Abstract
Since 2020, many local, state, and federal governments have passed resolutions or formal statements that officially declared racism as a public health crisis in their communities. These declarations have created opportunities to raise public awareness about the harmful consequences of racism and have provided a platform for advocates and policymakers to propose policies and allocate resources to address the problem. This study contributes to our understanding of these declarations in two ways. First, it examines the political and sociodemographic factors that predict the enactment of a declaration at the media market-level. Second, it analyzes the impact of such declarations on local TV news coverage of racism. We find that markets with a higher percentage of votes for Trump in 2016 were less likely to enact a declaration, as were markets with higher percentages of uninsured and Black populations. Time series models indicate that declarations were associated with more attention to racism on local TV news, particularly in the short term, which is an important political consequence of enacting declarations.
Introduction
Research in public health is clear: across numerous indicators, people of color and Indigenous people have worse health outcomes than do white people in the United States. While public health researchers have been documenting racial disparities in health for decades, scholarship identifying the causal attributions for these disparities has been shifting. Explicit articulation of the fundamental cause of these differences as attributable to systemic or structural racism, in particular, has been relatively recent (Bailey et al. 2017; Phelan and Link 2015) and carries important political implications.
Historically, scholars from various disciplines have argued that racist beliefs and practices can be embedded in institutions and help create biased outcomes that are used to reaffirm the existing racist beliefs (Bonilla-Silva 2006; Fields and Fields 2022; Jackman 1994; Mills 2019). More recently, scholars in public health have undertaken a deeper exploration of this structural form of racism, and how it has shaped patterns of inequality in health outcomes. Bailey et al. (2017) define structural racism as “the totality of ways in which societies foster racial discrimination, via mutually reinforcing inequitable systems, that in turn reinforce discriminatory beliefs, values, and distribution of resources (Bailey et al. 2017, 1455).” Camara Jones (2018) defines racism in such a way so as to include the whole society, both advantaged and disadvantaged groups: [racism] is “a system of structuring opportunity and assigning value based on the social interpretation of race, that unfairly disadvantages some individuals and communities, unfairly advantages other individuals and communities, and saps the strength of the whole society through the waste of human resources (Jones 2018).” Despite these important definitions by leaders in the field of public health, explicit attention to structural racism within public health was quite limited prior to 2020, as documented in a literature review published in 2018 (Hardeman et al. 2018). Similarly, structural racism was also rarely mentioned in reports written for state or national public health governmental bodies. In a rare exception, the 2014 Advancing Health Equity report by the Minnesota Department of Health specifically aimed “to be explicit about race and structural racism, especially the relationship of race to the structural inequities that contribute to health disparities” (2014, 6).
Attention to racism in the field of public health has accelerated in recent years, particularly following the 2020 murder of George Floyd and the COVID-19 pandemic. One study noted a 50-fold increase in citations in the medical literature examining “structural racism” and health between 2017 and 2021 (Dean and Thorpe 2022). The accumulating evidence surrounding structural racism, the real-time recognition of dramatic inequities in COVID-19 transmission, infection, and mortality rates that emerged in 2020, as well as the broad public reckoning related to racism in May and June of 2020 combined to create the political conditions for local and state governmental jurisdictions to pass formal declarations that racism is a public health crisis. These declarations align with broader public health strategies that leverage moral suasion as a tool to prompt action, emphasizing the potential of persuasive communication within a fragmented governance system like the U.S. (Greer and Singer 2017). While few such declarations existed prior to 2020 (Mendez et al. 2021), they increased rapidly in 2020 and into 2021. The CDC declared racism “a serious public health threat” in April 2021 (Wamsley 2021).
When local and state jurisdictions recognize racism as a public health crisis, they are formally acknowledging the detrimental impact it has on the health and well-being of people of color. These declarations have been issued by a wide range of public entities, including city and town councils, county boards, governors, state legislatures, education boards, health associations, and public health departments, highlighting a collective effort to address systemic inequities and their public health implications. Mendez et al. (2021) note that most of these declarations were initiated by local municipal governments, with systemic racism, the COVID-19 pandemic, social determinants of health, and specific health outcomes emerging as common themes.
