Coverage and Framing of Immigration Policy in US Newspapers

Abstract

Policy Points.

• Although immigration policy is recognized as a social determinant of health, less is known about how mechanisms, such as news coverage of policy, influence intermediary and proximal health processes like seeking health care.

• The extent of news coverage of federal, state, and local exclusionary or integration immigration policies can influence public agendas regarding immigrant inclusion and exclusion.

• Exclusionary federal immigrant policies have dominated the news across the United States over the past ten years, despite active immigrant integration policymaking at national, state, and local levels.

Context

Immigration policymaking at federal, state, and local levels in the United States has proliferated in the past decade. While evidence demonstrates that immigration policy is a determinant of health, there has been limited examination of the mechanisms by which policy influences proximal health processes. News coverage has served as a central platform for debates over restrictive and inclusive immigration policies and may constitute an important health mechanism by shaping public agendas, influencing support for immigrant exclusion or inclusion, and framing policy issues, thereby influencing immigrants’ social climates. This study sought to examine the extent of news coverage of exclusionary and inclusive immigration policy at federal and state levels and variations in messages about immigrants during two periods of extensive policymaking.

Methods

We conducted a quantitative content analysis of newspapers’ coverage of immigration policy between 2010 and 2013 and between 2017 and 2019. We conducted a systematic NewsBank search of articles covering legislation, lawsuits, and other policies related to immigration (n = 931). Articles were coded for policy type and level, positive or negative framing of immigrants, and other characteristics. Our analysis then compared the patterns of the two periods.

Results

In both periods, the majority of coverage focused on exclusionary policies at the federal level, despite a significant increase in integration policies between 2017 and 2019. We found significant shifts in both the negative and positive framing of immigrants, from the dominant negative messages of immigrants as an economic drain to immigrants as criminals and the dominant positive messages of immigrants’ economic contributions to immigrants as families.

Conclusions

Since 2010, coverage of exclusionary federal policy has consistently dominated the news, as messages have increasingly described immigrants as either criminals or part of families. We discuss the health implications and future research directions of news coverages’ role in influencing the immigration policy and social contexts that have been linked to health outcomes.

Immigration policymaking at the federal, state, and local levels has proliferated in the past decade. ¹ At all levels, it has focused on surveillance, policing, and deportation policies shaping the extent to which immigrants experience enforcement or live in “sanctuary” jurisdictions. Federal policies have also imposed more restrictions on new migrants. ² At the same time, high‐profile state and local policies that expand immigrants’ rights, such as workplace protections, and eligibility for public benefits, such as Medicaid, have been adopted. ³ , ⁴ Currently, the legislation, regulations, and court cases that constitute immigration policy in the United States form a patchwork of distinct exclusion and integration policies across states and localities. ³

Immigration policy is a determinant of health. ⁵ , ⁶ Exclusionary policies generally are linked to worse physical and mental health and to specific outcomes like poor birth outcomes. ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ In contrast, integration policies are linked to better access to health insurance and better socioeconomic outcomes, such as lower levels of poverty and access to higher education. ¹² , ¹³ , ¹⁴ Immigrants’ health in the United States is dually shaped by these policies of exclusion and integration that regulate the extent to which current immigrants experience restriction through surveillance, policing, or deportation and the extent to which they are granted rights and access to resources, such as health care, workplace protections, and education.

Like other social determinants of health, immigration policy is understood to shape health through structural mechanisms—such as social contexts and attitudes—that influence proximal processes like stress, health behaviors, and care seeking. ⁵ , ¹⁵ What is not understood as well are the specific intermediary mechanisms by which the nation's patchwork of immigration policy contexts may influence these health processes and, ultimately, outcomes. ³

Recent studies of the health impact of immigration policy point to the communication of policy as a potential health pathway. For example, a study that examined birth outcomes associated with the anti‐immigrant climate produced by Arizona's SB 1070 found that the period of policy enactment, not its brief implementation, was associated with worse birth outcomes for pregnant foreign‐born Latinas. ¹⁰ When examining the media environment during these periods, the authors found more extensive coverage of the policy during its enactment than its implementation, potentially contributing to the heightened awareness and stress among the state's immigrant communities.

