Few social or political issues are as salient in the contemporary United States as immigration (Chavez 2008; Kubrin et al. 2012; Román 2013). Research examining public opinion on immigration thus has considerable societal relevance and import. Scholars in diverse disciplines have developed an extensive body of knowledge about popular attitudes toward immigrants and immigration policy (Brader et al. 2008; Ellison et al. 2011; Filindra and Pearson-Merkowitz 2013; Hainmeuller and Hiscox 2010; Hopkins 2010; Wang 2012). Informed by minority group threat theory, studies have documented that public perceptions of both immigrants and Latinos as socially threatening are associated with the desire for reduced immigration levels (Citrin et al. 1997; Espenshade and Hempstead 1996; Fetzer 2000; Hawley 2011; Hood and Morris 1997; Lu and Nicholson-Crotty 2010; Watson and Riffe 2013). Threat perceptions have also been linked to increased support for specific anti-immigration policies, such as building a border fence and denying illegal immigrants work permits and emergency healthcare (Berg 2013; Buckler et al. 2009; Chiricos et al. 2014; Stupi et al. 2014; Wilson 2001).

In recent years, however, immigration enforcement in the US has changed dramatically (Chacón 2009, 2012), undergoing a “historical about-face” (Coleman and Kocher 2011: 229). Specifically, the deportation of noncitizens—documented and undocumented—with criminal arrest records—misdemeanor or felony, remote or recent—has become “the driving force in American immigration enforcement” (Eagly 2013: 1128, see also Inda 2013). In turn, “the criminal justice system has become the primary means to locate, remove, and permanently banish immigrants from the United States” (Vazquez 2011: 642). Scholars have characterized the growing focus on the policing of “criminal aliens” as “crimmigration” (Stumpf 2006), a “criminal removal system” (Eagly 2013: 1128), and the “governing [of] immigration through crime” (Inda and Dowling 2013: 2). Historically, federal authorities have maintained a monopoly over the policing of immigration (Coleman and Kocher 2011). Under crimmigration, by contrast, state and local police have emerged as gatekeepers of immigration enforcement, and “the key decision-making moment has been the initial identification of a potentially removable noncitizen by some form of arrest” (Motomura 2011: 1858). The result has been that “entire categories of police encounters” have been infused “with immigration-related meaning and potential consequences” (Kalhan 2013: 1152). Put simply, “mundane, everyday policing … can now lead to detention and eventually deportation” (Coleman and Kocher 2011: 230).

While considerable research exists on the antecedents of popular views about immigration reform, less is known about how perceptions of either immigrants or Latinos influence public support for police policies and practices that, in the context of crimmigration, may significantly shape immigration enforcement. This is notable because many scholars contend that crimmigration has increased racial profiling of Latinos and other abuses of police power, as well as the deportation of legally present noncitizens for minor offenses, such as traffic violations (Chacón 2012; Golash-Boza 2012; Golash-Boza and Hondagneu-Sotelo 2013; Inda 2013). Vasquesz (2011: 674) argues that “crimmigration has become the current mechanism to extend the longstanding subordination and marginalization of Latinos in the United States.” Recent studies also suggest that in the same way crime and welfare are race-coded issues, immigration is an ethnicity-coded issue—support for punitive immigration policies is often a way to covertly express anti-Latino sentiments (Ayers et al. 2009; Hartman et al. 2014; Lu and Nicholson-Crotty 2010; Valentino et al. 2013).

The current study builds on both the scholarship on crimmigration and the literature linking immigration preferences to underlying anti-Latino attitudes. I analyze public views about police powers that while ostensibly race-neutral, influence officers’ ability to identify undocumented immigrants, and to investigate both legally present noncitizens and Latino citizens for evidence of criminal wrongdoing. Using national survey data, I examine whether perceptions of Latinos as politically and economically threatening are associated with support for expanding such powers.

Crimmigration: Emergence and Criminal Justice Implications

Eagly (2013: 1142) documents that as a result of crimmigration there has been a two-fold increase in criminal alien removals in recent years, which was juxtaposed against a decline in the use of other forms of removal. This section briefly reviews some of the key statutory and procedural changes that have contributed to this shift in immigration enforcement. More detailed discussions of the historical development of crimmigration are available elsewhere (see Chacón 2009, 2012; Provine and Doty 2011; Stumpf 2006). My objective here is merely to outline how local police patrols and investigations, even those not targeted specifically at immigration enforcement, have emerged as important means through which anti-immigrant (or anti-Latino) sentiments among members of the public may be transmitted into discriminatory social controls.

The roots of crimmigration trace back to the 1980s, when Congress began taking steps to make it easier to deport noncitizens for criminal offenses and moved to criminalize behaviors associated with immigration, such as hiring undocumented citizens (Chacón 2009; Stumpf 2006). State and local officials subsequently continued the criminalization process, targeting such behaviors as loitering, failing to register employment of day laborers, and using a false identity (Chacón 2009; Varsanyi 2010). An example of nonfederal anti-immigration legislation is Arizona’s SB 1070. The law, which was enacted in 2010, made it a crime for undocumented immigrants to engage in various employment-related activities, and also required police officers to investigate, during lawful stops or arrests, the immigration status of persons who they suspect may be undocumented immigrants (Chacón 2012; Golash-Boza 2012).Footnote 1

In terms of federal initiatives, Congress passed both the Illegal Immigration Reform and Immigrant Responsibility Act (IIRIRA) and the Anti-terrorism and Effect Death Penalty Act (AEDPA) in 1996, which allowed for the deportation of legally present noncitizens for misdemeanor and felony offenses that carry one-year sentences (Morawetz 2000; Inda 2013). These laws also reduced judicial checks on and opportunities for relief from deportation (Vazquez 2011). The result was an expansion of the influence of local criminal justice policies and police practices on immigration enforcement (Golash-Boza 2012; Morawetz 2000). Discussing the “practical meaning” of the 1996 laws, Morawetz (2000) explained that:

The likelihood of deportation is greater in communities that are subject to elevated levels of police activity and in which people are more likely to be arrested and prosecuted. For example, a person of color who is more likely to be stopped while driving is also more likely to be saddled with a drug possession conviction if someone in the car is carrying drugs. That conviction, even if it involves only a small amount of drugs, could mean mandatory deportation (pp. 1945–1946).

