Context: Carceral institutions are among the largest clusters of COVID-19 in the United States. In response, activists and detainees have rallied around decarceration demands: the release of detainees and inmates to prevent exposure to COVID-19. This article theorizes the compounding racial vulnerability that has led to such a marked spread behind bars, mainly among race-class subjugated (RCS) communities.
Methods: The authors provide an in-depth account of COVID-19 in American correctional facilities and the mobilization to reduce contagions. They also use two survey experiments to describe public support for harm reduction and decarceration demands and to measure the effects of information about racial inequalities in prison and poor conditions inside migrant detention centers.
Findings: The authors found only one-third to one-half of respondents believe that response to COVID-19 in prisons and immigrant detention centers should be a high priority. They also found Americans are more supportive of harm reduction measures than decarceration efforts. Information about racial disparities increases support decarceration. They did not find any significant effect of information about poor conditions in migrant detention centers.
Conclusions: The conditions in carceral institutions during the pandemic—and public opinion about them—highlight the realities of compounding racialized vulnerability in the United States.
Racism, specifically, is the state-sanctioned or extralegal production and exploitation of group-differentiated vulnerability to premature death.—Ruth Wilson Gilmore, Golden Gulag: Prison, Surplus, Crisis, and Opposition in Globalizing California
Crisis and Resistance in San Quentin State Prison
Christopher Hickson tested positive for COVID-19 on a Saturday morning. “I felt devastated. I have watched on television and read about it. I feared for my life because of how many precious lives this virus has taken,” said Hickson in an interview. It was June 27, 2020, when a guard notified him of his status and instructed him to remain in his cell (Moreno Haines 2020). Later that day, Hickson and 60 other COVID-19–positive prisoners in San Quentin State Prison in California were moved to Badger Section, where they were forced to withstand isolation without cleaning supplies or electrical power. Hickson and 19 other prisoners launched a hunger strike to draw attention to their plight. The group demanded improved conditions in Badger Section and in San Quentin as a whole as well as a halt to all transfers between and inside California prisons. They claimed that, rather than containing the virus, shifting COVID-19–positive individuals between facilities simply guaranteed its spread (Moreno Haines 2020).
Available data backs the assessment of the hunger strikers. San Quentin had managed to remain free of COVID-19 until May 30, when 121 people were transferred from the California Institution for Men in Chino (Moreno Haines and Weil-Greenberg 2020). At the peak of San Quentin's pandemic, 1,636 prisoners—one-third of the prison's population—were COVID-19 positive and 72 of them had died (Clayton 2020). In a matter of days, San Quentin became the largest cluster of COVID-19 in the country (New York Times2020). The hunger strikers were not merely denouncing the conditions at Badger Section: they were putting forth an analysis of a growing crisis across the country. The story of San Quentin—the protest of those inside, the reticence of the administration, the harrowing effects of COVID-19—mirrors the tragedy inside countless prisons, detention centers, and jails across the United States. According to data from the Marshall Project (2020), as of January 28, there have been more than 365,924 cases of COVID-19 and 2,314 deaths reported in carceral institutions. The five largest US clusters of COVID-19 since the outbreak of the pandemic have been correctional institutions (Williams, Seline, and Griesbach 2020). It seemed, however, that the American public had either forgotten about the people behind bars or, worse, accepted the inevitability of their suffering and death.
The COVID-19 crisis behind bars is one whose roots lay within America's twin crises of mass incarceration and structural racism. In this article, we seek to illuminate the compounding racial vulnerability brought by COVID-19 to carceral institutions. The first part clarifies the significance of racialized vulnerability in the United States, and it identifies how patterns of systematic racism set the stage for disproportionate exposure to harm during the COVID-19 crisis. The second part provides an in-depth account of the harm done by COVID-19 inside jails, prisons, and immigrant detention centers. Consequently, it outlines the policy demands expressed by advocates both inside and outside prisons. The third part describes public opinion on COVID-19 in prisons and migrant detention centers. Our two survey experiments were fielded in the early spring—prior to the wave of racial justice protests—measuring support for the policies pushed by advocates. We exposed some of our respondents to informational treatments about racial inequities in prisons and poor conditions inside immigrant detention centers. We argue that our survey findings demonstrate public indifference toward the spread of COVID-19 in prisons. Though awareness of racial injustice may bring greater attention to the crisis, it is not enough to address the racial vulnerability faced by affected populations.
