Abstract
Immigration is at the forefront of the nation’s public policy agenda. Despite research finding the contrary, a primary concern held by some Americans is that immigration increases crime in the United States. The current study embeds an information experiment in a survey of 430 U.S. adults to assess whether fact-based information on the immigration-crime relationship influences attitudes toward immigration and its effect on crime. Results reveal that the presentation of fact-based information reduced negative perceptions toward the immigration-crime relationship only when participants were exposed to the information for 10 seconds or longer. Additionally, geographical, political, and emotional factors shaped attitudes. Policy implications are discussed.
Within the U.S., immigration policy is highly debated in public discourse. For more than 40 years, Americans have been asked their thoughts about the level of immigration into the U.S. (Gallup, 2024). Desires to reduce immigration steadily increased from 1965 to 1995, with a high of 65% of Americans expressing immigration should be decreased in 1995. This desire to reduce immigration then tumbled from 1995 to 2020 reaching a low of 28%. However, recent polls reveal that Americans’ desire to reduce immigration had its sharpest increase in the poll’s history between 2020 and 2024, with 55% wanting immigration levels to decline in 2024. Furthermore, recent Gallup polls indicate that a record number of Americans (55%) perceive immigration as a “critical threat” to the U.S., and more than one in four Americans (28%) rank immigration as the most important problem facing the nation (even ranking it higher than the economy, inflation, and crime; Gallup, 2024; Jones, 2024). Although some Americans are concerned with the impact immigration may have on their jobs and American culture generally (Chiricos et al., 2014), most of the fear surrounding the topic stems from concerns about safety and rising crime rates (Chouhy & Madero-Hernandez, 2019; Stupi et al., 2016). It is these concerns and fears that are often reinforced by politicians and the media. Consequently, these sources exacerbate the American public’s apprehension surrounding immigration (Chouhy & Madero-Hernandez, 2019; Grigorieff et al., 2020).
Despite immigration being a central issue in the news media and in political discourse, the public remains highly misinformed about the size and characteristics of the immigrant population as well its potential impact on crime in the U.S. (Grigorieff et al., 2020). Previous research suggests that these misperceptions are likely influenced by conservative, political, and negative media rhetoric (Grigorieff et al., 2020; Herda, 2019). For instance, in June 2015, while campaigning for president, Donald Trump made the following statement about immigrants: When Mexico sends its people, they’re not sending their best. They’re sending people that have a lot of problems, and they’re bringing those problems with us. They’re bringing drugs. They’re bringing crime. They’re rapists. And some, I assume, are good people. (Mark, 2018)
Additionally, in 2019, Trump characterized the U.S.-Mexican border crisis as a “growing humanitarian and security crisis” (Clark, 2019; Varvel, 2022). These comments were not only confined to solely the leader of the Republican party but were also articulated by other public officials and segments of the media. Furthermore, immigration is regularly characterized as a “Mexican invasion,” despite the majority of immigrants crossing the border coming from other Central American countries rather than Mexico (Varvel, 2022). These political and media portrayals of immigration not only criminalize individuals crossing the border but also (1) intensify the political divide on immigration issues; (2) garner support for restrictive anti-immigration policies; and (3) perpetuate the myth of an immigration-crime link (Mills et al., 2023).
Despite the negative rhetoric espoused by the media and politicians and the use of derogatory terminology such as “illegal aliens” when describing immigrants, empirical research consistently demonstrates that the relationship between immigration and crime is null and, in some cases, negative (Bersani & DiPietro, 2016; Chouhy & Madero-Hernandez, 2019; Ousey & Kubrin, 2018). Similar to the current study, prior work examines how aspects of immigration (e.g., the general size and characteristics of the immigrant population) impact attitudes toward it (Abascal et al., 2021; Carnahan et al., 2020; Grigorieff et al., 2020). Although these studies demonstrate that the presentation of some information can reduce negative attitudes toward immigrants, no study, to our knowledge, has provided information to respondents showing the lack of an immigration-crime link and assessed subsequent attitudes toward immigrants and immigration. This oversight is critical because U.S. immigration laws and the American public’s view of immigrants tend to be driven by the false belief that immigrants increase crime in the U.S. (Chouhy & Madero-Hernandez, 2019).
Within this context, the current study extends the existing literature in the following three ways. First, the study experimentally examines how factual information depicting the immigration-crime relationship impacts public attitudes about immigration. Second, the study assesses whether relevant social and demographic characteristics are related to attitudes about crime and immigration. Finally, the duration of information exposure on attitudes toward a perceived immigration-crime link is explored. The relevant background literature is discussed below.