During 1996, Congress also added Sect. 287(g) to the Immigration and Nationality Act (INA). Section 287(g) authorized “Memorandum of Understanding” (MOU) agreements between non-federal law enforcement and federal immigration authorities, such that after a brief period of training, state and local police could essentially operate in the capacity of immigration agents, investigating and arresting suspects for immigration offenses (Kalhan 2013; Vazquez 2011).Footnote 2 Although there were no MOU agreements signed until 2002 (Golash-Boza and Hondagneu-Sotelo 2013), more than seventy police agencies eventually entered into these agreements, which in turn resulted in nearly 200,000 additional noncitizens being identified for deportation (Coleman and Kocher 2011; Inda 2013). Many police agencies that have obtained 287(g) agreements have identified removable noncitizens primarily by policing minor offenses, such as driving infractions and violations of open container laws, and have concentrated their immigration enforcement efforts in Latino neighborhoods (Coleman and Kocher 2011; GAO 2009). Lacayo (2010:1-2) has argued that by providing “perilously unchecked authority to local law enforcement,” the 287(g) program “exacerbated racial and ethnic targeting of Hispanics at the local level … and created a threatening and insecure environment for all Latinos.”

More recently, the Immigration and Naturalization Service (INS) developed the Secure Communities program, which automated and routinized the checking of criminal arrestees’ fingerprints against the Department of Homeland Security’s (DHS) immigration database (Kalhan 2013). Implementation of the program began in select jurisdictions during 2008 and was completed in January 2013, at which point it was fully operational in all 3,181 jurisdictions in the US (ICE 2013). In the 4 years following the initial implementation of Secure Communities, there was a five-fold increase in the percentage of all deportations occurring through the program, which amounted to more than 180,000 removals (GAO 2012). The available data reveal that noncitizens deported through Secure Communities have most frequently been identified because of arrests for traffic offenses (GAO 2012). The significance of Secure Communities has been that it has given local police officers the ability “to arrest individuals for the very purpose of booking them and having their immigration status screened—without regard to whether that arrest leads to any criminal prosecution” (Kalhan 2013: 1153).

Latino Threat, Immigration, and Criminal Justice Attitudes

The more Hispanic America becomes, the more Democratic America becomes… The naturalization and registration of 500,000 to 1 million immigrants each year is thus locking up the future for the Democratic Party… If the GOP does not do something about immigration, immigration will do something about the GOP: Turn it into a retirement home of America’s newest minority.

—Pat Buchanan (2006: 59–61).

They come here, they have their babies, and after that they become citizens and all those children use those social services.

—“Save Our State” organizer (cited in Kadetsky 1994: 418).

My son, Alexander, is a white male with blue eyes and blond hair… public policy now discriminates against him. The sheer size of the so-called ‘protected classes,’ that are now politically favored, such as Hispanics, will be a matter of vital importance as long as he lives. And their size is basically determined by immigration.

—Peter Brimelow (1995: 11).

The previous section detailed how routine policing by local law enforcement has become a central component of contemporary immigration enforcement (Kalhan 2013; Vazquez 2011). This section draws on minority group threat theory and prior empirical work to identify the potential social foundations for this policy approach. Minority group threat theory suggests that perceived intergroup threat may increase support for discriminatory social controls directed at outgroup members (Blalock 1967, Liska 1992). Specifically, threat theorists argue that when faced with the possibility of reduced racial privilege or power due to increased intergroup competition, dominant group members often attempt to defend the racial hierarchy by engaging in ideological, behavioral, and political forms of discrimination against outgroup members.

A key premise of minority group threat thoery is that the criminal justice system constitutes an important line of defense against threats to the racial status quo (King and Wheelock 2007; Ousey and Unnever 2012). Studies find, for instance, that aggregate indicators of minority threat are positively associated with police force size, police brutality, and arrest rates (Eitle and Monahan 2009; Jacobs and O’Brien 1998; Kane et al. 2013; Smith and Holmes 2014; Stults and Baumer 2007). The key assumption underlying these studies is that contextual factors such as racial composition and changes in racial composition are positively associated with perceived minority threat and outgroup animus (Pickett et al. 2012). Perceived threat, however, is the key “action” variable in minority group threat theory (King and Wheelock 2007; Pickett et al. 2012), and is believed to be the most proximate cause of the mobilization of formal social controls. Theoretically, empowering or mobilizing the police should help to neutralize outgroup threat by increasing the probability that even low-level deviance by racial others will be detected and punished (Johnson and Kuhns 2009; Weitzer and Tuch 2006). Perceived threat, by increasing outgroup animus, may also reduce social concern for the discrimination and hardship faced by outgroup members as a result of aggressive policing (Unnever and Cullen 2009).

Supporting minority group threat theory, previous studies show that support for punitive immigration policies is higher among persons who perceive immigrants to be culturally, criminally, and economically threatening (Berg 2013; Chiricos et al. 2014; Espenshade and Hempstead 1996; Fetzer 2000; Stupi et al. 2014; Wilson 2001). For example, Chiricos et al. (2014) find that persons who agree that “illegal immigrants” pose a cultural (e.g., “damage the social fabric of America”) or economic (e.g., “increase the demand for social services, which raises taxes”) threat to the US are more likely to favor punitive border (e.g., “erecting a wall along the border”) and internal (e.g., “send anyone the police can identify as an illegal immigrant back to their native country”) controls to stem the flow of immigration.

Importantly, however, recent research demonstrates that a principal antecedent of perceived immigrant threat is animus toward Latinos (Hartman et al. 2014; Lu and Nicholson-Crotty 2010). Other investigations reveal a strong relationship between anti-Latino prejudice and views about immigration policies (Ayers et al. 2009; Burns and Gimpel 2000; Valentino et al. 2013). The implication is that the underlying concerns that energize the immigration debate largely have to do with Latinos generally, regardless of their citizenship status, rather than with undocumented immigrants per se (Hartman et al. 2014). This fact is evident in the quotes that lead off this section.

The conclusion that attitudes toward immigration largely reflect broader fears about Latinos is further supported by Chavez’s (2008: 42) analysis of immigrant threat discourse. He finds that such discourse “does not imagine Latinos, whether immigrants or US-born, as part of the national community,” but rather depicts them “as an internal threat to the larger community.” Discussing one example of the social construction of immigrant threat, Chavez (2008: 30) notes:

[I]t was not just recent Mexican immigrants who posed a threat but even those Americans who were descended from the first Spanish-speaking explorers of the Southwest. Not even 400 years of living in the Southwest, and over 150 years of that period as US citizens, reduced the threat posed by Latinos.