Theory and Background: Racism as Exposure to Premature Death
In this article we put forth the notion of “compounding racial vulnerability” to describe the disproportionate effects of external shocks, such as COVID-19, on race-class subjugated (RCS) communities. We argue that public health crises are exacerbated among RCS communities because of the arrangements of structural racism and by governmental responses to the crisis. Our conceptualization of compounding racial vulnerability has three elements: (1) the existence of health care inequities along lines of race and class, (2) an external shock that disproportionately affects RCS communities, and (3) a governmental response that heightens the unequal impact of the shock. We argue that the disproportionate spread of COVID-19 in carceral institutions is caused by the existing inequities of structural racism. These structures are in part sustained by public opinion and public preferences along the proverbial “color line” (Du Bois  2003) of race. Our survey analysis reveals a disregard for the welfare of inmates—who mostly come from RCS communities—and a lack of political will to address the causes of the crisis, mostly among white Americans but still present to a lesser degree among nonwhites.
Our surveys also show that public indifference may be ameliorated with information about unequal racial outcomes in the criminal justice system, though not enough to prompt our respondents to prioritize COVID-19 preventive measures behind bars, support decarceration, or address the systemic roots of the crisis. The theoretical contributions made by this article are also in line with other analyses of structural racism in the context of COVID-19 (Pirtle 2020; Sewell 2020). Furthermore, we believe the notion of compounding racial vulnerability describes the consequences of other disasters among RCSs; one example other than COVID-19 may be found in the government's and the public's response to Hurricane Katrina in New Orleans. As was the case with COVID-19, Katrina largely affected the Black community, and its impact was made worse by the racially unequal distribution of resources in the city and the political indifference of government agencies (BondGraham 2007; Frymer, Strolovitch, and Warren 2006; Stivers 2007).
Understanding the impact of COVID-19 in carceral institutions and public opinion requires placing the pandemic within America's larger context of structural racism, “the totality of ways in which societies foster racial discrimination through mutually reinforcing systems of housing, education, employment, earning, benefits, credit, media, health care, and criminal justice” (Bailey et al. 2017: 1453). In contrast to the view that racism is limited to ideational prejudices or to the remnants of pre-1965 de jure segregation, the notion of structural racism describes resource and power inequities consistently perpetuated along lines of race (Bonilla-Silva 1997; Omi and Winant 2015). Bailey and coauthors (2017) argue that a theoretical focus on structural racism explains persistent inequities in public health outcomes through myriad pathways, including environmental health (e.g., the lack of drinking water in Flint, Michigan), psychosocial trauma, targeted marketing of health-harming substances, inadequate health care, maladaptive coping behaviors, and stereotype threats.
One of the most radical expressions of structural racism is the centrality of the criminal justice system to the lives of nonwhite communities in the United States. To describe the importance of both structural racism and capitalism, Joe Soss and Vesla Weaver (2017) coined the term race-class subjugated communities (RCS). These communities, they argue, are politically characterized not as formal actors in government institutions but as targets of state policies. Rather than alleviate inequities, state actors perpetuate them through practices of “coercion, containment, repression, surveillance, regulation, predation, discipline, and violence” (567).
Law enforcement agencies have been key actors in the American systematization of racism (Hadden 2003). These patterns have evolved over time into the present era of mass criminalization (Gottschalk 2015; Lerman and Weaver 2014; Western 2006). Incarceration and policing are employed in the United States to respond to the very conditions of poverty and unrest prompted by structural racism; in return, criminalization has worsened the inequalities it is meant to address (Alexander 2010; Western 2006). As the criminal justice system has expanded, it has also worsened the social vulnerability of migrant populations (Macias-Rojas 2016). Since the late 1990s, immigration and criminal justice policies have taken a punitive turn through local-federal enforcement collaborations (e.g., the Secure Communities [S-Comm] program), increased deportations, and increased migrant detention (Golash-Boza 2015; Lopez 2019; Sampaio 2015).
One need not look behind bars to see the disproportionate impact of COVID-19 on race-class subjugated communities. Data from the Color of Coronavirus Project (collected mid-March 2020 through February 4, 2021) shows an age-adjusted COVID-19 mortality rate among Blacks, Latinxs, and Native Americans twice as high as that among white people. At one point (October 13, 2020), the rate among Native Americans was three times as high as that of white Americans. Nationwide, Black Americans represent 15.7% of all deaths of known race despite being 12.4% of the population. For their part, Latino Americans have experienced 18.1% of all deaths of known race and represent 16.3% of the population (APM Research Lab 2020). These racial disparities have also been present in vaccination rates. Data from the CDC (2021) shows that as of February 26, 2021, 64.4% of those vaccinated are non-Hispanic white, while only 8.7% are Latino and 6.5% are Black.