A key feature of several declarations is the adoption of an “equity in all policies” approach, which emphasizes evaluating existing policies and designing future programs through a racial equity lens (APHA, 2021). For instance, on June 12, 2020, then-Mayor Martin Walsh of Boston, Massachusetts, formally recognized racism as a public health crisis through an executive order. This order outlined eight strategic priorities to address the impact of racism on residents' health and well-being. To support these efforts, Mayor Walsh reallocated roughly $3 million from the Boston Police Department's overtime budget in the budget resubmission, directing these funds to the Boston Public Health Commission to combat racial health disparities (APHA, 2021). However, many declarations have been largely symbolic—that is, not necessarily accompanied by a formal allocation of resources (Mendez et al. 2021). Thus, questions remain about the meaning of and consequences of these local and state government declarations of racism.
Such declarations may serve an important agenda-setting function in local politics: they can communicate that a community takes racism seriously, and consequently shape the political dynamics in that jurisdiction to more openly discuss racism (such as in local news media), create and sustain new coalitions or partnerships organized around the crisis, and contribute to the passage of more policies to dismantle racist policies and conditions in those communities. However, whether or not these declarations actually increase attention to racism and its pernicious impacts is an empirical question, one that we begin to address in this study.
Background
Local governments and local media interact to draw public attention to some issues over other issues—a social and political process known as agenda-setting. Both local governments and the media are public arenas which have finite attention they can devote to particular issues, or social problems, in society (Hilgartner and Bosk 1988). Under this model, racism is an issue that gets constructed as a social problem (through assignment of causal attribution, see, e.g. Stone (1989)) by political actors in the executive or legislative branches of local government as well as in the news media (Hilgartner and Bosk 1988). The news media, in turn, through their sustained attention to particular issues, can either keep those issues on the policy or public agendas, or given the finite capacity of the media in sustaining attention, allow it to drop. When news media do expend attention on a particular issue, that attention can shape both public and policymakers' attitudes about what of many issues are the important ones to address (McCombs and Shaw 1972).
Passing a local policy (like a declaration of racism as a public health crisis) can also have consequences on the political trajectory of that community. Policy feedback theory describes the conditions through which an existing policy can shape the attitudes of the public, which can have “consequences for subsequent … policymaking” (Campbell 2012, 334). The media is one institution which can mediate this pathway through which policies can change local politics. Thus, through both agenda-setting as well as policy feedback theories, the local media are an important institution for determining the political prospects that a declaration of racism as a public health crisis might have meaningful consequences.
The introduction of declarations of racism by local governments began in May 2019, when Milwaukee County in Wisconsin adopted such a resolution, followed shortly by Cook County Illinois and Milwaukee City, respectively, in June and July 2019 (Mendez et al. 2021). Numerous cities and counties followed their lead, especially in light of the broader national discourse on racism following the murders of Breonna Taylor and George Floyd by police. At the time of this writing (December 2024), the American Public Health Association had identified 268 declarations across all jurisdiction levels (24 state, 95 county, and 149 city).
There have been only a handful of previous studies examining the local political agenda or policy implications of these declarations. One study, which reviewed the declarations passed in the earliest phase (through September 2020; N = 125), analyzed the geographic distribution and content of such declarations (Mendez et al. 2021). They found that declarations offered only limited information about specific action steps or resources to actually tackle racism. Another study, offering a critical view on the promise of declarations, pointed out that merely naming racism is insufficient for actually addressing and dismantling structural racism (Paine et al. 2021). These authors conducted a case study of three communities which passed racism resolutions, assessing the extent to which declarations went beyond performative communication into actually addressing processes, community engagement, and offering steps toward solutions. Another critical perspective analyzing the declarations identified places where they fell short, including in their lack of attention to the role of power distribution and limited work to center the perspectives of people of color (Pierson-Brown 2022). Another qualitative case study, examining the declarations in the Maryland-Virginia-Washington DC area, found that public health officials generally agreed that explicitly naming racism as a root cause of racial health inequities is a significant step forward and that the declarations can contribute to changes within public health practice, such as hiring health equity focused positions and organizational equity, diversity, and inclusion efforts. However, they offered less optimistic views on how these declarations might reach the broader community and cross into other sectors, and the qualitative interviews the authors conducted were silent on whether these efforts might shape media attention (Lamberti 2022).