The Public Salience of Immigration
The U.S. is home to more than 328 million people, with 45 million (or 13% of the total population) classified as immigrants (Bernat, 2019; Budiman, 2020). Among immigrants, it is estimated that approximately 75% are legally living in the U.S. (Bernat, 2019; Budiman, 2020). By 2065, the U.S. foreign-born population is expected to grow significantly, reaching an estimated 78 million (Budiman, 2020). As a greater number of immigrants seek to enter the country both legally and illegally, the salience of immigration among American public attitudes and national policy will only continue to grow.
Previous research shows that older, male, conservative, and less educated Americans are more likely to possess anti-immigration attitudes than other members of the public (C. Jackson & Newell, 2018; Pryce, 2018; Stupi et al., 2016). More specifically, these groups are more likely to believe that immigrants increase the crime rate, take jobs away from native-born citizens, and undermine American culture (Pryce, 2018). Beyond these factors, political ideology and party identification heavily shape immigration attitudes (C. Jackson & Newell, 2018). In a polarized political environment, like the American political system, Lupu (2015) and Levendusky (2009) argue that partisanship acts as a heuristic for citizen’s attitudinal and behavioral cues. Thus, citizens are more likely to hold attitudes and behave in ways similar to the parties they identify with. Although both Democrats and Republicans hold strong opinions about immigration, they differ in what they perceive as important for U.S. immigration policy. For example, Republicans generally support broader efforts aimed at reducing immigration and addressing illegal immigration, whereas Democrats prioritize the creation of easier paths for legal status and sponsorship of family members who immigrate (Oliphant & Cerda, 2022).
Previous research also finds several spatial and contextual factors that moderate the effect of party identification on attitudes toward immigration (Gravelle, 2016). One such factor is proximity to the U.S.-Mexico border. Some scholars have theorized that proximity to the border, or exposure to immigrants in general, facilitate contact with immigrant populations and reduce prejudice toward them (Cortina, 2020; Walker et al., 2022). Greater exposure to immigrants and direct experience with the border can result in people becoming more comfortable with these populations, reducing the need to rely on media and political rhetoric for understanding them and their experiences (Cortina, 2020). Additionally, increased exposure may work to diminish Americans’ perceptions of immigrant groups as criminal threats, which hold some of the greatest influence on support for more punitive immigration policies (Chiricos et al., 2014; Stupi et al., 2016). Of the few studies examining these relationships, mixed results regarding the effect of geographic proximity or immigrant contact on immigration and policy attitudes were found (Gravelle, 2016; Walker et al., 2022). However, when geographic proximity to the border is assessed alongside party identification, the results become clearer. For instance, Cortina (2020) found that as the distance from the U.S.-Mexico border increases, Republicans are more likely to support building a wall due to lack of direct contact with immigrants. In general, it appears that experience with the border, regardless of partisanship, shapes public attitudes toward immigration. However, if there is distance from the border, the public relies more on partisan cues (Cortina, 2020).
The Myth of Immigrant Criminality
The idea that immigrants commit more crime than native-born Americans is grounded in the belief that immigrants themselves possess inherent criminality or are “bad” people (Chouhy & Madero-Hernandez, 2019). Former President/President-elect Donald Trump, for example, referred to migrants entering the U.S. as “people coming in from prisons and jails,” “long-term murderers,” “animals,” and “not people” (Barkan & Rocque, 2024). Among the public, in a survey of White adults, respondents high in Hispanic resentment and White nationalism consistently perceived immigrants as criminogenic (Kulig et al., 2021), while a survey of college students associated the word “immigrant” with phrases like “illegal, foreign, and Mexican” (Alcalde, 2016). In addition to being portrayed as dangerous and criminal, much political discourse has described immigration on the southern border of the U.S. as an “invasion” (LeVine & Sacchetti, 2023; Schladen, 2024). These depictions fuel the negative perceptions of immigration and immigrants (Alcalde, 2016; Kulig et al., 2021). As posited by group threat theorists, when people perceive minority groups (who are already portrayed in a negative light) to be more sizable, perceptions of the minority group tend to be less favorable and illicit greater fear (Chiricos et al., 2020; Hopkins et al., 2018; Stupi et al., 2016)
Notably, this tendency to villainize immigrants has not targeted all immigrants entering the U.S. Instead, the negative rhetoric generally is reserved for those coming from Mexico and other Central America countries. These portrayals have fed into the “Latino threat narrative,” which posits that Latinx people are not like previous immigrant groups who assimilate into the national culture. Instead, Mexican and Latinx immigrants are viewed as unwilling or incapable of integrating into the national culture, leading people to view them as more threatening and equating their “Mexicanness” with illegality and dangerousness (Alcalde, 2016; Chavez, 2008; Golash-Boz & Hondagnue-Sotelo, 2013; Jones, 2024; Pickett, 2016).