Consistent with earlier studies of the news coverage of policies, we argue that news coverage—a critical determinant of health ¹⁶ —of immigration policies may be a mediating factor between the policies themselves and health. A large body of communication research has documented the influence of news coverage on population health outcomes through both direct and indirect mechanisms. ¹⁷ For example, content analyses of the media coverage of specific health policies (e.g., the Affordable Care Act [ACA]) and taxes on sugar‐sweetened beverages), theorize that the focus on and messaging of news coverage may directly inform the public about aspects of policy implementation, such as the process for enrolling in the ACA insurance exchanges. ¹⁸ , ¹⁹ The potential for this direct influence, however, may also vary as there is evidence of a marked absence of such health policy information in newspapers reaching underserved populations. ²⁰

How policies are framed may more indirectly influence support for policy responses to different types of social inequities. For example, earlier evidence shows that the media are a key source of messages regarding policies that address the social determinants of health. ²¹ Yet multiple content analyses of news coverage have found that behavioral, and not policy or structural, solutions to disparities were dominant. ²⁰ , ²² This focus on individual‐level behaviors serves to frame policy solutions at similarly individual levels: educational campaigns, for example, rather than regulations against predatory marketing. This evidence suggests that news coverage of policy may influence its ultimate impact and points to the importance of understanding all aspects of news coverage of a policy.

News coverage has served as a platform from which the debates and tensions over exclusion and integration policies have played out, from the coverage of the Deferred Action for Childhood Arrivals (DACA) program in 2012 to the first Muslim ban in 2017 and from the coverage of Arizona's SB 1070 in 2010 to California's state sanctuary policy in 2017. To date, however, news coverage of immigration policy has received little attention in the research on immigration policy and health. Instead, most examinations of news and media coverage of immigration issues have focused on how immigrants themselves are covered, rather than on the policies. ²³

News coverage of the enactment and implementation of exclusionary and integration policies may shape the public agenda and frame policy issues, ²⁴ thereby ultimately serving as a mechanism by which policy influences the proximal health processes that shape immigrants’ health outcomes. By focusing attention—devoting extensive coverage to specific issues at the expense of other issues—the news shapes public perceptions about the relative importance of different topics. News coverage also determines how specific issues are framed—the ways in which specific features of an issue are presented ²⁵ —which influences policy support and health behaviors. ¹⁸

The extent to which exclusionary, compared to integration, policies receive coverage is an indicator of how news coverage may influence public agendas. Evidence of the news media's agenda‐setting role shows that the level of attention given policy issues influences public opinion and legislative action. ²⁶ , ²⁷ For example, one study found that newspapers in border, compared to non‐border, states covered more immigration topics, and in turn, border residents were more likely to report immigration as the “most important problem” in opinion polling on state issues. ²⁸ Another study found that the volume of coverage of Mexican migration in local Canadian newspapers produced political urgency and legitimacy for restricting migration, which compelled lawmakers to change visa policies. ²⁹ Moreover, the potential impact of news coverage on policy is not trivial, as there also is evidence that immigration policies in one state influence the enactment of “copycat” legislation in other states. ³⁰

The valence of coverage—the positive or negative “direction” of coverage—can also shape public perceptions. ³¹ Evidence consistently shows that immigrants are portrayed in a negative light and that framing immigrants as criminals is common. ³¹ , ³² Negative framing, whether it is messages of immigrants as economic burdens or as criminals, is associated with anti‐immigrant sentiment. ³³ , ³⁴ , ³⁵ But the framing of immigration coverage is not uniformly negative. News coverage that includes positive messages about specific populations (e.g., DACA recipients, asylum seekers) can influence perceptions of some immigrants as “desirable” or sympathetic. ³⁴ , ³⁶ There is also evidence that some positive frames can be influential in garnering support for more inclusive immigration policies. ³⁷

A descriptive analysis of the patterns of immigration policy coverage can improve our understanding of the policy contexts in which immigrants live and help identify ways in which policy may influence immigrant health. To this end, we conducted a content analysis of news coverage of immigration policy during two periods of extensive policy movement over the last decade, between 2010 and 2013, the middle of the Obama administration, and between 2017 and 2019, the beginning of the Trump administration. We chose these periods because they are representative of recent patterns in the different types and levels of immigration policymaking. During these times, immigration policies were enacted and implemented at the federal, state, and local levels.