Few studies have examined news media response to the declarations. Researchers with the Berkeley Media Studies Group conducted an analysis in print media of racism declarations and found that while media attention to declarations was fairly high (with 800 articles identified in their study), coverage declined over time, and relatively little coverage discussed community members and community organizations and the role they can play (Barna 2023; Mahmood et al. 2025). While not a study specifically on the declarations per se, another analysis of the official Twitter accounts of state health departments from October 2020 to February 2021 (during a time period of vaccine roll-out, but also the proliferation of these declarations) found that only 0.06 percent of all tweets referenced racism, suggesting an exceptionally low and inconsistent use of these terms in external communication by health departments (Bradford et al. 2022). Another study from a slightly earlier period (March to June 2020) found that only 9 percent of local TV news stories reporting on health disparities referenced racism specifically (Xu et al. 2023). The full extent to which local media, in contrast to official public health communication, reported on racism during the time period of most of these declarations is yet unclear, which is one of the motivations for the present study.
Research Gaps and Specific Motivation
Like others in this special issue (Jarman et al. 2024; Kuo and Kelly 2024; Sullivan and Strach 2024), this study examines local politics, as opposed to examining national-level public health politics; specifically, we examine the local politics of racism declarations. Local declarations of racism as a public health crisis have important political and communicative implications, and, as noted above, a few studies have begun to examine them. Yet there has been limited research examining the factors that contributed to cities, counties, and states passing such declarations—which areas were more likely to identify racism as a public health crisis specifically on the public agenda, and which were not? While the studies noted above have examined the content of declarations and/or examined the stakeholders involved in the political process, only one study, to our knowledge, has examined predictors of their passage. Farris, Holman and Sullivan (2022) examined racism declarations by cities, specifically (and not the other levels of government), to assess the characteristics of local policymaking that were associated with the adoption of racism declarations. They found that cities with Black mayors, and those cities with a higher percentage of Black residents, were more likely to have adopted a racism declaration (Farris, Holman and Sullivan 2022). They focused on the characteristics of the elected officials, and no other components of the population (besides its racial make-up) that might explain the geographic variation in declarations.
There has also been limited research examining the consequences of such declarations. Existing research looking at what has happened in communities following these declarations have largely been case studies, which can offer important insights into the local politics of response to racism but are limited in their contribution to our understanding of the trends across many local areas. One criticism of declarations to date is that too few offer meaningful commitment in terms of policy proposals or resources devoted to addressing racism (Mendez et al. 2021; Paine et al. 2021). One potentially significant outcome of a declaration—even a declaration that does not have these otherwise important components—is to draw more news media attention to racism. Such attention, especially if sustained over time, could keep these issues on the public agenda, continuing to inform the public that racism is an issue in their community that demands policy action. Of course, news media attention to racism is not necessarily in agreement with the premise underlying these declarations—namely that racism is prevalent and harmful to community health. It is quite possible that some coverage was critical of these initiatives, or questioned the link between racism and health outcomes. That said, the volume of attention in news (regardless of content) contributes to public evaluation of how salient or important that issue is (McCombs and Shaw 1972). Thus, understanding whether a relationship exists between declarations of racism and news media attention to racism offers important insight into another political consequence of these symbolic pronouncements, regardless of how these efforts were portrayed.
In this study, we focus on local TV news attention to racism, and whether such attention was responsive to declarations (in the short and long term). We focus on local TV news because, despite huge changes in the news media ecosystem over the last 10 years, local TV news remains one of the most widely used news sources (Lopes et al. 2023; Pew Research Center 2024). It is an important source of health information, popular across partisan lines (Gollust et al. 2023; Gollust, Fowler and Niederdeppe 2019), and is particularly trusted for health information (Lopes et al. 2023). Further, for an assessment of the local politics of policy actions, local media are essential. Local news media, and TV in particular, have also been facing important threats to their ability to cover local issues and maintain their local agenda-setting function. Specifically, there has been a great deal of consolidation across stations, leading to more nationalized and often, more conservative-leaning, content, such as the Sinclair Broadcast Group's continued acquisitions of stations (Levendusky 2022; Martin and McCrain 2019; Neumann et al. 2024). These dynamics could constrain the potential for declarations to contribute to local attention to racism, a possibility we address in this study by attending to station ownership in our analyses.
Research Questions
Based on the above motivation, in this study, we examine three specific research questions:
What is the geographic distribution of declarations of racism as a public health crisis between 2020 and 2022? What local (media market-level) characteristics predict which communities made racism declarations in those communities? Were racism as a public health crisis declarations associated with attention to racism on local TV news?