Despite criminogenic depictions of immigrants, much of the empirical research has debunked this myth by finding that immigrants, both those residing in the U.S. legally and illegally, commit crimes at a significantly lower rate than native-born citizens (Chouhy & Madero-Hernandez, 2019; Mears, 2002; Ousey & Kubrin, 2018). For example, one study found that not only did undocumented immigrants have a criminal conviction rate 50% lower than native-born Americans, but legal immigrants also had a criminal conviction rate 66% lower than native-born Americans (Nowrasteh, 2018). Similarly, Light et al. (2020) found that U.S.-born citizens are more than two times as likely to be arrested for violent crimes, 2.5 times more likely to be arrested for drug crimes, and more than four times as likely to be arrested for property crimes than undocumented immigrants.
Furthermore, research has shown not only a negative correlation between immigration and crime but also, in some cases, a protective effect—where communities with higher immigrant populations experience lower crimes rates and are better able to prevent criminal behavior among its residents (Macdonald & Sampson, 2012; Ousey & Kubrin, 2018; Wolff et al., 2018; Wong, 2017). For instance, Wolff et al. (2018) found that immigrant concentration in neighborhoods had a direct negative effect on youth recidivism. Additionally, it increased the likelihood of the young justice-involved individuals living with both parents, having more community relationships, and having lower familial incarceration history (Wolff et al., 2018). Moreover, sanctuary cities, or cities with laws to protect undocumented immigrants from deportation and prosecution, have been found to have lower crime rates than non-sanctuary cities (Chouhy & Madero-Hernandez, 2019; D. E. Martinez et al., 2018; Wong, 2017).
Overall, the immigration-crime research has reached similar conclusions: “immigrants nearly always exhibit lower crime rates than native groups” (R. Martinez & Lee, 2000, p. 496). However, a more nuanced exploration finds that the protective impact of immigration on crime diminishes the longer immigrants and their children live in the U.S. (Kubrin & Miouduszweski, 2018). This phenomenon, also known as the “immigrant paradox” or the “assimilation paradox,” generally argues that as immigrants assimilate into U.S. culture, their developmental outcomes (e.g., educational, academic, and social) decline (Marks et al., 2014; Peguero, 2012; Peguero et al., 2016). Furthermore, the paradox argues that the children of immigrants who are born in the U.S. exhibit higher offending rates than their parents (Rumbaut & Ewing, 2007). Additionally, the risk of incarceration is higher not only for the children of immigrants but also for immigrants themselves the longer they reside in the U.S. (Rumbaut & Ewing, 2007). In other words, the protective impact of immigration dissipates the longer immigrants and their children reside in the U.S. and become more Americanized.
Despite research overwhelmingly finding no evidence that immigrants commit more crime than native-born Americans, the public still links immigration to crime (Jiang & Erez, 2018). As a result, the enforcement of immigration laws has been heavily tied to public safety (Chouhy & Madero-Hernandez, 2019). The steady increase in deportations since 2003 and the strict enforcement of various immigration policies that have expanded the scope of deportable offenses, removed immigrants through interior and administrative arrests, and separated families at the border illustrate this point (Chouhy & Madero-Hernandez, 2019; Pickett, 2016). Consequently, the American public and political systems continue to perpetuate the misconception that immigration is linked to crime, further reinforcing policies that assume immigrants are a major contributor to crime in the U.S. (Adelman et al., 2021).
Information Effects and Treatment Dosage
The breadth of research on the null/negative immigration-crime relationship has led some researchers to claim there is an “emerging scholarly consensus” on the matter (Bersani & DiPietro, 2016). Despite such findings, many Americans still embrace these myths and misperceptions which in turn influence their attitudes regarding immigration. Some researchers have attempted to provide factual information to U.S.-based samples about immigrant populations to assess whether it can alter attitudes (Abascal et al., 2021; Carnahan et al., 2020; Grigorieff et al., 2020; Hopkins et al., 2018; Sides & Citrin, 2007). Although some research has found evidence that factual immigration information challenges immigrant misbeliefs and misinformation (Abascal et al., 2021; Carnahan et al., 2020; Facchini et al., 2022; Grigorieff et al., 2020), other research has not (Hopkins et al., 2018; Sides & Citrin, 2007).