The immigration policymaking during these periods also represents two spectra of immigrant exclusion and integration. ³ Exclusionary immigration policies include both policies that regulate entrance into the United States and those that determine the extent of policing, surveillance, and possible deportation of those noncitizens already living in this country, either by expansion through the authorization of immigration enforcement or the protection of immigrants through “sanctuary” type policies. During the Obama era, federal policies focused on enforcement through collaboration with state and local law enforcement. In reaction to federal policies, some states enacted highly restrictive policies, such as Arizona's SB 1070. Other states, however, sought to expand benefits and opportunities for noncitizens, such as increased worker protections in California. During the Trump era, a period we defined as the start of his presidency until the date of our data collection for this study, federal policy expanded both the enforcement and the restriction of immigrant admissions, like the “Muslim ban.” At the state and local levels were high‐profile state level efforts to enact “sanctuary” type policies that protected some immigrants from enforcement. Within these policy contexts, we hypothesized that despite the proliferation of policies in both areas, there would be more coverage of exclusionary than integration policies. We also expected to find a higher level of negative, compared to positive, framing of immigrants.

Methods

We conducted a quantitative content analysis of newspapers in a selection of US states in the years 2010 through 2013 and 2017 through 2019 to examine the extent of coverage of exclusionary and integration policies and the positive and negative framing of immigrants in this coverage.

Search Criteria

We defined coverage of immigration policy as an article that mentioned legislation, regulation, lawsuits, ordinances, or regulation of migration or immigrants’ rights in the United States. Our search terms were “immigr*” and “policy OR policies,” and we set a parameter that the search terms be within a distance of 20 words from each other. ³⁸ We arrived at this parameter through a series of tests for which we set a distance of 5, 10, 20, and 50 words between search terms, and then we took 10 articles from a state not included in the study. We reviewed each article to determine whether the article covered immigration policy. We found that the smaller distances were too restrictive and missed immigration policy, but that the greater distance identified articles that discussed immigrants and policies that were not always explicitly focused on immigration policy.

Sampling

To create a representative sample of news coverage during the periods 2010‐2013 and 2017‐2019, we searched a purposive sample of state newspapers in NewsBank, a database of newspapers across the United States. ³⁹ Because immigration policy is enacted and implemented across multiple levels of government, state newspapers serve a critical function by reporting on subnational issues, including local and state policies. ²⁰ , ²²

We first identified and then included those states in which 5% or more of the population was foreign born (n = 30). We then identified the newspapers of each state's capital, assuming that these would be the most likely to cover federal as well as state and local policies. We applied our search term for the selected newspapers and further excluded the states with more than 100 articles published in each time period (n = 15). For the final sample of states, we generated a census of 12,575 articles. The number of articles in this census was large because of one outlier state, New York, whose state capital newspaper contained 2,277 and 5,902 articles during the time periods one and two, respectively. We aimed to have a sample of approximately 1,000 articles for power for analyses and logistical feasibility. To produce this sample, we randomly selected one‐half of the sample from articles published between 2010 and 2013 and one‐half between 2017 and 2019, stratified by state, for a final sample of 931. Online Appendix A lists the final sample of states, the census of articles for each, and the final number of articles selected for the study analysis.

Codebook and Coding Procedure

The codebook captures the characteristics of coverage, including section, policy level discussed (e.g., federal, state, or local), enactment of the policy (e.g., legislation, executive order, lawsuit), demographic focus of the immigrants described (e.g., race/ethnicity, country of origin, citizenship), type of policy (e.g., exclusionary or integration), and positive and negative messages. Online Appendix B provides the codes and definitions.

