*4.1. Perceived Risk of Infection*

First, Table 2 presents results from regressions on the perceived risk of infection among participants in our panel. To our mind, there are several important takeaways from this analysis.

First, in Model 1, a model with no control variables, identifying as Hispanic/Latinx was associated with a greater perceived risk of infection. However, this association was no longer present when controlling for alternative explanations in Model 2, suggesting other variables better explain variation in perceived risk of infection. Similarly, being a first- or second-generation immigrant did not appear to be associated with a greater perceived risk of infection as seen in Models 1 and 2. These results are counter to our expectations, suggesting that identification as Hispanic/Latinx or as a first- or secondgeneration immigrant do not account for risk perception by themselves.

Models 3 and 4 suggest that the interaction of these factors might be associated with greater perceived risk of infection. Identifying as a first-generation Hispanic/Latinx individual was associated with an 0.191 increase in the perceived risk of infection (*p* = 0.084) compared to individuals who are neither Hispanic/Latinx nor an immigrant, although this result did not hold at standard thresholds of statistical significance at *p* < 0.05.

It is also important to note that the results for other variables included in the models largely fall in line with what we might expect across both Models 2 and 4. For instance, we found a positive correlation between anxiety, discrimination, and taking the survey in Spanish with perceived risk of infection. It is also worth noting the size of these correlations, with Model 4 reporting a one-unit increase in anxiety associated with a 0.253 increase in perceived risk of being infected (*p* = 0.000), a one-unit increase in the discrimination index was associated with a 0.083 increase in the perceived risk of infection (*p* = 0.001) and taking the survey in Spanish was associated with a 0.539 increase in perceived risk of infection (*p* = 0.002).


**Table 2.** Perceived Risk of Infection.

Standard errors in parentheses <sup>+</sup> *p* < 0.1, \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001.

Taken together, Table 2 presents mixed results for our hypotheses, with inconclusive results about the relationship between identification as Hispanic/Latinx and as a firstand second-generation immigrant with perceived risk of infection. To better understand these results, we conducted further tests to examine differences among different ethnic subgroups within individuals identifying as Hispanic/Latinx.

Table 3 presents results from subgroup analysis that demonstrates the heterogeneity between different subgroups within the broader Hispanic/Latinx label. Model 2 demonstrates that identifying as Mexican was associated with a 0.160 increase in perceived risk of infection from COVID-19 (*p* = 0.45). However, identifying as Cuban was associated with a 0.600 decrease in the perceived risk of infection from COVID-19 (*p* = 0.008), and identifying as Central/South American was associated with a 0.280 decrease in the perceived risk of infection (*p* = 0.095). These results help to explain the findings in Table 2—there is a lot of variation within the Hispanic/Latinx community, which arguably reflects the cumbersome

all-encompassing label for a diverse group of individuals. At the same time, nativity does not appear to be meaningfully associated with perceived risk of infection in Model 2.


**Table 3.** Perceived Risk of Infection, by Subgroup.


**Table 3.** *Cont.*

Standard errors in parentheses <sup>+</sup> *p* < 0.1, \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001.

Delving deeper into the data, Models 3 and 4 show the interaction between ethnicity and migration. Again, these results demonstrate the complexity of identity and perceived risk of infection during the pandemic, with individuals identifying as first-generation Mexican associated with a 0.439 greater perceived risk of infection (*p* = 0.000) and those identifying as first-generation other Spanish individuals associated with a 0.678 increase in perceived risk of infection (*p* = 0.003) compared to individuals not identifying as Hispanic/Latinx or as an immigrant.

Finally, Table 3 reports similar findings to Table 2 in illustrating the positive relationship between anxiety, discrimination, and taking the survey in Spanish with perceived risk of infection from COVID-19. These results are consistent across Models 2 and 4, and the size and statistical significance of the results mirror those presented in Table 2, illustrating the importance of anxiety, discrimination, and taking the survey in Spanish on how individuals perceive the risk of becoming infected from COVID-19.