Furthermore, and a key focus of the current study, is a methodological assessment of how the duration of exposure to such factual information affects attitudes toward the immigration-crime link. Of the existing research on immigration information provision experiments, only two have fully explored the topic related to exposure to and duration of treatment. These studies both examined the lasting effect of information effects on immigration attitudes (Carnahan et al., 2020; Facchini et al., 2022). For example, Facchini et al. (2022) found that the effects of exposure to factual information on immigration (e.g., immigration can help address the labor shortage in certain sectors) persisted 10 to 12 days after the original intervention. The second study found that the effect of factual information on immigration attitudes can last up to 4 weeks (Carnahan et al., 2020). On a related inquiry, other scholars have explored the topic of repeated exposure on policy attitudes. It is generally accepted that repeated exposure to information is more impactful than a single dose of exposure (Mutz, 2008). This applies for both misleading and truthful information. For example, Foster et al. (2012) examined whether individuals would find information more credible if it came from multiple sources or if it was repeated. They found that people were more misled by repeated misinformation rather than misinformation coming from multiple sources. Foster et al. (2012) argue that this is likely because the repeated information/misinformation is being interpreted as markers of accuracy.
Although few studies have explored the enduring effects of information exposure, no study to date (to our knowledge) has explored the impact of information exposure duration (i.e., the amount of time spent consuming and engaging with factual information) on immigration-crime attitudes. Put differently, it is theorized that more fully engaging with factual information (i.e., spending more time with it) may result in the information having a greater impact on respondents’ attitudes compared to individuals who simply skim or glance over it. The results of this assessment have implications for the way that the media or other forms of news portray information on immigration (e.g., short form media [e.g., social media posts] or long form media [e.g., lengthier, several paragraph stories]). Thus, the current study allows for an examination of this underexplored inquiry.
Research Hypotheses
The current study investigates whether factual information about the immigration-crime relationship impacts individuals’ attitudes toward immigration and its effect on crime. Given that previous research finds immigration attitudes are affected by information on other immigration topics (e.g., the size and characteristics of immigrants, their help with labor shortage in certain sectors; Abascal et al., 2021; Carnahan et al., 2020; Facchini et al., 2022; Grigorieff et al., 2020), the current study hypothesizes that:
H1: Exposing individuals to factual information about the immigration-crime relationship will be associated with a decrease in respondents’ attitudes that immigration increases crime.
Additionally, the study explores whether the extent of exposure to treatment duration impacts these attitudes. This hypothesis is postulated due to prior research indicating that increased and repeated exposure to information content creates longer-lasting effects than a single dose of exposure (Carnahan et al., 2020; Facchini et al., 2022; Mutz, 2008). Thus, it is hypothesized that:
H2: Spending at least 5 seconds viewing the factual information will be associated with a decrease in respondents’ attitudes that immigration increases crime in the U.S.
H3: Spending at least 10 seconds viewing the factual information will be associated with a decrease in respondents’ attitudes that immigration increases crime in the U.S.
H4: Spending at least 15 seconds viewing the factual information will be associated with a decrease in respondents’ attitudes that immigration increases crime in the U.S.
H5: Within the treatment group, exposure to treatment duration will be associated with a decrease in respondents’ attitudes that immigration increases crime in the U.S.
In summary, significant and negative results for all five hypotheses would convey that exposure to information about the immigration-crime relationship decreases the respondents’ attitudes that immigration increases crime in America.
Methods
Data Collection and Sample
The current study relies on survey data that included a variety of questions regarding beliefs, attitudes, perceptions, and opinions on immigration in the U.S. Items are also included in the survey to capture the respondents’ demographics. The survey was reviewed and approved by The University of Texas at Dallas’ Institutional Review Board (Approval #IRB-22-2150). The survey was then fielded to individuals (n = 430) on Amazon’s Mechanical Turk (MTurk) platform on January 17, 2022.
MTurk is extensively used in social science research because it allows researchers to collect large datasets quickly and inexpensively. Moreover, MTurk samples are more representative of the U.S. population’s demographics compared to other commonly used samples in the social sciences, such as college student samples (Cheung et al., 2016). As such, researchers in criminal justice use MTurk to explore diverse topics such as public opinion toward policing policies, the belief that people with criminal records can change, and punitive criminal justice policies, among others (Burton et al., 2020; Graham et al., 2020; Shi, 2022).
