115: 201-225.

Butcher, Kristin F. and David Card. 1991. “Immigration and Wages: Evidence from the 1980s,” American

Economic Review 81: 292–296.

Citrin, Jack, Beth Reingold, and Donald P. Green. 1990. “American Identity and the Politics of Ethnic

Change.” Journal of Politics 52: 1124-1152.

Conover, Pamela Johnston and Stanley Feldman. 1984. “How People Organize the Political World: A

Schematic Model.” American Journal of Political Science 28: 95-126.

Cohen, Jacob, Patricia Cohen, Stephen G. West, and Leona S. Aikin. 2003. Applied Multiple Regression /

Correlation Analysis for the Behavioral Sciences. Mahwah, N.J.: Lawrence Earlbaum Associates.

de la Garza, Rodolfo O. 1992. Latino Voices: Mexican, Puerto Rican, and Cuban Perspectives on

American Politics. Boulder, Colorado: Westview Press.

28

de la Garza, Rodolfo O., F. Chris Garcia, John Garcia, and Angelo Falcon. 1993. “Attitudes toward

Immigration Policy.” Migration World Magazine 21: 13-17.

de la Garza, Rodolfo O., Jerry L. Polinard, Robert D. Wrinkle, and Tomás Longoria Jr. 1991.

“Understanding Intra-Ethnic Attitude Variations: Mexican Origin Population Views of Immigration.”

Social Science Quarterly 72: 379-387.

Giles, Micheal and Melanie Buchner. 1993. “David Duke and Black Threat: An Old Hypothesis Revisited.”

Journal of Politics 55: 702-11.

Glazer, James. 1994. “Back to the Black Belt: Racial Environment and White Racial Attitudes in the

South.” Journal of Politics 56: 21-41.

Goldberg, Jonah. 2007. “The Wrong Side of the Immigration Story.” National Review Online (June 6).

http://article.nationalreview.com/?q=ZDgzNWEyZWRjMDJkYTQ4ZWEzYmFiZDM3YjdiYzQzNzQ

Gordon, Milton. 1961. “Assimilation in America: Theory and Reality.” Daedalus 90 (2).

Graves, Scott and Jongho Lee. 2000. “Ethnic Underpinnings of Voting Preference: Latinos and the U.S.

Senate Election in Texas.” Social Science Quarterly 81: 227-236.

Hood, M.V. III, Irwin L. Morris, and Kurt A. Shirkey. 1997. “‘Quedate o Vente!’: Uncovering the

Determinants of Hispanic Public Opinion toward Immigration.” Political Research Quarterly 50: 627-

647.

Huntington, Samuel. 2004. Who Are We? The Challenge’s to America’s National Identity. New York:

Simon & Schuster.

Kelly, Nathan J. and Jana Morgan Kelly. 2005. “Religion and Latino Partisanship in the United States.”

Political Research Quarterly 58 (1): 87-95.

Leal, David. 2002. “Political Participation by Latino Non-Citizens in the United States.” British Journal of

Political Science 32: 353-370.

Newton, Lina Y. 2000. “Why Some Latinos Supported Propostion 187: Testing Economic Threat and

Cultural Identity Hypotheses.” Social Science Quarterly 81: 180-193.

Orrenius, Pia M. and Madeline Zavodny. 2003. “Does Immigration Affect Wages? A Look at

Occupational-Level Evidence.” Federal Reserve Bank of Dallas Research Department, Working

Paper 0302.

29

Pettigrew, Thomas. 1986. “The Contact Hypothesis Revisited.” In M. Hewstone and R. Brown (eds.),

Contact and Conflict in Intergroup Encounters. Oxford: Blackwell.

Pettigrew, Thomas. 1998. “Intergroup Contact Theory.” Annual Review of Psychology 49: 65-85.

Pew Hispanic Center. 2006a. “From 200 Million to 300 Million: The Numbers Behind Population Growth.”

Pew Research Center: Washington, D.C.

Pew Hispanic Center. 2006b. “Hispanics at Mid-Decade.” Pew Research Center: Washington, D.C.

Pew Hispanic Center. 2006c. “The Size and Characteristics of the Unauthorized Migrant Population in the

U.S.: Estimates Based on the March 2005 Current Population Survey.” Pew Research Center:

Washington, D.C. http://pewhispanic.org/files/reports/61.pdf.

Polling Report. 2007. “Immigration.” PollingReport.com. http://www.pollingreport.com/immigration.htm.

Powers, D. A. and C. G. Ellison. 1995. “Interracial Contact and Black Racial Attitudes: The Contact

Hypothesis and Selectivity Bias. Social Forces 74: 205-26.

Primo, David M., Matthew L. Jacobsmeier, and Jeffrey Milyo. 2007. “Estimating the Impact of State

Policies with Mixed-Level Data.” State Politics and Policy Quarterly 7: 446-59).

Sum, Andrew, Paul Harrington, and Ishwar Khatiwada. 2006. “The Impact of New Immigrants on Young

Native-Born Workers, 2000-2005.” Center for Immigration Studies.

http://www.cis.org/articles/2006/back806.pdf

Suro, Roberto. 2005. “Attitudes toward Immigrants and Immigration Policy: Surveys Among Latinos in the

U.S. and Mexico.” Pew Hispanic Center. http://pewhispanic.org/reports/report.php?ReportID=52

Swers, Michelle. 2002. The Difference Women Make: The Policy Impact of Women in Congress. Chicago:

University of Chicago Press.

Thomas, Sue. 1994. How Women Legislate. New York: Oxford University Press.

Welch, Susan and John R. Hibbing. 1992. “Financial Conditions, Gender, and Voting in American

National Elections.” Journal of Politics 54 (1): 197-213.

U.S. Census Bureau. 2006. “U.S. Hispanic Population 2006.” Current Population Survey, Annual Social

and Economic Supplement.

Voss, D. Stephen. 1996. “Beyond Racial Threat: Failure of an Old Hypothesis in the New South.” Journal

of Politics 58: 1156 1170.