In the view of many non-Latinos, the growing number of Latinos—both immigrants and US citizens—is threatening because it escalates intergroup competition over social and economic resources, such as jobs, results in tax dollars being spent on racial others, and reduces non-Latinos’ influence over political outcomes (Chavez 2008; Román 2013).

If, in fact, immigration functions as a coded issue that allows for the veiled expression of broader anti-Latino sentiments, then the focus of prior research on perceived immigrant threat and public views about immigration policies may be too narrow, especially given crimmigration.Footnote 3 Rather, there is warrant for examining whether, as minority group threat theory would predict, broader criminal justice attitudes, particularly attitudes toward aggressive policing, are shaped by perceptions of Latinos, and not just of immigrants, as threatening. Intensifying state and local law enforcement efforts may provide a particularly attractive route to mobilizing against perceived Latino threat. Instrumentally, aggressive policing, by uncovering Latino criminality, can lead to deportation for noncitizens (Eagly 2013), but, as importantly, it can also result in the incarceration and disenfranchisement of Latino citizens (Stumpf 2006). More generally, scholars argue that local police involvement in immigration enforcement has resulted in police discrimination against Latinos, and has exacerbated socio-economic disadvantage in Latino families and communities (Golash-Boza 2012; Golash-Boza and Hondagneu-Sotelo 2013; Inda 2013; Vazquez 2011). Indeed, survey estimates suggest that in the wake of crimmigration, each year approximately 5–9 % of Latinos in the US are stopped and questioned by legal authorities about their immigration status; nearly one-third of Latinos personally know someone who was recently detained or deported as a result of immigration enforcement (Lopez et al. 2010:11–12). For these reasons, Vazquez (2011: 674) contends that “crimmigration has proven to be an effective vehicle for modern Latino oppression.”Footnote 4

Many non-Latino Americans are aware that crimmigration policies that facilitate aggressive immigration policing by local law enforcement authorities result in discrimination against Latinos, and yet they support them anyway (Nill 2011). As discussed above, minority group threat theory would suggest, as would scholarship on the coded nature of the immigration debate (Ayers et al. 2009; Lu and Nicholson-Crotty 2010), that an important source of support for such policies may be perceptions of the threatening characteristics of Latinos. However, very little evidence exists about whether perceived Latino threat influences support for either local policing of immigration or for police practices, such as racial profiling, that may contribute to discrimination against Latinos in the context of crimmigration. Indeed, to my knowledge, no prior studies have analyzed whether perceptions about Latinos affect attitudes toward either aggressive police practices or toward police investigative powers. More broadly, there has been a neglect of Latino issues in policing scholarship (Martinez 2007; Weitzer 2013).

Although no previous research has assessed the influence of views about Latinos on policing attitudes, a handful of studies have explored whether stereotypes about Latinos and perceptions of Latino threat impact other types of criminal justice attitudes, such as views about drug rehabilitation and capital punishment (Barkan and Cohn 2005; Johnson et al. 2011; Nielsen et al. 2010; Stewart et al. 2015; Unnever and Cullen 2012; Welch et al. 2010). These investigations, though, save for two exceptions (Johnson et al. 2011; Stewart et al. 2015), all drew on relatively dated data—the respective surveys were conducted between thirteen and twenty-four years ago. As important, the results from this empirical literature have been mixed, with many studies reporting null effects (Barkan and Cohn 2005; Nielsen et al. 2010; Unnever and Cullen 2012). Nielsen et al. (2010), for example, found no evidence that either anti-Latino prejudice or negative stereotypes about Latinos were associated with views about drug rehabilitation.

The studies by Johnson et al. (2011) and Stewart et al. (2015), which drew on the same data, both reported significant effects of perceived Latino threat on, respectively, (1) support for the use of ethnicity in criminal sentencing, and (2) support for harsh punishments for Latino offenders. However, the outcome variables in these two studies explicitly focused on sentences for Latino offenders. For example, Stewart et al. (2015) found that respondents who agreed with statements such as “Hispanics hurt the US by committing more violent crimes than other racial or ethnic groups” were more likely to express punitive Latino sentiment, as measured with agreement with statements like “In the US Hispanic offenders commit most of the violent crimes and should be punished severely.” There are two potential issues with this approach. First, if both the independent and dependent variables explicitly refer to Latinos, it raises the possibility that the findings may simply reflect the correlation of alternative measures of anti-Latino attitudes, rather than the effect of perceived threat on support for punitive social controls. Second, criminal justice policies are publically debated in the US in ostensibly race- and ethnicity-neutral terms (see King and Wheelock 2007; Pickett and Chiricos 2012; Unnever and Cullen 2010), and by law cannot be applied differentially based on a defendant’s race or ethnicity. Thus, the most relevant question is whether perceived threat influences support for ostensibly race- and ethnicity-neutral crime policies. Finally, in the context of crimmigration, the specific type of criminal justice attitude that is arguably the most theoretically germane and politically relevant for understanding the criminal justice consequences of anti-Latino sentiments is support for aggressive policing.

Similar to the case of views about Latinos, there has actually been very little research that has explored whether attitudes toward Blacks or other minorities are associated with support for police powers. Rather, most prior research on racialized fears and perceptions has focused either on punitive attitudes, such as support for the death penalty or juvenile transfers (see, e.g., King and Wheelock 2007; Pickett and Chiricos 2012; Unnever and Cullen 2010), or on views about rehabilitation (Nielsen et al. 2010; Pickett et al. 2014). Indeed, to my knowledge, no previous studies have examined whether perceptions of either Black economic or political threat are associated with policing attitudes. A few studies have assessed whether stereotypes about Blacks and general racial prejudice are associated with increased support for aggressive or biased policing, but this literature has yielded equivocal findings. For example, Barkan and Cohn (1998) found that stereotyping Blacks as violent was positively associated with Whites’ support for police use of excessive force, but was not significantly associated with support for police use of reasonable force. Interestingly, those authors also found that general racial prejudice was not associated with Whites’ support for either type of force (see p. 749; see also Arthur and Case 1994). Jonson and Kuhn (2009) found that stereotyping Blacks as violent was associated with Whites’ support for police use of force specifically against Black suspects, regardless of whether the force was excessive or reasonable, but was not associated or was negatively associated with support for using force against White suspects. By contrast, Johnson et al. (2011) found that stereotyping Blacks as violent was not significantly associated with support for racial profiling. Finally, Peffley et al. (1997) found that stereotyping Blacks as violent was positively associated with Whites’ support for allowing police officers to search Blacks if they were using foul language in public, but not if they were well behaved.