Amazon MTurk also offers the ability for researchers to use qualification requirements (e.g., place of residence, number of approved tasks) within surveys to determine which MTurk workers are eligible to participate (Cheung et al., 2016). For this study, the following eligibility requirements were used: MTurk workers had to be 18+ years old and reside in the U.S., have a 95% or higher human intelligence task (HIT) approval rate, and have 500 or more approved HITS were allowed to participate in the survey. A HIT is a single, virtual task that an MTurk worker can work on, submit an answer, and collect a reward for completing. These qualifications were set in place to ensure that the data came from a reliable sample. All participants who completed the survey, passed all attention checks, and provided accurate survey codes and MTurk worker IDs were paid $1.00. This incentive is higher than what is recommended for standard behavioral science studies, which is 10 cents per minute (Moss & Litman, 2019). Most participants completed the survey within 5 to 7 minutes, meaning they were compensated at a rate approximately 0.3 to 0.5 cents higher per minute than the typical 10 cents. Those who failed to complete the survey in its entirety, pass all attention checks, and provide accurate codes/IDs were rejected and not paid for their work. The final sample size was 430 participants.
Experimental Procedure
This study uses an information provision experimental design, which is commonly used in economic and policy research to measure the attitudes and perceptions of people on real-world phenomena (Coppock, 2022). The information in this study is a fact sheet depicting recent immigration-crime research. It includes facts dispelling the misperceptions that immigrants and immigration bring more crime to the U.S. (see Appendix A for the full fact sheet). The experimental aspect of the study involved assigning the fact sheet to half the sample at random. After randomization, 212 participants received the fact sheet (treatment group), and 218 participants did not (control group). By randomizing the survey respondents into these treatment and control conditions, the causal effect of the information on the outcome can be estimated.
On average, the participants who received the fact sheet took 6.4 minutes to complete the survey while those who did not receive the fact sheet took 4.9 minutes. The treatment group read the fact sheet and then answered a series of Likert-scale questions tapping into their attitudes, opinions, and beliefs toward immigration and immigrants in the U.S. The control group just answered the questions. These measures are described below.
Measures
Dependent Variable
Immigration increases crime is the primary dependent variable that captures whether respondents believe immigration increases crime in the U.S. It is a four-item mean index (α = .93) that assesses respondents’ levels of agreement (1 = strongly disagree, 5 = strongly agree) to the following statements: (1) Foreign-born people are more likely to commit crime than U.S. born people; (2) Foreign-born people are more violent than U.S. born people; (3) Crime will increase as more foreign-born people come to the United States; and (4) Foreign-born people are more likely to be arrested for violent, property, and drug crimes than U.S. born people. The index is adapted from and informed by the Gallup and Ipsos surveys on American attitudes of immigration (Gallup, 2024; C. Jackson & Newell, 2018). Higher values on this index correspond to greater levels of belief that immigration increases crime in the U.S.
Independent Variables
The primary independent variable is the randomized treatment variable. The informational fact sheet is a one-page document discussing contemporary research on the immigration-crime relationship (see Appendix A). It includes several bullet points highlighting the negative relationship between immigration and crime. Most of the research in the fact sheet highlights how immigrants are less likely to commit crimes than U.S.-born people (Light et al., 2020; Light & Miller, 2018; Nowrasteh, 2018; Ousey & Kubrin, 2018; Ramos & Wenger, 2020; Wolff et al., 2018). The measure is coded such that 1 indicates receiving the fact sheet and 0 indicates not receiving the fact sheet. It appears in the tables as Tx = 1.
Comfort with immigrants is a five-item mean index (α = .74) that assesses respondents’ levels of support (1 = strongly oppose, 5 = strongly support) for the following statements and policies: (1) Immigration is good for the United States; (2) I am satisfied with the level of immigration in the country today; (3) I feel safe around foreign-born people; (4) Allowing immigrants living in the U.S. illegally to become U.S. citizens if they meet certain requirements over time; and (5) Immigrant housing. The index is adapted and influenced by Gallup and Ipsos surveys (Gallup, 2024; C. Jackson & Newell, 2018). Higher values on this index correspond to greater feelings of comfort with immigrants.
Political views are measured in two ways. The respondents were asked, “Generally speaking, do you think of yourself as a. . .? (1) Democrat; (2) Republican; (3) Independent; (4) Other; and (5) Not sure. Thus, Republican is dichotomously recoded so that 1 indicates Republican and 0 indicates any other party affiliation. In addition to measuring political party, a variable is included to measure political ideology (1 = very liberal, 5 = very conservative). Specifically, respondents were asked, “In general, how would you describe your own political viewpoint?” Conservatism was created from this question and higher values on the variable correspond with greater levels of conservative values.