To address these questions, we estimate multivariate regression models to examine (i) the predictors of the passage of declarations of racism as a public health crisis and (ii) the impact of these declarations on local TV news attention to racism. The explanatory variables included in the analyses reflect key factors linked to structural racism and its intersection with public health inequities and local sociopolitical contexts. For instance, to predict which communities made racism declarations, we incorporate Trump 2016 vote share as a measure of partisan lean, reflecting differing partisan perspectives on race and racism. We expect that conservative-leaning areas, characterized by higher Trump support, may be less likely to adopt declarations addressing structural racism (Gollust, Fowler and Niederdeppe 2019; Levendusky 2022).
Additionally, we include key sociodemographic variables such as the percentage of Black residents and uninsured individuals, as these indicators reflect community vulnerability to systemic inequities and health disparities, which are theorized to drive policy attention. Specifically, we include the percentage of Black residents as an explanatory variable, guided by Mendez et al. (2021), who highlight that many declarations explicitly reference the historical and contemporary oppression of Black communities. While some declarations use broader language to address racism's impact on multiple groups (e.g. Black, Indigenous, and Latinx populations), they disproportionately emphasize Black communities (Mendez et al. 2021, 9). Similarly, Rakotovao et al. (2024) underscore that systemic anti-Black racism is a central focus of public health declarations and initiatives across North America, often prioritizing the disproportionate effects of racism on Black communities. Infant mortality rate serves as a proxy for structural racism's impact on health, with prior studies linking higher rates to greater structural inequities (Wallace et al. 2017). Control variables, including population density, percentage of elderly, percentage of homeowners, percentage of rural residents, and whether there is a state capital in the market, are added to account for factors influencing both structural racism and local media narratives. These variables ensure that the analysis robustly isolates the relationships between structural inequities, public health crises, and media attention.
Data and Methods
To address our research questions, we gathered data on declarations of racism as a public health crisis from a comprehensive dataset maintained by the American Public Health Association (https://www.apha.org/). 1 Our dataset covers the period from January 2020 to March 2023 and includes declarations made by cities, counties, and states in the United States. Throughout this period, 250 declarations were enacted, with 20 by states, 89 by counties, and 141 by cities. 2 Most declarations were made by city or town councils, totaling 118. This was followed by health or public health entities, which issued 52 declarations, and county boards, which passed 49 declarations. Additionally, 16 declarations were part of statements made by governors or mayors, while eight were approved by state legislatures. Education entities made five declarations, and two were issued by entities classified as “other.” Nineteen states enacted at least one declaration at the state level, with California leading the way with two. Twenty-five states made at least one declaration at the county level and 33 at the city level. Nine states, including Idaho, Louisiana, Montana, New Hampshire, New Mexico, North Dakota, South Carolina, South Dakota, and Wyoming, did not enact any declarations during this period.
We also compiled an extensive collection of daily newscasts from all media markets and local TV stations processed by the Wesleyan Media Project from the company TVEyes (https://tveyes.com/). We searched through the closed captioning transcriptions of these newscasts to identify any mentions of “racism.” We analyze the volume of media attention to racism as an outcome because prior research on media agenda-setting suggests that increased news coverage of an issue elevates its perceived importance among the public (Iyengar and Kinder 1989; McCombs and Shaw 1972). When the media devotes greater attention to a topic, it enhances the salience of that issue in public discourse, shaping how people prioritize and understand societal challenges. By examining the volume of coverage, we aim to capture the extent to which declarations of racism as a public health crisis contribute to making racism a more salient issue in the media and, consequently, in public perception.
We then combined this dataset with demographic information from the 2018 American Community Survey available from IPUMS for each media market and election results from David Leip's Atlas of U.S. Presidential Elections for 2016 to capture the percentage of the two-party vote that Donald Trump received in each market. Finally, we created two binary indicator variables: one to indicate if a market is a state capital and another to capture whether a news station is owned by Sinclair, which has a reputation for news coverage (Gollust, Fowler and Niederdeppe 2019; Levendusky 2022; Martin and McCrain 2019; Neumann et al. 2024). The media market refers to a geographic area where all residents have access to the same local television programming. Media markets are designed as exclusive and exhaustive geographic units, meaning they do not overlap, and every county is assigned to a single media market (i.e. designated market area or DMA). To investigate the association between declarations and the coverage of racism on local television, we conducted our analysis at the media market-level, as this is the level at which the news broadcasts are disseminated, thus providing a relevant and consistent unit of analysis for our study. We analyzed all declarations within each media market and tested their relationship to the volume of news coverage in that same market. This approach provides a coherent framework for linking public health declarations to media attention within clearly defined, non-overlapping geographic areas.