Jose F. Figueroa and colleagues (2020) find that in the state of Massachusetts, “independent predictors of higher COVID-19 rates include the proportion of foreign-born noncitizens living in a community, mean household size, and share of food service workers.” Nonwhites are more likely to work in occupational sectors keeping them in close contact with others (Oppel et al. 2020). High residential density has also led to a quicker spread of COVID-19 among RCS: “Latino people are twice as likely to reside in a crowded dwelling—less than 500 square feet per person—as white people, according to the American Housing Survey” (Oppel et al. 2020). Resource scarcity and preexisting health issues have worsened the pandemic in areas highly populated by RCS communities, such as the Southern borderlands or Indian Country (Dickerson 2020; Walker 2020).
On top of a virus that has exacerbated structural vulnerability, governmental responses have either disregarded or harmed RCS communities. The federal government pushed for a speedy reopening of the economy and more lax public health measures merely months into the crisis, even as the contagion and death rate among Black Americans and other people of color rose (Adamy 2020). These inequalities, and the government's disregard, were made evident during the racial justice protests that ensued in the summer of 2020 and were widely met by the force of local and federal law enforcement (Krieger 2020). The federal government pushed to keep meat-processing plants open, exposing nonwhite migrant workers to COVID-19 despite rapid spread in these facilities (Swanson and Yaffe-Bellany 2020). The marginality of migrants was further affirmed through their exclusion from CARES Act relief funds, along with the exclusion of their US citizen relatives living in the same household (National Immigration Forum 2020).
The Crisis and Its Solutions
No Exit: COVID-19 in Carceral Institutions
The pandemic has infected and killed people in prisons and detention centers at far higher rates than it has affected the overall population. People in prisons and detention centers have been infected by COVID-19 at more than five times the rate of the overall population. This disparity has not been evenly distributed. Some states have avoided significant outbreaks, while a majority have seen widespread infections and deaths. Fifteen states have infection rates in prison that are at least seven times higher than the rates in the total state population (see fig. 1). As a result of these outbreaks, the pandemic has disproportionately exposed prisoners and detainees to premature death. The death rate in prisons as of October 2020 was 105 per 100,000, compared to 66 per 100,000 in the general population. By January 2021, the overall death rate had skyrocketed to 135 per 100,000, but the death rate in prison had risen to at least 188 per 100,000.1
This death rate differential, however, understates the disparity. The pandemic poses the greatest harm to older populations, especially those 65+ years old. But this demographic constitutes a relatively small percentage of the prison population. In June 2020 Brendan Saloner and colleagues adjusted the COVID-19 deaths in prison to account for the age demographics. They estimated how many prison deaths would occur if the death rate in prisons were the same within each age group as in the general population. After adjusting for age and gender, prison deaths were three times higher than in the overall population. We replicated this analysis in October 2020, and the gap had persisted, despite national deaths surpassing 200,000. After adjusting for age, prisoners are more than three times more likely to die from COVID-19. We also collected age distributions of prison populations and COVID-19 death rates by age within each state to conduct this analysis on the state level. We found a similar trend: at least seven states have exposed incarcerated people to rates of death at least five times more than their peers in the state as a whole (see fig. 1). Once outbreaks spread in prisons and detention centers, infections spread at a deadly pace, leaving the disproportionately Black and Latinx people in these facilities facing compounding racial vulnerability.2
At the root of the COVID-19 crisis inside carceral institutions lies a combination of policy failures, institutional abandonment, and government malfeasance. A report from the Marshall Project and Vice News described conditions inside Federal Correctional Institution at Elkon, Ohio, this way: “The conditions there were ripe for an outbreak, as dozens of men were packed into a dorm with stacked bunk beds, no hot water, and no access to the outdoors or sunlight for weeks.” Against CDC recommendations, prison authorities have continually mixed sick and healthy individuals with little physical distance between them (Blakinger and Hamilton 2020). Despite the CDC's call to provide personal protective equipment (PPE), face masks were slow to arrive, and when they did, they were often of poor quality.