A measure for Southern border states is also included, which captures whether respondents reside in states that are located along the U.S.-Mexico border. These include Arizona, California, New Mexico, and Texas. The measure is dichotomously coded such that 1 indicates residing in any of the four southern border states and 0 indicates any other state. Employment assesses the employment status of respondents. Respondents were asked their employment status from a selection of eight items: full-time employment, part-time employment, contract/temporary work, unemployment, student, military, retired, and disabled. The item is dichotomously recoded so that 1 indicates full-time employment and 0 indicates all other options. Finally, as race is a strong correlate of policy attitudes, a measure is included to assess the race of each respondent (C. Jackson & Newell, 2018; Pryce, 2018). The item is dichotomously recoded so that 1 indicates White and 0 indicates all other races.
Control Variables
Several demographic variables are controlled for in the study’s models that might impact the relationships between the independent and dependent variables. Specifically, the models control for gender (1 = Female), Age (in years), marital status (1 = Married), education level (1 = eighth grade, 8 = Doctorate degree), and income (1 = Less than $25,000, 5 = $200,000+; see Table 1 for descriptive statistics of all variables included in the analyses). Note that Table 1 is a balance table which allows for a comparison of the demographics between the treatment and control group. Although the treatment was randomized and thus should, theoretically, create perfect balance between the groups, that does not always occur. As a result, the treatment group was significantly higher regarding their comfort with immigrants and significantly lower in their levels of Republicanism and conservatism. Note, however, that these differences were small (Z/t were less than or equal to 2.25).
Descriptive Statistics for All Study Variables, by Treatment and Control Conditions (N = 430).
p < .05. **p < .01. ***p < .001 (two-tailed).
Analytic Strategy
First, the percentage of the sample that agree with of each item in the dependent variable is calculated. This is done separately for the treatment and control groups, and then Z-tests are performed to assess whether a treatment effect exists at the individual-item level. Second, a series of ordinary least squares (OLS) regression analyses are used to test the primary hypothesis on whether factual information about immigration and crime impacts attitudes toward the issue. Finally, additionally analyses are conducted to test the treatment dosage hypotheses.
Results
Does factual information depicting the relationship between immigration and crime impact attitudes toward this relationship? As shown in Table 2, the control group has slightly more negative attitudes regarding the immigration-crime relationship than the treatment group across the four items that comprise the outcome measure of the study. However, none of the differences observed are significant (Z < 2.00).
Public Opinion Toward Immigrant-Crime Relationship, by Treatment and Control Conditions (N = 430).
Note. Although these groups are balanced on the dependent variable, living within southern border states, employment status, race, gender, age, marital status, education level, and income level, they significantly differ on their feelings of comfortability with immigrants, political ideology, and political party affiliation.
p < .05. **p < .01. ***p < .001 (one-tailed).
Similar results are shown in the OLS models presented in Table 3. Model 1 presents the results of the model with the treatment indicator variable on the outcome. As shown, whether someone in the sample received the fact sheet or not had no impact on their attitudes toward the immigration-crime relationship (b = −0.105, p > .05). Model 2 presents the same results, but with the full model including control variables. Although the coefficients for both models are in the theoretically expected direction, these results do not support the study’s first hypothesis, as the effect was not statistically significant. Additionally, the results show that respondents who are conservative, live in a southern border state (i.e., Arizona, California, New Mexico, and Texas), are employed full-time, older, married, and have higher incomes are significantly more likely to possess negative attitudes regarding the immigration-crime relationship.
OLS Regression Examining Sources of the Belief that Immigration Increases Crime.
Note. The current study assessed variance inflation factors and found that values ranged between 1.03 and 1.30 (M = 1.14). Thus, multicollinearity is not present in the model. b = unstandardized regression coefficient; RSE = robust standard error.
p < .05. **p < .01. ***p < .001 (two-tailed).
This study also assesses whether the duration of information exposure affects attitudes toward the notion that immigration increases crime in the United States. To our knowledge, no methodological research has assessed the impact of information duration cut off points in survey research experiments. Thus, the exploratory analysis arbitrarily chose the natural cut off points of 5, 10, and 15 seconds, which are reasonable amounts of time that researchers might consider survey takers would view the information in an online survey.
The results from this analysis are presented in Table 4. As shown, receiving information did not influence the respondents’ immigration-crime attitudes when exposure was set to 0 and 5 seconds. However, when exposure reached 10 seconds or longer (the two righthand columns) the information effect was significant. In other words, the fact sheet did not matter in influencing attitudes regarding the immigration-crime relationship unless the respondents spend at least 10 seconds (b = −0.307, p < .01) or 15 seconds (b = −0.347, p < .01) on the information page. At 0 and 5 seconds, respondents likely did not fully read or engage with the information and, thus, were not exposed to the information long enough for it to matter. These results support H3 and H4, which predicted that spending at least 10 seconds and 15 seconds, respectively, viewing the factual information would decrease respondents’ attitudes that immigration increases crime in the U.S. However, H2 is not supported, as no significant effect was found at 5 seconds.