Below, I advance the existing Latino threat literature and the scholarship on crimmigration by analyzing data from a recent national survey to provide the first examination of whether perceived Latino threat is related to attitudes toward aggressive policing, which I operationalize as support for expanding police powers. Specifically, building on minority group threat theory, I test the following hypothesis:

Hypothesis #1

Perceptions of Latino economic and political threat will be positively associated with support for expanding police investigative powers.

To the extent that group threat theory is correct, and support for aggressive policing functions as a defensive reaction to perceived outgroup threat, then some types of police policies should receive greater levels of threat-based support than others. Specifically, perceived threat should be most strongly correlated with those policies that provide the police with the greatest discretion to differentially investigate and patrol outgroup members, and that also insulate police officers from charges of discrimination. For this reason, I test the following hypothesis:

Hypothesis #2

The relationship between perceived Latino threat and support for expanding police powers will be strongest in the case of investigative powers, such as police profiling, that have the clear potential to result in discrimination against Latinos.

Data and Methods

I test my hypotheses using data from a large nationally representative telephone survey conducted in the summer of 2010. To generate the sample for the survey, a list of randomly generated telephone numbers was purchased from Survey Sampling International (SSI), and a single adult (18 and older) respondent was randomly selected from within each sampled household using the “most recent birthday” method. A team of trained interviewers employed by Oppenheim Research, a public opinion polling firm in Florida, fielded the survey using computer assisted telephone interviewing (CATI). To allow supervisors to closely monitor interviews for consistency and accuracy, interviews were conducted from a single call center equipped with multiple call stations. In addition, as a check against data entry errors, supervisors called back 12 % of respondents to verify their responses.Footnote 5

In total, 961 respondents participated in the survey. This sample had the following characteristics (as compared to the population of US adults in parentheses): whites, 78 % (75 %); blacks, 12 % (12 %); Latinos, 8 % (14 %); males, 44 % (49 %); age 65 or older, 16 % (17 %); college graduates with at least a four-year degree, 37 % (25 %); and annual household income of $100,000 or more, 19 % (21 %). After listwise deletion, the analytic sample size ranges between 929 and 932. That the sample is more educated than the US public is normal in telephone surveys (Lavrakas 1993). I attempt to address this issue by controlling for education in the models. Additionally, prior studies find that education is not strongly or consistently correlated with views about police powers (Barkan and Cohn 1998; Johnson and Kuhns 2009; Weitzer and Tuch 2002, 2006).

Dependent Variable

As discussed above, routine policing by state and local police is now the driving force of interior immigration enforcement and is also a key source of discrimination against Latino citizens (Vazquez 2011). For example, Kalhan (2013: 1155) explains that the implementation of programs such as Secure Communities has cast “virtually all routine law enforcement activities at least potentially with immigration enforcement significance.” Moreover, under 287(g) agreements, “the general investigatory powers that police possess to stop and question individuals on criminal law enforcement grounds can be used for immigration status inquiries under the guise of public safety” (Coleman and Kocher 2011: 233).

Accordingly, a central question, which I address in the analysis, is whether perceived Latino threat influences views about expanding police investigative powers within the context of routine law enforcement activities. Respondents’ support for expanding police powers is measured using a three-item index (α = .824). Each respondent’s score on the index (Expand Police Powers) is equal to his or her average level of support (1 = not at all supportive, 10 = very supportive) for the following three policies: (1) “Allowing police officers to stop and question individuals based on the way they look”; (2) “Making it easier for police officers to search individuals’ cars and homes”; (3) “Allowing police officers to use more force against suspects.” The descriptive statistics for this variable, and those described below, are included in Table 1.

Table 1 Descriptive statistics and bivariate correlations between independent and control variables and support for expanding police powers
Full size table

Independent Variables

There are two key independent variables in this study, which separately capture the extent to which respondents perceive Latinos as politically and economically threatening. I focus on these two forms of intergroup threat for several reasons. First, these two forms of threat were those originally delineated by Blalock (1967), but have received less attention in the empirical literature than criminal threat or violence stereotypes. These forms of threat are also central themes in anti-Latino discourse (see, e.g., Brimelow 1995; Buchanan 2006).

Latino Political Threat is measured as a respondent’s average level of agreement (1 = strongly disagree, 5 = strongly agree) with the following two statements: (1) “When Latinos vote in local or national elections, they can have too much influence on the outcome;” (2) “As a result of recent events, too many Latinos will vote in upcoming elections.” The responses to these two items are highly correlated (r = .592, α = .743), which supports their integration into a single measure of perceived political threat. Latino Economic Threat is measured as respondents’ level of agreement (1 = strongly disagree, 5 = strongly agree) that: “Latinos take away economic resources that should go to others, like jobs and welfare.”Footnote 6

Control Variables

To reduce the potential for omitted variable bias, I control for factors that extant theory and research suggest may be correlated with views about Latinos and criminal justice attitudes. Previous studies show that demographic characteristics, especially age and race, are associated with attitudes toward police powers, such as racial profiling and the use of force against suspects (Barkan and Cohn 1998; Gabbidon and Higgins 2013; Higgins et al. 2008; Johnson and Kuhns 2009; Johnson et al. 2011; Weitzer and Tuch 2002; 2006). Accordingly, I control for respondents’ race (Black = 1), ethnicity (Latino = 1), gender (Female = 1), Age, Education, and household Income.Footnote 7 Age, Education, and Income are ordinal variables; the response categories for each variable are presented in Table 1.

There is evidence that political ideology, media consumption, and crime salience influence attitudes toward the police (Callanan and Rosenberger 2011; Gabbidon et al. 2009; Higgins et al. 2010; Reitzel and Piquero 2006; Weitzer and Tuch 2002, 2006). I therefore control for respondents’ political ideology (Moderate = 1, Conservative = 1, Liberal = 0), exposure to television news and crime programs, perceived victimization risk, and actual victimization experience in the 5 years preceding the survey (Household Victim = 1). I measure media consumption using three questions that asked respondents to estimate the weekly number of hours they spend (1) “watching national evening news like CNN,” (2) “watching local television news,” and (3) “watching crime programs (such as Law and Order, CSI, or COPS).” Research suggests that survey respondents cannot accurately recall their exact amount of media consumption (Prior 2009). For this reason, I recoded responses to the above questions to generate three binary controls (National News, Local News, Crime TV) that differentiate respondents who reported no weekly exposure (coded “0”) from those who reported a nonzero amount of exposure (coded “1”).Footnote 8 Perceived Risk is a six-item index (α = .902) equal to the average across each respondent’s assessments of the probability (1 = not at all likely, 10 = very likely) of experiencing six different types of victimization—auto theft, burglary, assault, robbery, rape, and murder—in the following year.