To identify which local characteristics are associated with the enactment of declarations, we estimate binary logistic regressions of declarations on a set of market-level characteristics. Our dependent variable is whether a declaration has been enacted in a particular media market, and we estimate a separate model for each type of declaration. We included several explanatory variables, including Trump 2016 vote share 3 which served as a measure of the market's partisan lean given differing partisan perspectives on race and racism. 4 More specifically, we expect that Trump vote share will be inversely related to the number of declarations in the media market.
We also included a few key sociodemographic and health-related factors as important independent variables related to structural racism and public health in a given community. These included the percentage of Black residents and the percentage of uninsured individuals in the market. We expect both the percentage of Black residents and the percentage of uninsured to be positively associated with declarations. Additionally, we incorporated infant mortality rate as previous work has found a relationship between more structural racism in a geographic area with higher infant mortality rates (Wallace et al. 2017). We expect infant mortality to be positively correlated with declarations, if this evidence of poorer community health prompts more policy attention.
We add additional demographic characteristics as control variables, including population density, percentage of elderly, percentage of homeowners (an indicator of community wealth), percentage of rural residents, whether the market is a state capital, and regional indicators. These important indicators are linked to the degree to which discussions of racism emerge in news, and the broader political context. As such it's important to rule out the possibility that these factors could explain the connection between the declarations of racism as a public health issue and the broader discussion of this topic in local media.
We next conducted an analysis to investigate the relationship between declarations of racism as a public health crisis and the coverage of racism on local TV news. To understand how this relationship has changed over time, we estimated autoregressive distributed lag and partial adjustment models. These models helped us determine both the short-term and long-term effects of changes in declarations on the media attention of racism. For this analysis, our main explanatory variable is whether a declaration was enacted in a specific market during a particular month. However, we also include in our models a variety of controls that we believe may be related to news attention to racism.
First, it is well-established that news attention is responsive to real or perceived differences in audience demand, which can vary with the demographics of a market (Gans 1979; Hamilton 2004; McManus 1995). Political parties have very different perspectives on issues of race, which was further exacerbated during the Trump presidency that frequently featured rhetoric that activated racism and sexism (Cassese and Barnes 2019; Schaffner 2018; Schaffner, MacWilliams and Nteta 2018; Sides, Tesler and Vavreck 2018). As such, we include a control for the 2016 share of Trump in our news attention modeling, and we expect that as the percentage of citizens who voted for Trump rises in a media market that attention to racism in news will decrease.
As before in the models predicting declarations, we include measures of the percentage of Black residents and expect that attention to racism in news would increase as a function of larger communities of color. Media markets containing a state capitol have been shown to have residents who are more attentive to issues of policy (Carpini and Keeter 1996), and so we might also expect to find a positive relationship between having a state capitol and attention to racism. Similar to the previous model, we also control for population density, which have been shown in to affect the amount of time and attention news devotes to public affairs (Althaus and Trautman 2008; Vinson 2003), percentage of homeowners, percentage of rural residents, and regional indicators to account for historical differences with respect to race, particularly in the South. Finally, we also included a binary indicator to capture May and June 2020, which approximates the time of the protests and uproar that followed George Floyd's murder, so that we do not solely attribute the increased news attention during that time to public health declarations.
Results
In this section, we will first provide a description of the declarations of racism as a public health crisis, focusing on their spatial and temporal variation. Then, we will present the results of our models to predict the enactment of declarations. Lastly, we will show the results of our models to assess the impact of declarations on local TV news attention to racism.
Describing the Declarations of Racism as a Public Health Crisis
Figure 1 shows the number of declarations that have been enacted in different U.S. states over time from the study period of January 1, 2020 to March 2023, including those made at the city, county, and state levels. The majority of declarations were made between 2020 and 2021. In 2020, 203 declarations were put in place, followed by 36 in 2021, 11 in 2022, and none in the first 3 months of 2023. In 2020, none of the declarations were enacted before May. The highest number of declarations was made in June and July of 2020, totaling 151. This aligns with the period of protests following George Floyd's murder and also when the COVID-19 pandemic rose to a second peak in the country. 5 Among the states that made at least one declaration, California (37), Ohio (31), Connecticut (25), Massachusetts (20), and Michigan (20) had the highest number of declarations.