A set of codes captured exclusionary policies, from which we constructed three composite variables that categorized articles as discussing policies authorizing enforcement, protecting individuals from enforcement through “sanctuary” type policies, or restricting immigration. Each of the policies authorizing or protecting individuals from enforcement function within the same policy domain, but because of their distinct function and potential impact on immigrant communities, we placed each in a separate category. We classified articles coded as covering policies related to federal enforcement, E‐Verify, immigration raids or arrests, multilevel government collaboration, and legal status checks as e nforcement policies; articles coded for protective enforcement policies, including detainer noncompliance and “sanctuary cities,” as s anctuary policies (this code was mutually exclusive of Enforcement); and articles coded for having other exclusionary policies, such as the Muslim ban, as o ther restrictive policies. These three policy types represent the variation within exclusionary policies related to the restriction of migration and immigrants’ rights. Some focused on enforcement, others on limiting enforcement, and still others on restricting arrivals to the United States. Another set of codes captured specific integration policies, including health care, social services, public charge, and other inclusive immigrant policies, and together they formed a single compositive variable, i ntegration policies.

To capture the framing of each article, we used an existing instrument with positive and negative messages about immigrants. ²⁴ Positively valenced messages referred to immigrants as contributing to collective resources, being hard workers, having strong families, experiencing racism and discrimination, and being parents of children. Negatively valenced articles referred to immigrants draining collective resources, not working and living off welfare, stealing jobs from Americans, being overly fertile, being terrorists, committing crimes, and being religious minorities. We calculated an overall positive valence variable as the sum of all positive messages in each article, and a negative valence variable as the sum of all negative messages in each article.

The research team developed, tested, and refined the codebook. Because all articles mentioned immigrants or immigration, the goal of this process was to ensure that both the codes and the coding procedure captured the policy aspects of the news coverage. A team of five research assistants tested each iteration of the codebook by coding articles from the excluded states and then discussing and clarifying each code. They repeated this process three times to establish rigorous reliability between coders. ⁴⁰ Before coding, each member of the team coded a sample of 14 articles from the newspaper of an excluded state, after which we calculated the statistics of interrater reliability for each coder pair. The item‐specific raw agreement ranged from 62% to 100%. We computed Cohen's kappa, a measure of interrater reliability that adjusts for chance agreement between raters. ⁴¹ The κ statistics ranged from 0.75 to 0.88 with an average of 0.83, indicating a high level of reliability. ⁴² Each coder was then assigned three states to code.

Analysis

We cleaned and organized the data in Microsoft Excel and analyzed them in SAS 9.4. ⁴³ To assess the overall patterns in the characteristics of the policies covered, the extent of coverage of exclusionary and integration policies, and the valence of coverage, we first conducted a univariate analysis for the full sample. Then, to assess how these patterns varied between the two time periods, we conducted a bivariate analysis with Chi‐squared tests for each variable. Finally, to determine the variation in positive and negative messaging across coverage of policy type, for each subset of articles that discussed enforcement, sanctuary, other restrictive, and integration, we analyzed the proportion of articles that included each of the positive and negative messages.

Results

The overall patterns in coverage show that immigration policies at the federal level and exclusionary policies received the most coverage. Overall, there were more news articles than opinion and editorial pieces (Table 1). Half the articles covered federal immigration policies; 18% covered only state policies; and 5% covered only local policies. Exclusionary policies received the most coverage, with 33% of articles covering enforcement policy, 21% covering sanctuary policy, and 11% covering restrictive policy (Table 2). Contrary to what we expected, there was a higher positive valence across articles, with a mean of 0.9 positive messages per article, compared to 0.4 negative messages. There was, however, extensive negative coverage across multiple messages, such as immigrants as criminals (19.3% of articles) who drain collective resources (9.7%), take jobs from Americans (4.7%), and are terror threats (4.1%). The positive portrayals of immigrants most frequently characterized them as being hardworking (23.7% of articles), being parents of children (18.5%), having strong families (18.5%), and being victims of racism and discrimination (9.9%).

Table 1.