A handful of prior investigations have found that racially prejudiced individuals are more likely to favor granting police greater latitude in searching and using force against suspects (Peffley et al. 1997; Johnson and Kuhns 2009, but see Johnson et al. 2011). More broadly, an extensive literature demonstrates that racial concerns exert a strong impact on crime control preferences (Pickett and Baker 2014; Pickett et al. 2014; Unnever and Cullen 2010). In the analysis, I control for racial attitudes using responses (1 = strongly disagree, 5 = strongly agree) to five items from Henry and Sears’s (2002) Symbolic Racism 2000 scale: (1) “There is a lot of discrimination against Blacks in the US today, limiting their chances to get ahead”; (2) “It’s really a matter of some people not trying hard enough; if Blacks would only try harder they could be just as well off as Whites”; (3) Irish, Italians, Jewish, and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors”; (4) “Over the past few years, Blacks have gotten less than they deserve”; (5) “Generations of slavery and discrimination have created conditions that make it difficult for Blacks to work their way out of the lower class.” I recoded responses so that higher values indicated greater racism, and then averaged across the items to construct a measure of Symbolic Racism (α = .732).Footnote 9

Previous research documents a strong positive association between punitive attitudes and support for police powers (Arthur and Case 1994; Cullen et al. 1996). I control for Punitiveness using a seven-item index (α = .817) that measures respondents’ average level of support (1 = not at all supportive, 10 = very supportive) for the following “get tough” youth justice policies: (1) “Trying more juvenile offenders in adult courts”; (2) “Locking up more juvenile offenders”; (3) “Making sentences more severe for juveniles who commit crimes”; (4) “Making juvenile offenders work hard labor while they are in prison”; (5) Taking away television and recreation privileges from juveniles in prison”; (6) “Putting violent juvenile offenders in adult prisons”; (7) “Giving the death penalty to juveniles who commit murder.” Although the items focused on youth justice, I believe they provide an adequate control for punitiveness. Support for punitive youth justice policies is highly correlated with general punitiveness toward criminals (Mears et al. 2007). Indeed, the correlation is sufficiently large that separate items asking about juvenile justice and criminal justice policies are often combined into a single measure of punitiveness (Johnson 2009; Pickett and Baker 2014; Welch et al. 2010).

Finally, consistent with prior research (Burns and Gimpel 2000; Stupi et al. 2014; Wang 2012), I also control for the characteristics of respondents’ local environment. Percent Latino is the percentage of county residents who identified themselves as Latino in the 2010 US Census.Footnote 10 Latino Growth is the difference between 2000 and 2010 in the percentage of the county that was Latino. To control for the local political environment, I include an indicator of the percentage of the county that voted for John McCain in the 2008 Presidential election (Percent Republican). Percent Unemployed is an indicator of local economic conditions, and is measured as the average unemployment rate in the county in 2009, as documented by US Bureau of Labor Statistics. Homicide Rate is the average homicide rate in the respondent’s county between 2007 and 2009, the 3 years preceding the survey.Footnote 11 To correct for skewness, I use the natural logs of the Percent Latino, Percent Unemployed, and Homicide Rate variables in the regression models. I also control for whether the respondents reside in the four states—Arizona, New Mexico, Texas, and California—that border Mexico (Border State = 1).

Analytic Strategy

To examine the relationship between perceived Latino threat and support for expanded police powers, I use both bivariate and multivariate methods. I asses general support for expanding police powers, and then I separately investigate support for each of the three specific police powers that make up the Expanded Police Powers index. In the multivariate analyses, because the dependent variables are all approximately normally distributed continuous measures, I use ordinary least squares regression to estimate the models. Correlated errors are a concern due to the nested nature of the data—the 961 respondents are clustered in 596 counties. Therefore, all of the models were estimated using robust standard errors, clustered at the county level. This was done using Stata’s vce(cluster) command, which generates robust standard errors that are unbiased in the context of clustered data by generalizing the Huber/White/sandwich estimate of variance to account for clustering (see Rogers 1993; Williams 2000). The models also include a quadratic term for Age to account for this variable’s nonlinear relationship to views about police powers, which was detected in a preliminary analysis. At varying points during the review of the results, I discuss the indirect associations of different variables to the outcome. In accordance with the best practices in mediation analysis, and following scholars’ suggestions not to use Baron and Kenny’s (1986) causal steps procedure or the Sobel test, I assessed indirect associations using the product of the coefficients approach with percentile-based bootstrap (k = 5000) confidence intervals (Hayes 2009, 2013; Zhao et al. 2010).

Findings

As a first step in the analysis, I assess the bivariate correlations between perceptions of Latino political and economic threat and support for expanding police officers’ investigative powers. The correlations are presented in Table 1. As predicted, significant positive correlations (r = .195 and .204, respectively) exist between both perceived Latino Political Threat and perceived Latino Economic Threat and the desire to Expand Police Powers. These positive correlations provide initial evidence that respondents who conceive of Latinos as threats to their racial group’s political influence and economic interests are more supportive of taking steps to strengthen police officers’ ability to investigate suspects. Other significant findings also emerge. Most notably, the evidence suggests that support for increasing police powers tends to be significantly lower among Blacks, and significantly higher among older persons, political conservatives, punitive respondents, those who are racially resentful, and individuals who perceive a greater risk of criminal victimization.

I turn now to the multivariate results, which are presented in Table 2. Models 1 and 2 in Table 2 separately introduce the two measures of perceived Latino Threat to examine whether concerns about Latinos’ political influence and use of economic resources, respectively, are correlated with the willingness to provide police officers with greater investigative powers. The results are fully consistent with minority group threat theory. In model 1, perceived Latino Political Threat is significantly and positively (b = .212, p = .001) associated with support for expanding police powers. Likewise, in model 2, perceived Latino Economic Threat is significantly and positively (b = .172, p = .002) associated with support for expanding police powers. The relationships between perceived Latino threat and attitudes toward expanded police power are not trivial in magnitude. For example, setting the controls at their means, a small shift in threat perceptions from “somewhat agree” to “strongly agree” on the Latino Political Threat and Latino Economic Threat measures, respectively, increases the predicted level of support for expanding police powers by 5 % (4.154–4.365) and 4 % (4.052–4.225). The models account for approximately 25 % of the variation in views about police powers.