#### *4.2. Perceived Risk of Dying*

While it is important to understand individuals' perceived risk of infection, the COVID-19 pandemic has been particularly lethal for people of color. As a result, we turn next to analysis of individuals' perceived risk of dying from COVID-19 if they contract the virus to see whether there are systematic differences in perceived risk related to mortality from the virus.

Table 4 presents the results of models examining the relationship between identification as Hispanic/Latinx and migration on perceived risk of dying. Results from Model 2 with all covariates included show that identifying as Hispanic/Latinx was associated with a 0.256 increase in perceived risk of dying of COVID-19 (*p* = 0.001). Consistent with earlier results, no generational nativity status was statistically different from not being an immigrant.

Turning to Model 4, the interaction between Hispanic/Latinx and nativity appears to be positively correlated with perceived risk of dying. Identification as a first-generation Hispanic/Latinx individual was associated with a 0.262 increase in perceived risk of dying (*p* = 0.018), while identifying as a second-generation Hispanic/Latinx was associated with a 0.324 increase in perceived risk of dying from COVID-19 (*p* = 0.000) relative to non-Hispanic and non-immigrant individuals. In short, it appears that the combination of being Hispanic/Latinx and a first- or second-generation immigrant is related to increased perceived risk of dying from COVID-19.


**Table 4.** Perceived Risk of Dying.

Standard errors in parentheses <sup>+</sup> *p* < 0.1, \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001.

As with earlier models, results relating to our control variables largely fall in line with what one might expect, with anxiety, discrimination, Spanish language, and age associated with greater perceived risk of dying from COVID-19 if they contract the virus. Similarly, factors that might reduce health risks were associated with decreased perceived risk of dying, such as having health insurance, identifying as White, greater household income, and education. Identifying as male was also associated with decreased perceived risk of dying.

However, it is important to further analyze the perceived risk of dying by subgroup. Table 5 presents the results of this analysis. First, we found a large degree of heterogeneity in the results, with large variations in the relationships between subgroups and perceived risk of dying. Identifying as Mexican was associated with a 0.340 increase in perceived risk of dying from COVID-19 (*p* = 0.000) and identifying as Other Spanish increased perceived

risk of dying by 0.311 (*p* = 0.050). Like with previous models, nativity alone did not appear to be related to perceived risk of dying.

**Table 5.** Perceived Risk of Dying, by Subgroup.




Standard errors in parentheses <sup>+</sup> *p* < 0.1, \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001.

However, when turning to the interaction between ethnic group and immigration, we found that first- and second-generation Mexican individuals reported an increased risk of dying, by 0.541 (*p* = 0.000) and 0.377 (*p* = 0.001), respectively. Similarly, thirdgeneration Central/South American individuals were associated with a 1.629-fold increase in perceived risk of dying. The lower self-reported risk among first- and second-generation Mexicans might be explained by Latinxs individuals who feel closer to the "canonical immigrant" tending to underreport distress as a way of defying stereotyping [52,53]. As with previous models, we found that our control variables report results in line with how one might expect them to be related to perceived risk of dying from COVID-19.

Overall, while controlling for alternative explanations, identification as Hispanic/Latinx and being first-generation was associated with greater perceived health risks from COVID-19. This was particularly true of participants of Mexican descent. However, these results are complex, and speak to the importance of treating heterogeneity among the Hispanic/Latinx community seriously in social science and public health research. We also found that anxiety, discrimination, and taking the survey in Spanish were also consistently positively associated with both the perceived risk of becoming infected and the perceived risk of dying from COVID-19. In the next section, we discuss the implications of these results.

#### **5. Discussion**

The COVID-19 pandemic has affected people around the world, but its effects have been particularly acute for people of color, exacerbating inequities in healthcare that existed prior to the pandemic. In this article, our findings demonstrate that people identifying as Hispanic/Latinx, and as first-generation immigrants perceive a greater likelihood of getting infected and of dying from COVID-19 than other individuals.