Inmate transfers and admissions continued despite the pandemic raging inside. After a series of COVID-19 outbreaks triggered by inmate transfers in March, as was the case with San Quentin, the federal Bureau of Prisons placed a lockdown on new transfers across 122 facilities (Johnson 2020). In May these restrictions were officially lifted with 6,800 new admissions. Even during the active period of the March directive, however, transfers continued seemingly unabated across the country without proper testing or preventive measures (Phillips 2020). Detainee transfers had particularly harsh impacts in immigrant detention facilities. Between March and June, the Marshall Project and the New York Times counted more than 750 domestic flights carrying migrants under the custody of Immigration and Customs Enforcement (ICE). Despite the rapid spread rate, reports from both federal and state prisons show a lack of testing taking place behind bars (Kassie and Marcolini 2020; Rubin, Golden, and Webster 2020).
#FreeThemAll: Demands for Decarceration and Harm Reduction
The stories coming out of carceral facilities show both the deplorable conditions that led to COVID-19 clusters but also the centrality of inmate activism in advocating for better conditions. As of February 1, 2021, the University of California, Los Angeles, Law COVID-19 Behind Bars Data Project counts 229 “grassroots and other COVID-19 related efforts” and 75 “correctional population reduction requests” (Dolovich 2020). For its part, the organization Freedom for Immigrants (2020) reports 63 organizing campaigns within migrant detention centers related to COVID-19. The demands embodied by lawsuits, protests, and hunger strikes inside and outside prisons are ideologically diverse—going from reformism to prison abolitionism. An exploration of the various advocacy campaigns mentioned above reveals three main types of policy demands espoused by these campaigns:
First, there is a demand for harm-reduction practices inside carceral institutions. These include the provision of PPE, COVID-19 testing, social distancing, guaranteed services (e.g., commissaries and phone calls), and proper sanitation. The purpose of these policies is to prevent the spread of COVID-19, provide proper care for those with the virus, and procure safe living conditions in the middle of the pandemic. Harm-reduction practices are exemplified in a list of demands put together by the advocacy organization Survived and Punished (2020) to improve conditions inside the California Institution for Women: “Ensure that all incarcerated people, including those in COVID-19 quarantine, have daily access to food, clean water, and non-punitive medical care, including mental health care. . . . Immediately and freely distribute non-diluted cleaning and disinfecting supplies.”
Advocacy campaigns have also embraced demands around crisis decarceration. The purpose of these policies is to reduce the total population inside prisons and ameliorate the risk of COVID-19 spread (see Wang et al. 2020). These include calls to release elderly and physically vulnerable patients, including those with respiratory illnesses. Or, as the list of demands by Survived and Punished (2020) put it, “Release all elderly (50+) and medically high-risk people to safe quarantine outside of prison.” They also involve policies that expand release eligibility, such as those pushed and eventually implemented in the state of California (Myers and Willon 2020). Crisis decarceration demands also include a halt to new admissions and transfers across and within facilities. Pressed between a rock and a hard place, organizers inside migrant detention centers demanded expedited deportations for detainees with no more appeals or legal recourse (Narea 2020).
Finally, some organizations used the COVID-19 crisis to critique the criminal justice system's legitimacy through abolitionist decarceration demands. Whereas calls for crisis decarceration encompassed short-term policies, abolitionist decarceration demands asked for the release of all prisoners and for the abolition of incarceration as a tool of criminal justice altogether. Describing the effects of COVID-19 in carceral institutions, the Detention Watch Network campaign guidelines read: “Demand Freedom for All! This moment highlights why cages are a public health nuisance, people can't heal, recuperate, or avoid infection in jails and prisons. No one will get release unless we demand everyone be released” (DWN 2020: 2).
Public Views of the Crisis
We identified three phases of compounding racialized vulnerability: racially unequal exposure to death, an external shock that poses greater risk to vulnerable people, and poor responses to this external event that exacerbate harm. We investigate these dynamics in the context of public opinion during the COVID-19 crisis. We find that the general public places less priority on the lives of prisoners and shows reticence toward responses that involve releasing people from carceral institutions. In other words, there is reluctance to support responses that undo the disparities that created racialized vulnerability in the first place, even though there is general support for improving conditions during the crisis. However, information about the preexisting racial disparities increases support for more extensive solutions during the crisis. In this way, our survey offers a modest picture of public support for structural changes, and it also suggests that the pandemic can expose racial vulnerability such that appeals to racial disparities shift views toward more extensive responses.