Type of Article and Characteristics of the Policies Covered

	All Articles	Time Period 1: 2010‐2013	Time Period 2: 2017‐2019
	% of Newspaper Articles	% of Newspaper Articles	% of Newspaper Articles
	n = 931	n = 482	n = 449
Type of article
Newsa	69.9	61.4	79.1
Editorial/opinion piecea	30.1	38.6	20.9
Level of policy
Federal onlya	50.5	44.6	56.8
State onlya	18.2	23.4	12.5
Local only	4.7	4.6	4.9
Federal plus other	4.7	4.4	5.1
State and local	0.6	0.6	0.7
General policy (no specific level)	21.3	22.4	20.0
Means of enactment
Legislationa	28.9	35.3	22.1
Executive policya	25.6	16.6	35.2
Lawsuit	5.9	5.4	6.5
Ethnic/cultural background of immigrants covered
Latino	31.8	33.4	30.1
Asian	5.2	5.0	5.4
North African or Middle Easternb	3.7	1.9	5.6
Black	2.7	2.5	2.9
White	1.7	1.2	2.2
Muslima	4.6	0.8	8.7
Region of origin of immigrants covered
Mexico	19.6	19.3	19.8
Central Americaa	9.8	4.4	15.6
Caribbeanb	1.7	0.4	3.1
South America	2.7	2.5	2.9
East and Southeast Asia	3.5	3.1	4.0
Middle Eastb	4.0	2.3	5.8
Europeb	2.7	1.2	4.2
Legal status of immigrants covered
Undocumenteda	70.0	78.4	61.0
DACAa	11.3	7.3	15.6
Refugeea	7.2	2.1	12.7
Asyluma	6.9	1.2	12.9
TPSb	1.2	0.2	2.2

Investigator‐generated data set of newspaper coverage of immigrant policy.

^a p‐value ≤ 0.001

^b p‐value ≤ 0.01

Table 2.

Types of Policies and Valence of Immigration Policy Coverage

	All Articles	Time Period 1: 2010‐2013	Time Period 2: 2017‐2019
	% of Newspaper Articles	% of Newspaper Articles	% of Newspaper Articles
	n = 931	n = 482	n = 449
Exclusionary policies
Enforcement policy	33.0	34.7	31.2
Sancturary policy	21.0	19.1	22.9
Other restrictive policyc	11.0	12.9	8.9
Integration policy a	21.1	16.4	26.1
Overall valence of coverage
Mean # of negative mentions	0.4	0.4	0.4
Mean # of positive mentions	0.9	0.8	1.0
Negative coverage of immigrants
Take jobs from citizensc	4.7	6.0	3.3
Drain collective resourcesa	9.7	13.3	5.8
Do not work	0.5	0.8	0.2
Overly fertile	0.8	0.8	0.7
Commit crimesa	19.3	16.0	22.9
Terror threats	4.1	3.1	5.1
Religious minoritiesb	1.3	0.4	2.2
Positive coverage of immigrants
Contribute to collective resources	15.8	16.6	14.9
Hard workingb	23.7	27.2	20.0
Parents of childrena	18.5	14.1	23.2
Strong familiesa	18.5	14.5	22.7
Experience race racism/discrimination	9.9	7.5	12.5
Learning English	1.9	1.5	2.5

Investigator‐generated data set of newspaper coverage of immigrant policy.

^a p‐value ≤ 0.001

^b p‐value ≤ 0.01

^c p‐value ≤ 0.05

We found notable differences in the characteristics of coverage between the two time periods (Tables 1 and 2). Next, we describe how the coverage characteristics varied between 2010 through 2013 and 2017 through 2019.