Racism as a public health crisis declarations in the United States, January 2020–March 2023.
Figure 2 depicts the distribution of declarations made at the state, county, and city levels across the United States. Figure 2a shows that the majority of state-level declarations were enacted by states in the West (10), followed by states in the South (4), Midwest (3), and Northeast (3) regions. The spatial distribution of county- and city-level declarations follows a similar pattern, as illustrated by Figure 2b, c. Although some states such as Texas and Florida did not enact any declarations at the state level, they had declarations at the city and county levels. Most of the county-level declarations were enacted by counties in the Midwest (36) and West (31) of the country, whereas Northeast (56) and Midwest (46) regions concentrate most of the declarations at the city level. For a more detailed breakdown of the number of declarations at each level by state, please refer to Table A1 in the appendix.
Declarations of racism as a public health crisis across U.S. States, Counties, and cities.
Predictors of Declarations at the City, County, and State Level
In this section, we present the results from our multivariate regression analyses to examine which market-level characteristics predict the enactment of declarations of racism as a public health crisis.
First, we start with the results of a model to predict the enactment of declarations at the city level. Table 1 displays the results of logistic regression models, where the dependent variable is whether any declaration was enacted in cities within a media market. Column 1 exhibits the results without regional indicators, while column 2 includes them.
Market Predictors of Declarations at the City Level.
Note. State clustered standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.
Our results suggest that markets with higher shares of votes for Trump in the 2016 election are less likely to have city declarations. Larger markets are more likely to have city-level declarations, while markets with older adults are less likely to have one. On the other hand, markets with a higher infant mortality rate are more likely to have city declarations, suggesting possible policy responsiveness to markets with this salient indicator of worse public health. Interestingly, our results also uncover some counter-equity findings, such as the fact that markets with higher proportions of uninsured people and Black populations are less likely to have city-level declarations.
Table 2 presents the results of logistic regression models, where the dependent variable is whether any declaration was enacted in counties within a media market. Similarly to our findings for city-level declarations, markets with higher shares of votes for Trump in the 2016 elections and those with higher proportions of uninsured people are less likely to have a county-level declaration.
Market Predictors of Declarations at the County Level.
Note. State clustered standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.
Table 3 displays the results of logistic regression models that analyze market-level predictors of declarations made at the state, city, or county level (in other words, any declaration that would include that media market). Overall, the results are similar to those found for the city- and county-level declarations. Markets with a higher percentage of votes for Trump in the 2016 elections are less likely to enact a declaration. In contrast, the chance of having a declaration is higher in larger markets. Markets with higher percentages of uninsured people and Black populations are less likely to have declarations adopted at the state, city, or county level.
Market Predictors of any Declaration (State, City, or County).
Note. State clustered standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.
In the next subsection, we will explore the relationship between the enactment of declaration of racism as a public health crisis and media attention to racism on local TV news, testing whether the passage of these declarations is associated with the short- or long-term TV news media agenda.
The Association Between Declarations and Attention to Racism on Local TV News
The coverage of racism by local TV news has varied significantly across different media markets and over time. On average, racism was mentioned 46 times per month in a media market, as displayed in Table A2 (see Appendix for more information). It is worth noting that the majority of the variation in racism coverage occurred within a specific media market over time, rather than between different markets.
To better understand the variation in local TV news coverage of racism, Figure 3 presents the number of mentions over time across different states. There is a substantial heterogeneity across states, with some states having a higher coverage of racism while fewer in others. There was a substantial increase in the volume of attention to racism in June 2020, coinciding with the period of the highest number of enactment of declarations and protests following George Floyd's murder.
Local TV news attention to racism.
We now assess the relationship between the declarations of racism as a public health crisis and the attention given to racism in local TV news. In Column 1 of Table 4, we present the results of a partial adjustment model. The results show that declarations are a strong predictor of local TV news mentions of racism, controlling for political factors and social-demographic characteristics of the media market. All else equal, the enactment of a declaration produces a long-term increase of 294 (or between 247 and 341 using a 95 percent confidence interval) mentions of racism on local TV news. For robustness, we ran additional models including an autoregressive distributed lag model, a model with robust standard errors, with Newey-West standard errors, and with Driscoll-Kraay standard errors (see Table A3 in the Appendix).