Table 2 Regression of support for expanding police powers on perceived latino political and economic threat and controls
Full size table

Seven of the control variables also have significant (p < .05) associations with support for expanding police powers in both models. Respondents who older, politically conservative, punitive, and who perceive a greater risk of criminal victimization are all more likely to favor an expansion of police officers’ investigative powers. By contrast, Black respondents, those residing in a household where someone has been criminally victimized in the past 5 years, and those residing in counties with higher unemployment rates are less likely to support expanding police powers. Contrasting the bivariate findings, Symbolic Racism is no longer significantly associated with policing attitudes. The results from supplementary analyses (not shown) indicate that the relation of Symbolic Racism to policing attitudes is fully mediated by increased Punitiveness (indirect relationship: b = .212, 95 % CI = .140–.295).

The standardized (Beta) coefficients in Table 2 show that the two measures of perceived Latino threat both have smaller associations with policing attitudes than such factors as respondents’ race, political ideology, punitiveness, and perceived victimization risk. This fact, however, does not reduce the theoretical or practical importance of the findings. Minority group threat theory does not argue that perceived threat will be the strongest predictor of policy views, but only that it will influence such views. Additionally, from a policy standpoint, what is of utmost interest is the potential impact of public attitudes, and thus of determinants of such attitudes, on crime policy. However, it is not clear at what threshold changes in public attitudes translate into meaningful effects on crime policy. Indeed, political scientists suggest that very small changes in public opinion can have large effects on policymaking (Stimson 2004: 158). And there is strong evidence that public opinion exerts sizeable effects on both criminal justice policy (Nicholoson-Crotty et al. 2009; Enns 2014) and practice (Baumer and Martin 2013; Brace and Boyea 2008).Footnote 12 It is also important to point out that in addition to the direct relations of Latino Political Threat and Latino Economic Threat on support for expanding police powers, both variables also have significant indirect associations with these policy attitudes through Punitiveness (b = .079, 95 % CI .035 to .128, and b = .067, 95 % CI .031–.107, respectively).

Next, I unpack the relationship between perceived Latino threat and support for expanding police powers by separately exploring the influence of the former on views about the three forms of police power—police profiling, police searches, and police use of force—that collectively make up the Expand Police Powers index. The goal here is to investigate whether perceived threat has a similar influence on attitudes toward expanding each of these three forms of police power, or alternatively if the impact of threat is concentrated most heavily in the case of certain forms of investigative power. This portion of the analysis tests my second hypothesis, which predicts that perceived Latino threat will be most consequential for views about those types of police power that have the clearest potential to result in ethnic discrimination, such as the ability to stop individuals simply because of the way they look.

Tables 3 and 4 display the relevant regression results. Inspection of the Tables reveals that perceived Latino political and economic threat both have significant or marginally significant positive associations with support for all three forms of police power. However, for both measures of perceived threat, the coefficients are much larger, and are more significant, for police profiling than for police searches or police use of force. The evidence thus suggests that while respondents who perceive greater political and economic threat from Latinos are on average more supportive of expanding police powers generally, they are especially likely to favor allowing police officers to stop and question individuals based on the way they look.

Table 3 Analysis of the relationship between perceived latino political threat and support for expanding specific police powers
Full size table
Table 4 Analysis of the relationship between perceived latino economic threat and support for expanding specific police powers
Full size table

In the final portion of the analyses, I disaggregate the sample on the basis of race and ethnicity, and separately estimate the models for Whites, Blacks, and Latinos.Footnote 13 The disaggregated results are presented in Table 5. The findings reveal the perceived Latino threat is only a significant predictor of attitudes toward police powers among White respondents. Specifically, among Whites, both measures of perceived Latino threat are significantly associated with support for expanding police powers generally, as well as with views about police profiling and police searches specifically. Among Whites, as with the full sample, perceived Latino Threat appears to be more strongly related to views about police profiling, than to views about police searches or police use of force. Setting the controls at their means, a small shift in threat perceptions from “somewhat agree” to “strongly agree” on the Latino Political Threat and Latino Economic Threat measures, respectively, increases Whites’ predicted level of support for police profiling by 6 % (4.043–4.299) and 7 % (3.953–4.224).

Table 5 Analyses disaggregated by race
Full size table

Among Blacks and Latinos, by contrast, there are no significant, or even marginally significant, relationships between perceived Latino threat and support for expanding police powers. It bears noting that some of the coefficients that emerge for Blacks and Latinos are actually larger than those for Whites, while other coefficients are smaller than or even in the opposite direction of those for Whites. The relatively large standard errors for Blacks and Latinos reflect the smaller sample sizes for these groups. If the sample sizes for Blacks and Latinos were equal to that for Whites, the standard errors for Blacks and Latinos would be virtually identical to those for Whites, ranging from .055 to .093 for Blacks, and from .080 to .113 for Latinos.

It remains true, however, that the probability that the regression coefficients reported in Table 5 for Blacks and Latinos would be observed simply by chance as a result of random sampling error exceeds all conventionally accepted limits for significance (p > .10). Additionally, in the models for Latinos there are less than four respondents per variable, which falls under the recommended minimum sample size for multiple regression models (see Allison 1999: 9). And unlike the coefficients for Whites, which are consistently positive and are robust to model specification, the coefficients for Blacks and Latinos are sensitive to model specification, changing in size and direction across the various supplementary models that I estimated (e.g., if state-level controls are included, or if the control variables are dummy coded). Readers should thus use caution when interpreting the results for Blacks and Latinos. The most reliable inference from the data is that among Whites, there is a significant positive relationship between perceived Latino threat and support for expanding police powers, especially those powers used to stop and search suspects.

Sensitivity Analyses

In this section I report the results from several sets of supplementary analyses (not shown) that examine the sensitivity of the findings to alternative model specifications. First, it is possible that the broader state environment may be more important than county context in shaping attitudes, and thus that state-level contextual conditions may serve as key sources of omitted variable bias in the relationship between perceived threat and support for expanding police powers. Accordingly, I reestimated all of the models after replacing the county-level controls with their state-level counterparts (e.g., State Percent Latino, State Percent Unemployed). The findings regarding the relationship between perceived Latino threat and support for expanding police powers remained unchanged.