Changes in Characteristics of News Coverage

Not only was there more coverage of federal policy, as opposed to state and local policies, but the proportion of articles covering federal policy also increased between 2010/2013 and 2017/2019 (44.6% vs 56.8%, p < 0.01). The proportion of articles covering legislation was higher (35.3% vs 22.1%, p < 0.01), while the proportion covering executive policy was lower (16.6% vs 35.2%, p < 0.01) in 2010/2013 than in 2017/2019. The highest proportion of immigration news coverage during both time periods featured Latinos and immigrants from Latin America. We found, however, significant increases in coverage from 2010/2013 to 2017/2019 of Muslims (0.8% vs 8.7%, p < 0.01) and Middle Easterners (1.9% vs 5.6%, p < 0.01). Similarly, the highest proportion of articles overall mentioned Mexico as a country of origin, with little change between time periods (19.3% vs 19.8%, p = 0.8), but the coverage of immigrants from Central America was higher in 2017/2019 (4.4% vs 15.6%, p < 0.01). Undocumented immigrants had the highest proportion of coverage overall between time periods (78.4% vs 61.0%, p < 0.01), despite the higher coverage in 2010/2013 than in 2017/2019. The coverage of DACA recipients (7.3% vs 15.6%, p < 0.01), refugees (2.1% vs 12.7%, p < 0.01), and asylum seekers (1.2% vs 12.9%, p < 0.01), however, increased from 2010/2013 to 2017/2019.

Changes in Coverage of Exclusionary and Integration Policies

During the period 2010/2013, exclusionary policies received the greatest coverage, including enforcement (34.7%), sanctuary (19.1%), and other restrictive (12.9%) policies (Table 2). In contrast, during the same time period, integration policies accounted for only about 16% of coverage. Compared to 2010/2013, there was no significant change in the coverage of enforcement policies in 2017/2019 (31.2%, p = 0.3), other restrictive policies (8.9%, p = 0.05), or sanctuary policies (22.9%, p = 0.1). In contrast, in 2017/2019, there was a significant increase in the coverage of integration policies (26.1%, p < 0.01).

Changes in Valence of Coverage

While there was no significant change in either the positive or negative valence of articles between time periods (Table 1), there were notable changes in which messages appeared in each time period. In 2010/2013, negative messages of immigrants draining collective resources, taking jobs, and committing crimes were the most prevalent. Compared to 2010/2013, however, in 2017/2019, the number of messages about taking jobs from citizens (6.0% vs 3.3%, p = 0.05) and draining collective resources (13.3% vs 5.8%, p < 0.01) was lower, and the number of messages about immigrants committing crimes (16% vs 22.9%, p < 0.01) was higher. Among positive messages, overall, messages of immigrants as being hardworking, contributing to collective resources, being parents of children, and having strong families were mentioned the most during both time periods. Compared to 2010/2013, we also found in 2017/2019 increases in articles mentioning immigrants as having strong families (14.5% vs 22.7%, p < 0.01), being parents of children (14.1% vs 23.2%, p < 0.01), and experiencing racism/discrimination (7.5% vs 12.5%, p < 0.01). In contrast, we found a significant decrease in mentions of immigrants as hardworking (27.2% vs 20%, p < 0.01).

Online Appendix C lists these patterns across time periods within articles covering each type of policy. Similar to the overall pattern, in articles covering enforcement, sanctuary, and other restrictive policies, there were decreases in both negative and positive messages of immigrants related to jobs and collective resources and increases in negative messages of immigrants as committing crimes and in positive messages of immigrants as being parents and having strong families. In contrast, among the articles covering integration policies, there was an increase in all positive and negative messages.

Discussion

Within the nation's complex contexts of immigrant exclusion and integration, news coverage plays a critical role in setting the public agenda regarding immigration policy. News coverage also frames the ways in which audiences make sense of policy itself. These processes may help explain how immigration policy influences population health. In our content analysis of newspaper coverage of immigrant policy across a selection of US states in 2010/2013 and 2017/2019, we assessed which immigration policies received the most coverage and also the messages about immigrants in that coverage. Next, we look at the implications of these findings for processes through which immigration policy may influence health.

The majority of coverage focused on federal immigration policies rather than state or local policies, even though our sample was of state newspapers. During both periods, about one in three articles focused on enforcement policies, and about one in five focused on sanctuary policies. In contrast, in 2010/2013, only about one in six articles covered integration policies. This, however, rose to more than one in four during 2017/2019. These findings build on previous literature, indicating that the coverage of immigration generally, tends to focus on the national level and on conflict and disagreement over immigration issues.