OLS Regression Examining Immigration-Crime Attitudes by Time Dosage.
Note. b = unstandardized regression coefficient; RSE = robust standard error.
p < .05. **p < .01. ***p < .001 (two-tailed).
Additionally, to explore dosage efforts among those who were only in the treatment group (i.e., received the fact sheet), an analysis was done using the time spent on the fact sheet (in seconds) as the main independent variable. The results from this analysis are shown in Table 5. They show that for each 1 second increase in the time spent on the fact sheet, there is a negative and significant decline (b = −0.004) in the belief that immigration increases crime. This finding supports H5, which hypothesized that exposure duration within the treatment group would be associated with a decrease in respondents’ attitudes toward the immigration-crime relationship. Additionally, conservatism, full-time employment, marriage, and income were significant predictors on one’s belief in the immigration-crime relationship in the treated-only sample. These four variables hold significance throughout every one of this study’s models, which highlights the importance of their explanatory power on attitudes toward immigrants, immigration, and crime.
Dosage Effects Among the Treatment Group on Immigration-Crime Attitudes (N = 211).
Note. b = unstandardized regression coefficient; RSE = robust standard error.
p < 05. **p < .01. ***p < .001 (two-tailed).
Discussion
The growing concern over immigration in the United States is heavily influenced by the public’s misperception that immigrants bring more crime to the nation (Adelman et al., 2021; Chouhy & Madero-Hernandez, 2019; Jiang & Erez, 2018). Despite research finding the opposite to be true (Chouhy & Madero-Hernandez, 2019; Ousey & Kubrin, 2018), the immigration-crime relationship remained a significant issue in the 2024 presidential race (A. Jackson, 2024; Oppenheim, 2024). Previous research has examined how factual information might play a role in correcting misperceptions regarding immigrants, but, to our knowledge, no studies have explored how (1) information only discussing immigration-crime research and (2) the duration of exposure to this information may impact misperceptions (Abascal et al., 2021; Carnahan et al., 2020; Facchini et al., 2022; Grigorieff et al., 2020; Hopkins et al., 2018; Sides & Citrin, 2007). The current study expands on this research by addressing these two salient gaps in the literature.
Theoretical Takeaways
This study contributes several key findings, both theoretical and methodological. First, in virtually all the multivariate models, the results indicate that several theoretically germane variables, such as conservatism, income, and employment, are significant predictors of the belief that immigration increases crime. Previous research frequently finds that conservative individuals not only hold stronger punitive attitudes toward immigration but also are more susceptible to misinformation regarding immigrants (C. Jackson & Newell, 2018; Pryce, 2018). Second, previous studies report that individuals who are unemployed are more concerned about immigration than those who are employed (Becchetti et al., 2010). Given the stigmatizing rhetoric of immigrants “stealing jobs from citizens,” it is logical that unemployed people would hold more disdain for immigrants. However, research has shown that higher income has the opposite effect, which is confirmed in this study’s analyses. García-Muñoz and Milgram-Baleix (2021), for example, find that higher income countries are more conscious of the economic benefits brought by immigration and thus, have fewer anti-immigration opinions. It is possible that higher income is correlated with more positive views on immigration because of the increased economic security and exposure to diverse perspectives.
Although individuals earning higher income were less likely to perceive that immigration increases crime in this study, the opposite was found with employment. Unfortunately, there is limited research unpacking the relationship between employment and immigration. One dated study found that individuals who were employed were more likely to have unfavorable attitudes toward undocumented migrants compared to those who were unemployed (Espenshade & Calhoun, 1993). Being employed full-time does not guarantee economic stability, and individuals in low-wage jobs may feel threatened by immigrants, which can foster negative attitudes. Furthermore, the quality of employment matters. Those in higher-income positions may collaborate more with immigrants, contributing to greater comfortability with the population and more favorable perceptions. Although this study’s findings on income and employment are contradictory, it appears that they are tapping into different underlying constructs. Although income may reflect economic power that can lead to more positive views on immigration, employment can tap into feelings of insecurity and competition.