Data and Methods
We conducted two surveys of American adults fielded in April and May 2020. The surveys were fielded on Lucid, a web-based survey platform. These modules were included as a part of the Yale Cooperative Lucid Surveys. Respondents were recruited through a diverse array of online methods, including digital ads and mobile games. The surveys employed a quota-based sampling method, and this created demographically diverse samples that closely approximate the US population. We also incorporated weights to improve the correspondence of the data to the general population according to age, gender, household income, racial group, region, and education level (see appendix 2.1 in the online-only appendix). We present the weighted findings for our descriptive analysis, and we present our unweighted findings for our experimental analysis. We present alternative analyses in the online-only appendix, in addition to a table with demographic summaries of the survey sample (see tables 4 and 5).
The first survey measured views on COVID-19 in prisons and jails, and the second survey measured views on COVID-19 in immigrant detention centers. The first survey was conducted in April 2020, and 1,040 American adults completed it. In the first survey, all participants were given the following text before answering the questions:
Many public health experts have identified prisons and jails as sites vulnerable to COVID-19 outbreaks. Steps have been taken to reduce the prison population. While advocates have called for more people to be released, opponents of such measures argue that it is unsafe to release people convicted of crimes as well as unfair to crime victims.
After participants received this information, we asked a series of questions about respondents' views on COVID-19 in prisons and jails. We started with general questions, and then we presented a series of policy proposals to measure support for these proposals. We followed this with some general questions about incarceration in the United States (see appendix 2.2.1 in the online-only appendix for the full list of questions).
The second survey was conducted in May 2020, and 1,080 American adults completed it. In the second survey, all participants were given the following prompt:
Many public health experts have identified immigrant detention centers as sites vulnerable to COVID-19 outbreaks. Such outbreaks have begun in many immigrant detention centers across the country. At least 47 detention centers have witnessed cases of COVID.
After participants received this information, we asked respondents about their general views of COVID-19 in detention centers, and we followed with questions about policy preferences, concluding with questions related to broader views of immigration enforcement (see appendix 2.2.2 in the online-only appendix for the full list of questions).
Both surveys embedded experiments, so we could infer the causal effects of different types of information. In the treatment condition, the initial prompt was supplemented with additional information. In the first survey, we randomly provided information about racial disparities in prison. In the second survey, we randomly provided information about poor health conditions in immigrant detention centers. In this section, we provide a description of views on COVID-19 in prisons/jails and detention centers based on these surveys. Then we provide an analysis of the causal effects of the informational treatments.
Description of Views
Devaluation of Lives
In light of the risks posed by the pandemic, we investigated how people valued the lives of those in prisons, jails, and migrant detention centers. To do this, we began each survey with a question on general views on COVID-19 in prisons, jails, and immigrant detention centers. In the first survey, we asked participants, “Which of the following best describes what you think about COVID-19 in prisons?” Respondents were given five options: (1) “Since inmates are guilty of crimes, this is a part of the punishment for their crimes,” (2) “Prisons should not be a priority at this time,” (3) “This is a serious problem, but protecting those outside prison should be prioritized over protecting those in prison,” (4) “This is a public health emergency, and the well-being of prisoners should take a high priority,” and (5) “Don't know/No opinion.” A plurality of respondents (42.5%) said that COVID-19 in prisons is a serious problem, but that protecting those outside prison should take priority. Another 17.1% of respondents said it is not a priority (11.4%) or that it is a part of punishment for crime (5.7%). Less than a third of respondents (30.2%) said that this is a public health emergency and the well-being of prisoners should take a high priority.3
The second survey showed similar patterns with reference to immigrant detainees. We asked survey participants, “Which of the following best describes what you think about COVID-19 in immigrant detention centers?” Respondents were given four options: (1) “We should address COVID-19 inside detention centers to protect the human rights of detainees,” (2) “We should address COVID-19 inside detention centers to prevent spread to American citizens,” (3) “COVID-19 in detention centers should not be our priority because the detainees are not American,” and (4) “Immigrants in detention should bear COVID-19 as part of their punishment for coming to the US without authorization.” In this survey, around half of respondents (49.5%) said that we ought to address COVID-19 in detention centers because it is a human rights issue. The other half indicated that immigrants in detention should not be a priority or should be a priority to protect American citizens. Thirty-eight percent said that we ought to address COVID-19 in detention centers to prevent spread to American citizens; 7.9% said that immigrant detainees should not be prioritized because they are not citizens; and 4.2% said that COVID-19 is a part of their punishment for coming to the United States without authorization.