Furthermore, we find important differences between different subgroups, with individuals identifying as Mexican reporting greater perceived health risk than other subgroups within the Hispanic/Latinx community. Collectively, these results build on our understanding of perceived health risks during COVID-19, demonstrating the increased perceived risks of the Hispanic/Latinx community in the US, and of first-generation immigrants especially. Building on previous studies, our results also indicate that anxiety, discrimination, and completing the survey in Spanish are also correlated with greater perceived risks of becoming infected and of dying from COVID-19. Taken together, our results add to the collective understanding of migration and public health, illustrating how ethnic

identity and migration status influence individuals' perceptions of risk during public health emergencies, including COVID-19.

While this study focuses exclusively on perceived risks and not realized health effects of the pandemic among participants in the study, these findings are significant for several reasons. First, it is likely that risk perceptions affect people's behavior, including the adoption of protective behaviors, whether people remain in the labor force, and where people live [54,55]. Risk perceptions could exacerbate existing problems and induce the cycle of harm. For one example, perceived risks could be associated with vulnerable workers continuing to work at meatpacking plants in Nebraska where social distancing measures were insufficient, or the underground economy where lawful or safe employment is not possible [32,56].

Of course, workers at meatpacking plants and in other workplace environments that are particularly conducive to the spread of COVID-19 may have few alternative options for employment [57]. Further, an individual may not feel they have much choice but to work in an environment where there may be a heightened risk of exposure to COVID-19 because they need the income regardless of how they feel about the risks from the virus, and this is a very different calculation than a discretionary decision to drink at a bar. Understanding these nuances will be critical in future research to building a thorough understanding of risk perception, attitudes, and behavior in the COVID-19 pandemic.

In particular, further research should examine the relationship between perceived risks and behavior during the pandemic, especially among the Hispanic/Latinx and immigrant communities, to better understand the adoption of protective behaviors and continued employment, especially in essential services, and how it affects internal and external migration.

Second, building on the findings of this study, there remains scope for further exploring the mental health burden of the COVID-19 pandemic on Hispanic/Latinx and immigrant individuals. This is especially critical given the extended duration of the pandemic, and the fact that the long-term implications of the pandemic on individuals' health, society, and the economy remain unknown at the time of writing.

Third, a broader implication of our results is that the pandemic may only amplify existing stresses felt by Hispanic/Latinx and immigrant individuals in the US, creating a dual crisis. A heightened anti-immigration climate may increase the chronic fear of deportation, which in turn "may exacerbate current health conditions while increasing vulnerability to others" [58] p. 592. In a pandemic, Latinx communities are likely to be impacted more as they are more vulnerable to sickness and can be afraid of going to health centers which, in the case of a communicable disease, makes it more difficult to maintain public health. The role of community advocates and the separation of ICE from local police are important in reducing fear from deportation and thus allowing immigrants to have better access to health [34]. Further research should build on this study to further examine the relationship between perceived health risks of COVID-19 and the broader social and political environment in the US and its hostility to Hispanic/Latinx individuals.

It is also important to note that there are also some limitations of this study that future scholarship could address. First, our study raises important questions about the disproportionate effect of COVID-19 on at-risk populations in the United States. Unfortunately, we were unable to examine perceptions of risk with realized health effects in this article given the nature of the study, but this presents an important avenue for future research to examine. Scholars could examine the extent to which individuals' expectations about becoming infected and dying from COVID-19 matched data about the prevalence and impact of the virus.

Second, the purpose of this study is to understand perceived risks about COVID-19 among the US population and to examine how race, ethnicity, and nativity influence risk perception. However, we are unable to speak directly to what drives our results. It is possible that some populations worry more (or less) about COVID-19 due to dissociation from the crisis, fatalistic attribution, or other factors that were not measured in the data we used. We hope that further studies examine more of the causal mechanisms associated with perceived risks among different groups.

Third, studies could also examine how individuals responded to perceived risk in their behavior regarding COVID-19. For instance, higher risk perceptions regarding becoming infected and dying from COVID-19 could be associated with greater adoption of preventive behavior to reduce the risk of becoming infected. Alternatively, individuals with higher levels of perceived risk could adopt fatalistic attitudes about the virus and become more risk-accepting. Further research should build on this study to examine the consequences of risk perception about COVID-19 on behavior.