Finally, in the full sample, the study finds that exposure to factual information about the immigrant-crime relationship did not significantly change the respondents’ attitudes regarding the belief that immigration increases crime. There are at least two plausible explanations for these findings. First, it might be the case that the respondents who got the information spent such little time reading it, that it had no impact on their attitudes. Second, it might be the case that people are more likely to accept information that confirms their existing beliefs and political ideals, while dismissing information that contradicts them (Knobloch-Westerwick et al., 2020). In other words, when confronted with evidence that challenges their preconceptions, individuals may become more entrenched in their original views rather than adjust them to the truth. However, other research finds that exposure to information that is contrary to your attitudes does not increase opposition, but instead, it simply lessens support for the attitudes (Coppock, 2022). This might lead to small, and possibly insignificant, effects.
Methodological Takeaways
Although the information did not have a significant effect on immigration-crime attitudes in the main full model (Table 3, Model 2), the study finds that the duration of exposure to this information matters. Specifically, the results reveal that exposure to factual information regarding immigration and crime only mattered if an individual spent 10 seconds or longer reading it. The current study tested the length of time it took to read the information sheet, and it was slightly more than 1 minute. Thus, it appears that spending 10 seconds or longer with the information would allow for a quick skim, but any amount of time less than that was statistically equivalent to not reading the information at all.
To our knowledge, this study is one of the first to highlight that time spent reviewing information, in an information provision experiment, might have varying effects on the outcome—in this case views toward immigration and crime. That being said, it would behoove other researchers using experimental designs such as the current one to place holds in their surveys if the information given to respondents is lengthier (~1 minute read or longer) or complex (e.g., figures, graphs, academic text). Rather than allowing readers to skip or skim through the information, holding would force them to engage with the material more and potentially allow the ability for their attitudes to be altered (if they were going to).
Limitations and Future Research
This study is not without limitations, and four are addressed here. First, these data are drawn from Amazon’s MTurk platform, which is limited in its generalizability compared to other population-based survey vendors (e.g., YouGov; Follmer et al., 2017). Furthermore, other than the manipulated variable (the information sheet), the data are cross-sectional and thus, limited in the ability to establish causality. Future studies might consider a pretest-posttest or longitudinal design, which would help with causality via temporal ordering (Coppock, 2022). Second, there is possible omitted variable bias within the models. Other variables such as ethnicity or personal and vicarious experiences with legal and undocumented immigrants may also have been valuable to include in the study. Similarly, the Southern border states variable only examines states on the U.S.-Mexico border (i.e., Arizona, California, New Mexico, and Texas). This study’s results show that individuals living in these states are more likely to believe that immigration increases crime (see Table 3). However, previous research has shown that increased exposure to ethnic minorities is associated with a decline in the ethnic typification of crime (Cortina, 2020; Walker et al., 2022). A possible explanation for this contradictory finding is that these states have greater exposure to more conservative media that employs anti-immigrant rhetoric (Mills et al., 2023; Young et al., 2024). Future research should explore this possibility as well as include variables that examine more of the South rather than just the Southwest.
Third, the current study is unable to ensure that the participants actually read the fact sheet, which is why the information page in the survey was timed. As mentioned previously, future studies should consider adding time holds to their online surveys to encourage and ensure more engagement with the material. Fourth, only factual information on attitudes toward immigration and crime was used in this study. Future studies should explore how factual information on other topics of public concern impact views toward immigration, such as the economy and labor market. Despite these limitations, this study is among the first to explore the importance of information clearly depicting the null or negative crime-immigration link. It also tests whether the duration of exposure to such information impacts immigration policy attitudes. Future research should extend these findings while paying mind to the limitations of our study.
Conclusion
As the concern over immigration continues to grow among the American public (Jones, 2024), it is imperative that we as a nation get the facts straight on immigration’s effect on crime. Misinformation regarding immigrant crime has the potential to lead to violence and other dangerous outcomes. For example, the public saw the consequences of these misperceptions after false accusations were made regarding members of Springfield, Ohio’s Haitian immigrant community abducting and eating pets during the 2024 presidential debate. Stemming from these remarks, more than thirty bomb threats were made against Springfield’s schools, government buildings, and city officials’ homes. These threats led to evacuations and building closures (Rubinkam & Smyth, 2024). It is comments like these that further stigmatize immigrant groups, exacerbate the myth-based opinions of the public, and ultimately have real-world, criminal consequences for these populations (Mills et al., 2023).
In conclusion, as seen with Springfield, Ohio, misperceptions, negative political rhetoric, and punitive attitudes regarding immigrant populations are leading to further discrimination and criminal behavior. The current study reveals that factual information has the potential to be an effective tool in changing attitudes about the immigration-crime relationship. Thus, providing more accessible, engaging, and factual information to the public on politicized issues like immigration and crime might begin to have profound impacts on public policy and potentially crime reduction.
Footnotes
Appendix A
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