Across these two surveys, one major theme emerged from this question: a majority of people either give less priority or secondary priority to the lives of people in prisons, jails, and detention centers. Most respondents believe either that we ought not prioritize protecting people in these places or that it should be a priority to protect other people whom they may infect. Prisoners are among the most vulnerable to COVID-19; yet, on average, respondents view them as less worthy of care and protection, without the same inherent value as nonprisoners. Public opinion mirrors—and sanctions—the governmental responses described above and reinforces the process of compounding racialized vulnerability: the lives of the already vulnerable behind bars face compounded harm by the lack of priority given to their lives amidst the pandemic.
Policy Preferences on Harm Reduction and Decarceration
After measuring general views, we asked about support for a series of policy proposals roughly mirroring activist demands. These policy proposals included proposals related to harm reduction (increased sanitation, more testing, medical care) and decarceration (releasing already eligible, releasing vulnerable, and releasing as many as possible). Respondents were given the opportunity to say whether they agreed with each proposal on a five-point scale ranging from “Strongly Disagree” to “Strongly Agree.” We converted responses to a numeric scale from 0 to 4, with strongly agree equal to 4 and strongly disagree equal to 0. We provide the weighted means in our summary measures, and we provide the unweighted results in the online-only appendix (see figs. 14–15 and 23–24) with similar trends present. We provide the weighted means in our summary measures, and we also combine the proportion that agree and strongly agree to create a dichotomous measure of policy support. We provide the unweighted results in the online-only appendix. Three major themes emerged from these policy responses.
First, there was general support for harm reduction, but much less support for decarceration
The weighted mean support for better sanitation in prisons was 3.11, and mean support for increased testing was 2.9. More precisely, 76% of respondents either strongly favored or favored providing increased sanitation supplies and personal protective equipment in prisons and jails. A slightly smaller majority (61% of respondents) also expressed support for ensuring that people released from prison receive necessary preventative care and medical treatment (see fig. 2).
Similar patterns held true in questions about harm reduction in immigrant detention centers. The weighted mean response for providing increased sanitation supplies and personal protective equipment was 3, and the mean response for providing medical care for released people was 2.81. Seventy percent of respondents strongly favored or favored providing increased sanitation supplies and PPE in detention centers, and 62% of respondents favored providing medical care for released people (see fig. 2).
We found much less support for releasing incarcerated people or immigrants in detention. The weighted mean response was 2.43 for releasing those already eligible and 2.53 for releasing vulnerable populations (including elderly people and pregnant women). The weighted mean support for releasing as many as possible was 1.55. In other words, 56% of respondents said they strongly agree or agree with the proposal to release people already eligible for release, and only 25% of respondents agreed or strongly agreed with the proposal to release as many people as possible (see fig. 2).
Detention centers received similar responses. The mean 50% of respondents strongly favored or favored releasing vulnerable populations, and 35% strongly favored or favored releasing as many people as possible. Both surveys confirm strong support for harm reduction but deep reluctance toward releasing people from carceral institutions.
Second, there was a strong correlation between policy preferences and the valuation of the lives of people in prisons and detention centers
As people placed less value and priority on addressing COVID-19 in these institutions, they consistently expressed less support for policies that would promote the well-being of prisoners and detainees (see appendix figs. 16 and 25 in the online-only appendix for regression coefficient plots). The weighted mean support for better sanitation increased from 2.22 to 3.08 when moving from those who said it is a part of the punishment for crimes committed and those who consider prisoners' lives to be a high priority during the crisis. This difference was even more pronounced in preferences related to decarceration. For instance, the mean support for releasing the vulnerable was 1.8 among those who did not consider COVID-19 in prisons a high priority, while it was 2.78 among those who considered it an emergency and priority (see fig. 3).
This is true in connection to immigrant detention centers as well. Among respondents who said COVID-19 in detention centers is a human rights issue, 75% strongly agreed with increasing sanitation supplies and personal protective equipment. For everyone else, only 35% of respondents strongly agreed with this proposal (see fig. 3). These differences were starker for questions related to release. Respondents who indicate that people in prisons and detention centers should not be as prioritized as others are far more likely to say people should not be released. Only one policy preference did not show a correlation across general views of COVID-19 in migrant detention centers: expediting deportations (see appendix fig. 25 in the online-only appendix). We believe this unlikely convergence of preferences to be the result of divergent beliefs. On the one hand, respondents who prioritize the health of migrant detainees may be heeding the political demands of migrant detainees and advocates.4 Respondents regarding the health of migrant detainees as a low priority may favor deportations out of a nativist preference for a higher net rate of deportations.