In a second set of supplementary analyses, I reestimated all of the models after using dummy coding for the three ordinal control variables: Age, Education, and Income. Specifically, in these models I included dummy variables for each response category of the ordinal control variables, with the exception of the lowest category, which served as the reference group. The results were substantively identical to those reported in the manuscript.

I also estimated a set of supplementary models in which I disaggregated the index measuring perceived Latino Political Threat, and separately included each political threat item in the models. The results for each item were substantively similar to those obtained with the Latino Political Threat index. The coefficients for both measures were significant or marginally significant in all of the models in which the coefficients for the Latino Political Threat index were significant. For each measure, the coefficient was larger and more significant in the model predicting views about police profiling, than in the models for views about the other types of police power, and each measure was significantly related to views about police power only among whites.

In another set of supplementary analyses, I simultaneously entered the measures of Latino Economic Threat and Latino Political Threat into the models. The results again were similar to those reported in the paper. Both threat measures had significant or marginally significant associations with the Expand Police Powers index. They also both had significant or marginally significant associations with support specifically for Police Profiling, and there was a marginally significant relationship between the political threat measure and support for Police Searches. However, neither variable was significantly related to views about Police Use of Force. In the racially disaggregated analyses, the two threat measures only had significant or marginally significant associations with policing attitudes among whites.

Finally, because the Punitiveness, Symbolic Racism, and Perceived Risk indices were all constructed by averaging separate survey items, I reestimated all of the models after including alternative versions of the indices that were generated from the predicted component scores following a principal components analysis (PCA) of the items in each index. The results that emerged with the PCA-generated indices were substantively identical to those obtained with the mean indices, but the levels of explained variance in the models were generally lower.

Discussion and Conclusion

As a result of crimmigration, “immigration enforcement is now deeply intertwined with the local enforcement of criminal law” (Eagly 2013: 1130), such that “enforcement, detention, and removal of immigrants [currently] pervade every aspect of the criminal justice system” (Vazquez 2011: 642). State and local police increasingly function as the gatekeepers in this process—they exercise “the discretion that matters” by deciding who to investigate and arrest (Motomura 2011:1822). At the same time, mounting evidence indicates that public opinion on immigration largely reflects veiled sentiments about Latinos (Ayers et al. 2009; Hartman et al. 2014; Lu and Nicholson-Crotty 2010). In this context, it is unfortunate that previous studies have not assessed whether perceptions of Latino threat influence support for aggressive policing of crime or immigration, and more generally that policing scholarship has devoted so little attention to Latino issues (Martinez 2007; Weitzer 2013).

Informed by minority group threat theory, my analysis examined whether perceptions of Latino economic and political threat were associated with support for expanding police investigative powers among a nationally representative sample of Americans. Additionally, I separately explored the relationship between perceived threat and support for expanding three specific forms of police power—the discretion to stop, search, and use force against suspects. These forms of police power influence officers’ ability not only to identify undocumented immigrants, but also to engage in aggressive policing against legally present noncitizens and Latino citizens. Consistent with minority group threat theory, the results revealed that both forms of perceived Latino threat—political and economic—were associated with greater support for expanding police powers, at least among White respondents. The evidence also showed that this relationship was strongest in the case of attitudes toward police profiling. This finding suggests that perceived Latino threat may be most consequential for public views about police policies that have clear potential to result in discrimination against Latinos.

Previous studies of public views about police policies have yielded mixed findings about whether racial attitudes, political ideology and anxieties about victimization are related to support for aggressive policing, but have generally shown that respondents’ race and gender are relatively consistent and strong predictors of such attitudes (Arthur and Case 1994; Barkan and Cohn 1998; Gabbidon and Higgins 2013; Gabbidon et al. 2009; Higgins et al. 2010; Johnson and Kuhns 2009; Johnson et al. 2011; Weitzer and Tuch 2002). In the current study, perceived Latino threat was only slightly less strongly correlated with overall support for expanding police powers than respondents’ race, but was more strongly correlated with the former than respondents’ gender. In the case of police profiling specifically, perceived Latino threat was more strongly correlated with support for profiling than either respondents’ race or gender. Perceived Latino threat was also more strongly correlated with support for profiling than symbolic racism or perceived victimization risk, and was only slightly less strongly correlated with support for profiling than political ideology.

An important implication of the results is that widespread perceptions of Latino threat (see Chavez 2008, Román 2013) may constitute an important social foundation for crimmigration. That is, by increasing public support for aggressive policing, or, at minimum, by reducing opposition to discriminatory social controls such as police profiling, threat perceptions may increase the political attractiveness and viability of crimmigration as a “solution” to the immigration problem. To be clear, the findings do not suggest that individuals necessarily view aggressive policing as a mechanism for intensifying immigration enforcement, though some may. Rather, they suggest that some Whites understand that the negative consequences of aggressive policing fall disproportionately on Latinos, and that these Whites appear to become less concerned about such discrimination as they become more anxious about the threat posed by Latinos. Thus, in turn, high levels of perceived threat in society may create a social environment that is more favorable toward, or at least more acquiescent about, the types of policies and practices involved in crimmigration. It bears emphasizing, however, that growing evidence suggests that among many Americans, the “immigration problem” is simply one dimension of a larger diversity problem—the Hispanicization of America (Hartman et al. 2014; Lu and Nicholson-Crotty 2010). For such individuals, aggressive policing within the context of crimmigation may hold additional threat-reduction benefits not provided by more traditional forms of immigration enforcement—namely, the potential to facilitate the deportation (for noncitizens), incapacitation, and disenfranchisement of Latinos who are in the US legally.

To build on the findings presented herein, future research should examine the extent to which local criminal justice authorities are responsive to community members’ perceptions of Latino threat and levels of support for aggressive policing. Recent evidence suggests that popular support for punitive policies, such as the death penalty, can influence criminal justice practice and policy (Baumer and Martin 2013; Nicholoson-Crotty et al. 2009). However, less is known about how public opinion shapes local policing, especially in the context of immigration enforcement (but see Stults and Baumer 2007). Yet, local authorities have not responded uniformly to crimmigration (Ramakrishnan and Wong 2010). In some communities, local police have adopted anti-immigrant practices, whereas in others they have resisted involvement in immigration enforcement or have developed pro-immigrant procedures (Eagly 2013; Provine et al. 2012). It seems plausible that community-level attitudes may help to explain this variation in enforcement approaches across local jurisdictions.