For example, a recent study of the 2016 presidential election found that anti‐immigrant messages and discussions of policy received the most coverage, driven by Trump's presidential campaign. ⁴⁴ Similarly, a study of immigration coverage over multiple presidential elections from 1996 to 2006 found that a conflict frame has consistently been the most common, focusing on disagreements over policy reforms between government and immigrant groups. ⁴⁵ By comparing the coverage of federal policy to that of state and local policy, our study provides further evidence of the primary focus of news coverage on national‐level issues. Our findings also suggest that not only are messages and frames focused on the tensions between immigrant exclusion and inclusion but that the content of coverage is dominated by the exclusionary aspects of immigration policymaking. There is evidence, however, of a trend toward more coverage of policies related to areas of immigrant integration such as health care, nutrition assistance, and education.

Across both time periods, we found evidence of both negative and positive messaging regarding immigrants, that most coverage is of Latin American immigrant groups, indicating that messaging likely directly or indirectly relates to these groups. Contrary to our hypothesis and past research, these articles had an overall higher positive valence. This is consistent, however, with recent research on the 2016 election, which showed that despite the focus on President Trump's anti‐immigrant messaging, the majority of articles provided a “balanced” tone. ⁴⁵ Articles covering integration policy saw increases across all positive messages. Indeed, even coverage of enforcement and restrictive policies included some increases in positive messaging, indicating potentially sympathetic portrayals of immigrants in some of this reporting.

More important, our findings show changes in the content of positive and negative messages. We saw significant shifts from negative messages regarding immigrants as an economic drain in 2010/2013 to messages about immigrants as criminals in 2017/2019. Similarly, we saw a shift from positive messages of immigrants’ economic contributions in 2010/2013 to messages of immigrants as having families or being parents of children in 2017/2019. This suggests a decline in the economic framing of immigrants and a shift to what could be described as a values or moral frame, focused—at the positive extremes—on messages about immigrant families and—at the negative extreme—on messages about immigrant criminality. Another study of news coverage of immigration and crime from 1990 to 2013 found that the proportion of articles with the framing of immigrants as prone to crime has steadily risen, ³² suggesting a pattern that has been emerging over the long term.

These patterns point to how media coverage may influence the relationship between immigration policy and health by shaping the perceptions of and societal attitudes toward immigration policy. First, the significant focus on specific levels and types of policies reinforce the likely agenda‐setting influence of immigration policy coverage. The dominant focus of coverage indicated an emphasis on federal exclusionary policy over policies at other levels and integration policy. This focus, however, is not consistent with the actual range of federal, state, and local policies that span both exclusion and integration. In both time periods, integration policies were adopted, such as driver's licenses for undocumented immigrants or the extension of Medicaid eligibility in numerous states. ⁴⁴ For public health practitioners, this finding raises considerations regarding public awareness and, more specifically, immigrants’ awareness of the contexts in which they live.

For example, while research has shown a relationship between both exclusionary policies and the integration of state‐level policy contexts and health outcomes, ⁵ the news coverage of exclusionary policies may exacerbate this relationship through immigrants’ heightened awareness of and concern about the surveillance, policing, or deportation in their state or community, as well as the general public's support for these policies. Similarly, the positive health impact of integration policy may be attenuated by immigrants’ lack of awareness and knowledge of policies, despite their aim of creating inclusionary environments for immigrants.

In addition, these findings may provide insights into the emerging evidence that immigrants’ perceptions of the policy environment are associated with health outcomes. The well‐established “chilling effect” shows that immigrant communities’ awareness of and concern about the negative impact of immigration policies, such as deportation, are associated with their avoidance of public resources. ⁷ Recent studies show that immigrants’ perceptions of anti‐immigrant policy climates or disapproval of immigration enforcement policies are associated with fear and feelings of discrimination. ⁷ , ⁴⁷ , ⁴⁸ , ⁴⁹ Indeed, there is evidence that the extent of exclusionary immigration policies in states is associated with noncitizens’ perceptions that their state does not treat immigration favorably. ⁵⁰ One study found that states’ implementation of the Secure Communities enforcement program was associated with more reports of Latinos’ distrust of health information from the government. ⁵¹ Another study found that the announcement of the “public charge rule” was associated with a small decline in child enrollment in Medicaid. ⁵² There are likely related dynamics between policy, coverage that emphasizes exclusionary policy, and immigrants’ perceptions of policy and their environments—which is linked to health.