Association Between Declarations of Racism and Local TV News Attention to Racism.
Note. State clustered standard errors are in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.
Figure 4 shows the response over time to media attention to racism after a one-unit increase in declarations. A one-unit increase in declarations triggers a corresponding rise in media attention to racism. The period-specific effects are positive and decay substantially over the following months after the pulse. Given the large immediate response, the cumulative effect indicates a substantial corresponding rise in media attention to racism. These results confirm our theoretical expectations that the enactment of a declaration triggers a substantive large positive response in local media attention to racism.
Estimated lag distribution of changes in racism declarations on media attention to racism.
Our findings also show that local television news in areas with a higher percentage of the 2016 two-party vote share going to Donald Trump tended to mention racism less frequently. Additionally, areas with a greater proportion of rural residents were also less likely to mention racism. On the other hand, larger markets and those with a higher percentage of homeowners were more likely to discuss racism in their coverage. Additional models were estimated considering as explanatory variables whether there is a state capitol in the market and the percentage of Sinclair stations within a media market. As the estimated coefficients for these variables were not statistically significant, we chose to report these results in the appendix section (see Table A4).
Discussion
In 2020 and the following years, there was a significant increase in policy attention to racism as a public health issue, but this attention was not spread evenly across the United States. The American Public Health Association documented 251 declarations across various levels of jurisdiction from January 1, 2020, to March 2023. These declarations were made by 20 states, 90 counties, and 141 cities. The majority of state-level declarations came from states in the West (10), followed by states in the South (4), Midwest (3), and Northeast (3) regions. A similar pattern was observed in the distribution of county and city-level declarations. Although some states, such as Texas and Florida, did not enact any declarations at the state level, they had declarations at the city and county levels. Most county-level declarations were made by counties in the Midwest and West of the country. In contrast, the Northeast and Midwest regions had the most declarations at the city level. In this study, we focus on the 50 U.S. states, as our data source does not track declarations of racism as a public health crisis in U.S. territories such as Guam, American Samoa, or Puerto Rico, which are critical contexts for examining the intersection of structural racism and public health and merit further research.
The results from our multivariate regression analysis indicate that markets with a higher percentage of votes for Trump in the 2016 elections were less likely to enact a declaration and less likely to cover racism, which is consistent with theories of news production that suggest that news shifts depending upon the audience (Hamilton 2004; McManus 1995). Larger markets and those with a higher percentage of homeowners were more likely to discuss racism in their coverage, while areas with a greater proportion of rural residents were also less likely to do so. Notably, our results also show that the enactment of a declaration triggers a substantive large positive response in local media attention to racism. Additionally, we find evidence that markets with higher percentages of uninsured people and Black populations were less likely to have declarations adopted at the state, city, or county level. In contrast, larger markets and those with a higher infant mortality rate were more likely to have a declaration.
These findings add to the literature on how local jurisdictions vary in their concentration of public health resources—here, the resource is a policy marshaling attention to racism. Other studies also demonstrate how community-level factors can shape the distribution of resources—from public health education such as public service announcements (Gollust, Fowler and Niederdeppe 2019), to communication of health statistics at the local level (Hansen et al. 2021) to more distinct resources like mosquito control districts (Jarman et al. 2024), all of which vary across local geographies by the partisan or wealth composition of those areas. The strategic concentration of these resources in some areas, and not others, can exacerbate inequality (An, Levy and Hero 2018; Jarman et al. 2024).
Our approach differs from a previous study that examined city-level political factors associated with city declarations and found that a higher percentage of Black residents in a city is associated with higher odds of a declaration, all else equal (Farris, Holman and Sullivan 2022). Our approach at the market-level allows us to understand broader politics in a media environment, as even people living in the suburbs or far reaches of a city are exposed to that media environment. Moreover, our analysis demonstrates that issue attention catalyzed by a declaration enacted by a public official in a single jurisdiction can extend beyond local boundaries, influencing a broader media market and amplifying public health discourse (Greer and Singer 2017).