Another line of inquiry that merits attention in subsequent studies is exploring the antecedents of Latino threat perceptions. Few prior investigations have assessed the sources of perceived Latino threat. A larger literature, however, has explored the sources of narrower forms of threat perceptions, such as those focusing specifically on documented or undocumented immigrants. For instance, some researchers have found that contextual measures of threat, particularly Latino growth, are positively related to perceived immigrant threat (Newman 2013; Rocha et al. 2011), though the evidence has been far from consistent (see, e.g., Stupi et al. 2014; Timberlake and Williams 2012; Vallas et al. 2009; Wang 2012). Other investigations suggest that perceived immigrant composition (Wang 2012), media consumption (Dunaway et al. 2011; Watson and Riffe 2013), and beliefs about changing race relations (Higgins et al. 2010) may shape perceptions of immigrant threat. Again, however, perceived immigrant threat appears largely to constitute either a dimension or an outcome of attitudes toward Latinos, broadly (Lu and Nicholson-Crotty 2010; Valentino et al. 2013). The current study shows that broader perceptions of Latino threat are associated with support for aggressive policing, net of county- and state-level contextual factors. However, investigations are needed that explore whether factors such as contextual threat, perceived Latino composition, media consumption, and racial beliefs influence perceptions of Latino threat.

Additional studies are also needed that examine the association between perceived Latino political threat and criminal justice attitudes. Despite the centrality of political (or power) threat to minority group threat theory (see Blalock 1967), the current investigation, to my knowledge, constitutes only the third study to date that has included a measure of perceived outgroup political threat—whether in relation to Blacks, Latinos, or immigrants—as a potential predictor of criminal justice attitudes. And it is the first to show that perceived political threat is significantly associated with such attitudes. Johnson et al. (2011) and Stewart et al. (2015), by contrast, found that perceived Latino political threat did not exert a significant influence on support for allowing judges to consider offenders’ ethnicity when deciding criminal sentences.Footnote 14 However, both studies measured perceived political threat with two survey items that potentially conflated anxiety about Latinos’ political influence with favorable attitudes toward Latino political mobilization (e.g., “When Hispanics vote in local or national elections, they can influence the outcome”) (see Johnson et al. 2011: 414). In the current study, I used similar items to measure political threat, but I modified them slightly to reduce any ambiguity in meaning (e.g., “When Latinos vote in local or national elections, they can have too much influence on the outcome”). This may explain the divergence in findings across the three studies.Footnote 15 Even still, as Johnson et al. (2011: 427) stress, political threat has great theoretical relevance, and thus “a priority in future work” should be to examine alternative measures of the concept that capture perceptions about different forms of Latino political influence and mobilization.

Relatedly, there is a need for subsequent studies that examine whether other forms of perceived Latino threat influence support for expanding police powers. My objective in this study was to determine whether perceptions of Latino economic and political threat were associated with views about aggressive policing, rather than to comparatively evaluate the effects of different forms of threat perceptions. The findings indicate that both forms of perceived threat are associated with policing attitudes, though the evidence herein cannot speak to whether those associations would persist unchanged after controlling for other forms of perceived Latino threat. Indeed, to my knowledge, no previous study analyzing the relationship between perceived minority threat and criminal justice attitudes has included measures of all of the relevant forms of perceived threat—political, economic, criminal, and cultural, etc.

As noted above, I chose to focus on political and economic threat because they were both emphasized in Blalock’s (1967) original formulation of threat theory, and have also constituted key themes in public discourse on immigration (see, e.g., Brimelow 1995; Buchanan 2006). However, recent work on immigration attitudes suggests that other forms of perceived threat, particularly criminal and cultural threat (Chiricos et al. 2014; Stupi et al. 2014), may be of theoretical importance. Blalock (1967:166–168) also discusses the possibility that perceived criminal threat, which he describes as a “threat-oriented ideology,” may hasten and intensify dominant group members’ counter-mobilization against other non-criminal forms of threat: “If buses are integrated, the next thing you know they’ll be raping our women.” This may occur because perceived criminal threat may foster particularly high levels of fear and hatred, and, unlike other forms of threat, may imply the existence of internal racial differences (e.g., in personality or biology). Therefore, perceived criminal threat may contribute to the delegitimization of outgroup members’ political and economic advancements. Research is thus needed that examines the possibility that criminal threat perceptions may mediate or condition the relationships between political and economic threat and support for expanded police powers.

Although the negative consequences of crimmigration are well documented (Golash-Boza 2012: India 2013; Vazquez 2011), it is not clear that members of the public are informed about these adverse outcomes (Román 2013). For example, evidence suggests that crimmigration policies can have a detrimental effect on state and local economies (Román 2013). Recent research also shows that because immigrant concentration is negatively associated with crime (Ousey and Kubrin 2009; Stowell et al. 2009), deportations can, in some circumstances, actually increase crime rates (Stowell et al. 2013). Likewise, racial profiling resulting from local police involvement in immigration enforcement “cuts at the heart of [Latinos’] sense of belonging and essentially creates a caste system that demotes Latinos to second-class citizens” (Nill 2011: 55). Future research should explore whether informing the public about the negative effects of crimmigration policies has the potential to reduce support for aggressive policing, and to attenuate the relationship between perceived Latino threat and policing attitudes.

I close by emphasizing the need for continued research into the ways that both perceived immigrant threat and Latino threat shape attitudes toward the police, sentencing, and criminal sanctions. This is not only important because crimmigration has created a situation where “immigration law and the criminal justice system are merely nominally separate” (Stumpf 2006: 376). It is also critical because the US population will continue to diversify demographically in the coming decades (US Census Bureau 2008), which in turn will likely exacerbate perceptions of immigrant and Latino threat. For example, in a seminal study, Craig and Richeson (2014a) recently showed that among Whites, exposure to information about the US Census Bureau’s projections of demographic change increased perceived outgroup threat as well as racial and ethnic animus. A second study by these authors found that the same experimental manipulation, by increasing perceived threat, fostered more conservative policy preferences on a variety of fronts, including immigration (Craig and Richeson 2014b). To the extent that these experiments accurately forecast the public’s reaction to increased diversity, the findings in the current study suggest there may be a concurrent increase in support for aggressive police practices, such as police profiling. This possibility deserves close monitoring in future research.