Overall, these findings show the correlation between the priority given to the lives of people in carceral institutions and how people believe we ought to respond to this crisis. This corresponds to the work that shows a correlation between racial resentment and opposition to voting rights for people convicted of a felony (Wilson, Owens, and Davis 2011).
Third, white respondents were more likely than Black and Latinx respondents to support harm reduction measures, but Black and Latinx respondents were more likely to support releasing more people from prisons and immigrant detention centers
White respondents were more in favor of increasing sanitation supplies and PPE, but Black and Latinx respondents were much more supportive of proposals to release incarcerated people and slightly more supportive of releasing more detained immigrants (see table 1 and appendix fig. 20).5 Black-white differences in beliefs about policing and incarceration are well documented (Jefferson, Neuner, and Pasek 2020; Unnever 2008). We extend that work here to cover differences in beliefs regarding releasing people from prisons and jails during the COVID-19 pandemic.
The Effects of Information about Racial Disparities and Health Conditions
We also studied how public views of COVID-19 in prisons and jails were affected by informational treatments. To do this, we randomly assigned half of respondents in each survey to receive a short statement related to prisons/jails or immigrant detention centers. For the first survey, we examined the effect of information about racial disparities in prisons and jails. For the second survey, we examined the effect of information about poor health conditions in migrant detention centers.
Racial Disparities in Prisons and Jails
In the survey module on COVID-19 in prisons and jails, we added the following information to the survey prompt:
Black Americans are incarcerated in the US at a rate that is about 5 times that of White Americans. Thus, it is likely a far greater proportion of the Black population will die in prisons from COVID-19 than the proportion of the White population who will die because of COVID-19 in prisons.
This information communicated the first phase of compounding racialized vulnerabilities: preexisting racial disparities in outcomes. We analyzed the differences in views related to COVID-19 in prisons and jails between those in the control group and those who received this prompt. We conducted ordinary least squares (OLS) regressions, and we included a variety of pretreatment covariates as control variables (see appendix tables 6–15 in the online-only appendix for full regression tables).
In questions related to harm reduction, this information about racial disparities had no significant effect. But people who received the informational treatment were more likely to favor releasing more people. The treatment increased agreement with releasing the vulnerable by 0.11 on a five-point scale (though this was not significant). It significantly increased agreement with releasing those already eligible by 0.21, and it significantly increased agreement with releasing as many people as possible by 0.23 (see fig. 4).
These causal effects are large and important. In the control group, 58% of respondents disagreed or strongly disagreed with the proposal to release as many people as possible. In the treatment group, this fell to 48%. The effect size for releasing the eligible is equivalent to a shift of three points on a seven-point party ID scale, and the effect size for releasing as many people as possible is equivalent to a four-point shift in party ID (e.g., from a weak Republican to a weak Democrat).6 These effects partially closed the gap between support for harm reduction and support for decarceration during COVID-19.
This effect was concentrated among white and Latinx respondents (see fig. 5).7 In the control group, Black respondents were far more likely than white respondents to support releasing people from prisons and jails, and Latinx respondents were between the Black and white respondents. But the racial disparities treatment dramatically changed these differences. The treatment had a large and significant positive effect on white and Latinx respondents, increasing their support for releasing people from prisons and jails. But it had the opposite effect on Black respondents, who were less likely to support releasing people from prison as a result of the treatment (though this effect was insignificant).
Previous research has found that information about racial disparities in the criminal justice system does not change views on the criminal justice system (Hetey and Eberhardt 2018). Respondents can simply fit information about racial disparities within their views of the world. For instance, racial disparities can fit within conceptions of Black criminality and the dehumanization of Black people (Jardina and Piston 2019; Unnever and Cullen 2010). For others, these disparities reinforce existing opposition to systemic racism within the criminal justice system (Hetey and Eberhardt 2018). But in the context of COVID-19, information about racial disparities—with no additional commentary as to why these disparities exist—changed views. It would make sense that Black Americans' views do not change with this informational treatment because they are already aware of these disparities. For others, this stark disparity combines with the death resulting from COVID-19 to prompt reconsideration of our continued patterns of incarceration.