Our study also finds that declarations were associated with more attention to racism on local TV news, which is a political consequence of declarations that have not yet been explored. While the resources invested or community engagement processes that accompanied the declarations are important, examining the news media consequences is also crucial. We do find that markets that enacted a declaration saw more attention to racism, although we are unable to identify whether this attention was specific to the public health context, nor do we know this attention was positively or negatively valenced, as we describe as a limitation below. Our results also show that the declarations had a relatively steep drop-off in attention, further demonstrating the relatively short attention cycle of the media.
There are some limitations to our research that we would like to address. Firstly, we focused on declarations made between 2020 and March 2023, so any declarations made outside of this time frame were not included in our study. As we previously mentioned, there are only a few of them. Secondly, we were fairly limited in our inclusion of predictors of either declarations or media content; we included local demographic information about media markets that we could access, but we did not have data on other local characteristics that may be relevant, such as the activity of public health or health equity-related non-profit groups, organizers, coalitions, or multi-sector partnerships that might contribute to both the likelihood of a declaration being enacted or of sustained attention to racism on TV (Other work, such as Ku and Kelly in this issue, have demonstrated the importance of such partnerships in public health policymaking.). Thirdly, we did not investigate the diversity of content across the declarations, which has been done in previous studies (Paine et al. 2021). Fourth, the racism keywords we used for the local TV news search are not specific to the context of health, so we cannot confirm if news coverage discussed racism in the context of health or determine the tone of the coverage. We plan to address this issue in future work. Fifth, our study aimed to assess the association between declarations and media coverage of racism. Although our aim was not to establish a causal relationship between variables, we believe our study provides a valuable contribution to future studies that aim to establish a causal link. Sixth, our model does not account for other dynamic changes that happened over the study period of analysis—it considers only static factors as key independent variables (e.g. demographic composition of markets) and only declarations varied across time. This is important because other time-varying characteristics could have also affected either racism declarations or racism coverage, such as COVID-19 rates or other high-profile events, including police killings.
Finally, it is important to acknowledge the nuance that while our study finds an association between declarations of racism as a public health crisis and increased attention to racism on local TV news, we are only measuring the volume of this coverage, not its tone or content. This distinction is critical because “attention to racism” is not inherently positive or negative; the tone and framing of this coverage could vary widely, potentially influencing public perceptions in complex ways. Racism as a keyword could encompass a range of topics, from constructive dialogue to contentious or polarizing narratives, and our analysis does not determine which of these is reflected in the coverage. Given the potential for backlash or misrepresentation in public discourse about racism, it is not our aim to imply that increased attention is necessarily beneficial. Rather, we demonstrate that declarations are associated with more discussion of racism in the markets where they occurred, while recognizing that future research should explore the tone and framing of this coverage to fully understand its implications. Despite these limitations, we believe our study provides a valuable contribution to the existing body of research. In light of these limitations, future research should aim to address and account for these factors for a more comprehensive understanding of the local politics of public health declarations.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Robert Wood Johnson Foundation, (grant number 79754). The views expressed here do not necessarily reflect the views of the Foundation.
Notes
Author Biographies
Appendix
Market Predictors of any Declaration (State, City, or County).
| (1) | (2) | |
|---|---|---|
| Variables | Any declaration | Any declaration |
| Trump 2016 vote share | −0.0496** | −0.0669*** |
| (0.0214) | (0.0218) | |
| Population 100k | 0.0862*** | 0.0950*** |
| (0.0215) | (0.0235) | |
| Pct 65 and older | −0.0755 | −0.0232 |
| (0.0785) | (0.0932) | |
| Pct Black | −0.0621** | −0.0615** |
| (0.0274) | (0.0284) | |
| Pct homeownership | −0.0263 | −0.0575 |
| (0.0566) | (0.0597) | |
| Infant mortality | 0.400* | 0.352 |
| (0.228) | (0.237) | |
| Pct uninsured | −0.148** | −0.161*** |
| (0.0628) | (0.0621) | |
| Pct some college | 0.0407 | 0.0109 |
| (0.0283) | (0.0318) | |
| Pct rural | 0.00895 | 0.0183 |
| (0.0190) | (0.0188) | |
| Capital | 0.289 | 0.409 |
| (0.506) | (0.513) | |
| Midwest | 1.669** | |
| (0.710) | ||
| South | 0.954 | |
| (0.769) | ||
| West | 0.722 | |
| (0.784) | ||
| Constant | 0.854 | 3.835 |
| (2.901) | (3.598) | |
| Observations | 206 | 206 |
Note. State clustered standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.