Finally, the messaging of policy coverage may be a critical aspect of these processes, as the framing of coverage reflects the social environment and can reinforce social attitudes toward immigrants, which is translated into welcoming or hostile contexts. For example, the shifts in negative and positive messaging point to a divergence from or a greater polarization in immigration framing. The focus on immigrants as criminals in 2017/2019 is consistent with and likely a reflection of President Trump's and others’ increasingly hostile political rhetoric toward immigrants. ⁴⁴ Such rhetoric and Trump's election in 2016 have been linked to poor health outcomes, from the birth outcomes of Latina women to the emotional well‐being of youth. ⁵³ , ⁵⁴

The role of news and media coverage in mediating these relationships merits examination. In addition, the shift of positive messages from economic framing to more values framing likely has implications. There is mixed evidence regarding the potentially beneficial impact of the family framing of immigration messages. In a study testing the messaging of immigration policy, researchers found that the family frame was associated with greater support for immigrant legalization among conservative voters, but no immediate improvement in support from liberal voters. ³⁷ Despite the seemingly positive messages of families, these messages may reflect or reinforce broader social concepts regarding the value of different immigrant groups. This is consistent with the evidence of the ways in which concepts of “deserving” immigrants create or perpetuate inequities within immigrant populations. ⁵⁵ Messages of “deservingness” influence the extent of support for inclusion for some, ⁵⁶ thus potentially justifying the exclusion of others. ³ In the context of increased messaging about immigrants as criminals, the shift to a positive focus on immigrants as families is reminiscent of the false dichotomy between “good” and “bad” immigrants that has emerged in many immigration policies. ³ For example, President Obama's Deferred Action for Parental Accountability was introduced as a policy that would ensure the deportation of only “felons, not families.” ⁵⁷

As more knowledge is gained regarding the relationship between immigration policy and health, further examination of the role of news and media coverage can elucidate the pathways between the nation's complex patchwork of exclusionary and integration policies and the well‐being of immigrant populations. Our study provides a starting point for this inquiry, and its limitations can be addressed in future studies. Specifically, future research can link content analysis data with population health data across states and/or time periods. For example, research can determine whether the level of coverage of federal, compared to state immigration policy, is associated with variations in mental health or stress‐related conditions (e.g., hypertension) across states. Research can also examine the extent to which the changes in positive and negative messages are associated with changes in access to health care at different time periods. Content analysis data can also be linked to opinion polling of public perceptions of immigration policy to assess if changes in the coverage of exclusionary and integration policies is associated with changes in immigrants’ perceptions of the policy environment or the public's perceptions of immigrants. As newspaper readership continues to decline, other information sources, particularly social media, may provide insights into how communication about immigration policy can shape the public agenda and frame policy issues. For example, a study of tweets found that the implementation of Arizona's SB 1070 was associated with greater anti‐immigrant sentiment and messaging on the social media platform. ⁵⁸ Finally, the focus on English‐language news outlets may have missed patterns and messages of coverage in non‐English news outlets, which play a unique agenda‐setting and framing role.

Conclusions

In the United States, 40 million noncitizens live, go to work and school, and build communities under a patchwork of multiple levels and often discordant types of immigration policy. Understanding how the media communicate policy information and shape the support for these policies merits our continuing attention to better understand how policies of exclusion or integration affect immigrants’ health across the nation.

Funding/Support: None.

Conflict of Interest Disclosures: All authors completed the ICMJE Form for Disclosure of Potential Conflicts of Interest. No conflicts were reported.

Supporting information

Appendix A: States included in content analysis, census, and frequency in final sample

Appendix B: Codebook

Appendix C: Positive and negative messages in newspaper coverage of immigration