Poor Health Conditions and Immigrant Detention Centers
In the survey module about COVID-19 in immigrant detention centers, we randomly assigned half of respondents to receive this prompt:
Conditions inside immigration detention centers do not allow for social distancing or proper hygiene. Large groups of people are often placed in one room with 3 ft. between beds. Common complaints include lack of soap, malfunctioning bathrooms, and improper medical care. Researchers suggest that without proper measures, 70 percent of ICE detainees may become infected over a 90-day period.
We hypothesized that this information could prompt people to increase support for harm reduction, given the poor conditions that currently exist. We also hypothesized that this could increase support for releasing immigrants from detention, since it could be seen as a way to save lives.
However, this information did not have a significant effect on views of COVID-19 in immigrant detention centers (see fig. 4).8 For questions relating to increasing PPE, releasing as many as possible, medical care for the released, and general thoughts on COVID-19 in immigrant detention centers, the estimate of the effect was less than 0.03 from 0 and not significant. The estimates were slightly larger for questions related to releasing the vulnerable (t = 0.049), expediting deportations (t = 0.085), and whether the United States has detained too many people (t = 0.07). The treatment effects are consistently positive across all but one of the outcome variables, but these estimates are fairly small and statistically insignificant. They are suggestive that the treatment may shift preferences, but any shift would be small. And the lack of significance prevents us from making any conclusive claims about the positive effects. (See fig. 5 for heterogeneous effects by respondent racial group.)
We have sketched a picture of the politics surrounding COVID-19 inside carceral institutions through the lens of compounding racial vulnerability. The roots of the crisis predate the virus; mass incarceration and structural racism are the underlying conditions that have rendered nonwhite detainees “vulnerable to premature death” (Gilmore 2007: 28). Institutional responses at multiple levels have heightened structural marginalization to bring about new forms of compounding racial vulnerability, often sanctioned by an electorate unwilling to protect the health of inmates and detainees. However, we also hope to amplify the vigorous organizing happening inside and outside these institutions, centering decarceration demands that have become even more prominent in the aftermath of the marches that followed the killing of George Floyd. Our survey results illustrate compounding racial vulnerability by showing the low priority given to conditions inside prisons and detention centers compared to other aspects of public life during the pandemic. We also show widespread approval for harm-reduction measures but more opposition to decarceration. Our work shows that messages highlighting racial disparities may have an effect on public support for prisoner release. It is our hope that rather than seeing these findings as a display of static opposition or infeasibility of activist demands, advocates and academics alike may see our work as a springboard to inform campaigns around carceral and racial justice.
We would like to thank Ray Block Jr. for tremendous support through various stages of this project. We also want to thank others who provided helpful feedback and support: Danny Hirschel-Burns, Dasean Nardone-White, Gwen Prowse, panel participants at the 2020 APSA meeting, editors, anonymous reviewers, Alex Coppock, Kyle Peyton, Roselyn Cruz, and Lara Takasugi Denney. Lastly, we would like to acknowledge the organizers in prisons and detention centers who have inspired this work.
See the online-only appendix for more on data sources. Because the prison population dropped significantly after the pandemic began, the denominator (prison population numbers from earlier in the pandemic) likely deflates the death rate in prisons.
See section 1 of the online-only appendix for state-by-state data and more information about these data. The variation across states reveals the differences in how state governments protected inmates with varying levels of effectiveness. Such variation highlights our broader point, since many of the primary actors are at the state and local levels. While this variation across states deserves in-depth treatment, it is beyond the scope of what we can provide here.
When presenting summary descriptive measures, we include both the treatment and control groups. As the coefficient plots in figure 5 show, this does not significantly affect our results, except in the case of policy preferences related to decarceration. In this case, including only the control group widens the gap we highlight between support for harm reduction and support for decarceration.
Detainees in Bristol County, Massachusetts, for example, demanded expedited deportation for consenting migrants who had no further legal recourse as part of their protest efforts (Narea 2020).
This graph format featuring the scatterplot and subgroup means is adapted from Coppock 2019.
See appendix table 8, model 6, in the online-only appendix. The Lin estimator normalizes coefficients and allows for this kind of comparison.
In figure 6, we show heterogenous effects by respondent racial group. We include Black, white, and Latinx respondent groups in this analysis. Our small sample sizes are relatively small in other groups, and this limits the statistical power needed to include these other groups in this analysis.
See appendix tables 16–22 in the online-only appendix for full regression tables.