Fourth, while this article used panel data from a national probability sample, it is observational data; therefore, causal inferences are difficult to establish. As a result, our findings should be interpreted as correlational and not causal in nature. Further work should establish the causal mechanisms through which individuals report higher levels of perceived risk relating to becoming infected and dying from COVID-19, and longitudinal studies might shed some light on these processes. This is especially important because it is possible that there is a complex causal pathway from ethnic identification and nativity to perceived risks, and anxiety, discrimination, and other factors might serve as mediating or moderating variables through this process. Future scholarship should explore these complex relationships in greater detail and establish causal pathways leading to heterogenous perceptions of risk relating to COVID-19.

Finally, we are constrained in our analysis by the classification of the Hispanic/Latinx subgroups in the dataset, and these groups are not as specific or fine-grained as we would like. For example, it is difficult to interpret the 'other Spanish' classification in the data, and it would be a good practice to have additional follow-up questions to have a better understanding of the participants identifying as 'other Spanish.' We hope that future studies might have better data, but in our opinion this limitation further emphasizes the importance of understanding heterogeneity within racial and ethnic groups more broadly. We hope that datasets will have more fine-grained data along these lines in the future so that scholars can use them to better understand people's experiences of COVID-19.

#### **6. Conclusions**

In this article, we demonstrated the positive relationship between ethnic identity and immigration status on perceptions of harm from the COVID-19 virus in the United States. Hispanic/Latinx and first-generation immigrants reported higher perceived risk of becoming infected and dying from the virus. Further analysis illustrated that this association was especially true of individuals of Mexican descent in our sample, while other Hispanic/Latinx subgroups report mixed results. Further, we found that anxiety, discrimination, and taking the survey in Spanish were positively related to perceived risk of becoming infected and dying across all models.

Taken together, our results add further evidence about the heterogeneous impact of COVID-19 on vulnerable populations in the US. Beyond increased medical risks associated with the pandemic, these results suggest that Hispanic/Latinx individuals have a higher mental health burden than other individuals with an increased perception of health risks.

The COVID-19 pandemic has had a greater worldwide impact than other public health threats in recent memory, and its full effects are yet to be realized [1–4,59]. As such, we do not know to what extent our findings might generalize to other places, settings, and times. However, we hope this study helps build on the already-impressive scholarship on the pandemic to understand its effects on vulnerable populations and immigrant communities in the US and around the world.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/ijerph182111113/s1, Table S1: Ordered Logit Regression: Perceived Risk of Infection, Table S2: Ordered Logit Regression: Perceived Risk of Infection by Subgroup, Table S3: Ordered Logit Regression: Perceived Risk of Dying, Table S4: Ordered Logit Regression: Perceived Risk of Dying by Subgroup.

**Author Contributions:** Conceptualization, T.J., D.C., B.G.-A. and C.D.-R.; Formal analysis, T.J.; Methodology, T.J., D.C., B.G.-A. and C.D.-R.; Writing—original draft, T.J., D.C., B.G.-A. and C.D.-R.; Writing—review & editing, T.J., D.C., B.G.-A. and C.D.-R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by a Research Development Program Award from the University of Nebraska at Omaha.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Informed consent was obtained from all participants involved in the study. Households who log into the UAS website are asked to agree to an online consent before they take the first survey.

**Data Availability Statement:** The project described in this paper relied on public data from survey(s) administered by the Understanding America Study, which is maintained by the Center for Economic and Social Research (CESR) at the University of Southern California. The content of this paper is solely the responsibility of the authors and does not necessarily represent the official views of USC or UAS. The collection of the UAS COVID-19 tracking data is supported in part by the Bill & Melinda Gates Foundation and by grant U01AG054580 from the National Institute on Aging, and many others. This data is publicly available here: https://uasdata.usc.edu/index.php, accessed on 19 April 2021.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analysis, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.
