COVID-19 stressors, mental/emotional distress and political support

Abstract The effects of COVID-19 on democracy and mental health are still under investigation. In this article, it is considered that, on average, higher COVID-19 stressors and symptoms of distress are associated with lower political support and that higher COVID-19 stressors are associated with higher symptoms of mental/emotional distress. This formulation was tested by conducting two online surveys in Britain in August 2020 and March 2021. Strong support was found for this hypothesis. Greater worry about COVID-19 life changes is associated with a lower evaluation of government performance on the pandemic and with a lower perceived responsiveness of the political system; higher COVID-19 stress resulting from anti-pandemic measures is associated with a poorer evaluation of government performance and, subsequently, with less trust in government. It was also found that higher COVID-19 worry and stress were associated with more symptoms of mental/emotional distress. These findings highlight that pandemic-related stressors may influence people’s political engagement and mental health.


1
The COVID-19 pandemic has renewed scholarly interest in the effects of external stressors, like pandemics and disasters, on democracy.With some exceptions (Kritzinger et al. 2021;Amat et al. 2020;Altiparmakis et al. 2021;Jennings et al. 2021), investigators have found that the COVID-19 pandemic has strengthened, or at least has not undermined, core dimensions of political support such as institutional trust and satisfaction with government measures (Yam et al. 2020;Schraff 2020;Oude Groeniger et al. 2021;Esaiasson et al. 2020;Sibley et al. 2020;Eggers and Harding 2021;Graffigna et al. 2021;Jørgensen et al. 2021;Bol et al. 2021;Lupu and Zechmeister 2021).Some researchers have suggested that, in times of crisis, citizens share a common external threat that tightens bonds among them and leads them to abide with and approve institutional responses, which in turn has beneficial effects on political support (Bol et al. 2021).Other investigators suggest, instead, that the high level of uncertainty to which citizens are exposed increases anxiety and the need for security which, in turn, reinforce political trust (Schraff 2020), which researchers have labeled the "rally-round-the-flag" effect (Lambert et al. 2011).
We argue that, although valuable, this formulation may only apply to the beginning of the pandemic and ignores the role that psychological stressors and mental/emotional distress might play during a pandemic or a disaster.While previous studies have used sophisticated designs to model the effects of the COVID-19 pandemic or the subsequent anti-pandemic policy responses, by neglecting citizen emotional responses they offer an incomplete assessment of the situation.
The growing literature examining the negative consequences of disasters on psychological well-being (Norris et al. 2002;Bonanno et al. 2010), which have been well-documented during COVID-19 (O'Connor et al. 2021;Perlis et al. 2021;Zhao et al. 2021;Gotlib et al. 2021), supports this position.More specifically, we posit that psychological stressors originating from perceptions of policy measures and from life changes due to a pandemic may undermine rather than promote (or stabilize) political support, on average.This hypothesis builds on extant work examining the effects of anxiety on evaluations and attitudes (Marcus et al. 2000;Valentino et al. 2008;Albertson and Gadarian 2015;Huddy et al. 2007).Anxiety may decrease political support conditional on the origin of the threat and the expertise and relevance of the actor involved.Moreover, anxiety may foster blame attributions by facilitating higher levels of information processing.Thus, the rally-round-the-flag effect may decay at later stages of the pandemic (Kritzinger et al. 2021) when the emergency is no longer a shock and is external to the government.Consequently, citizens who are more vulnerable to COVID-19 anxieties may engage more in information-seeking and processing and attribute responsibility of the situation to the actors who are ultimately in charge of solving the problem.
We argue further that mental/emotional distress mediates the relation between pandemic stressors and political support.We believe that incorporating mental health is important not only because the research described above has documented that the pandemic has significantly affected people's psychological well-being, but also because poor mental health has been found to reduce political participation (Landwehr and Ojeda 2020;Couture and Breux 2017).More importantly, researchers have recently found that depression negatively affects people's perceptions of government responsiveness (Bernardi et al. n.d.).Based on findings of studies of cognitive aspects of depression (LeMoult and Gotlib 2019), negativity biases in the way depressed people interpret information might explain why people who suffer from depression perceive government responsiveness to be lower.We apply this argument to hypothesize a negative association between mental/emotional distress and political support.At the same time, however, we acknowledge the complexity of causal direction.For instance, it is likely that mental/emotional distress also exacerbates anxiety and stress related to COVID-19.Indeed, some research has reported bidirectional associations between COVID-19 and psychiatric disorders (Taquet et al. 2021).
Based on the above arguments, we formulated the following three hypotheses: H1: Psychological stressors will be negatively associated with political support; H2: Psychological stressors will be positively associated with symptoms of mental/emotional distress; and H3: Symptoms of distress will be negatively associated with political support.
To our knowledge, this study is the first to examine the relation between mental distress and political support in the context of the COVID-19 pandemic.

Methodology
We commissioned two online surveys, conducted in August 2020 and March 2021, of a demographically and politically representative sample of the GB adult population (aged 18+) to the polling firm YouGov using their 'Political Omnibus' approach (N~1,600).
Ethical approval was previously obtained and details are reported in Section S6.The samples were recruited from an online panel using active sampling based on quotas relating to age, gender, social grade, education, region, political attention and the 2016 EU Referendum and 2019 General Election votes.The quotas were based on the following publicly available data: ONS mid-year estimates, The Census, Election and Referendum Results, and British Election Study face-to-face study.
Our survey questionnaire includes several questions about factors relating the COVID-19 pandemic that might have generated worry and stress among citizens.
Response options range from 1 (very worried / stressed) to 4 (not at all worried / stressed).We recoded the variables so that higher values denote higher worry / stress.Specifically, we asked respondents whether they were worried that they would become seriously unwell or die and whether they had the same feelings for their family and friends, and whether they were worried about their finances and about the long-lasting, negative effects of the pandemic.We also asked respondents whether they were stressed about restrictions on leaving their home, reduction in contacts with people outside their household, and wearing a face mask in public spaces.While the former set of questions assess fear and anxiety around COVID-19, the latter questions assess people's perceptions of 'anti-pandemic' measures.We created two summative indices: COVID-19 worry (range: 4-16; Cronbach's alpha = 0.69 for both August and March surveys); and COVID-19 stress (range: 3-12; Cronbach's alpha = 0.71 for August and 0.74 for March surveys). 1Depression was measured with the 9-item form of the Center for Epidemiologic Studies Depression Scale (Radloff 1977).Respondents were asked about their feelings in the past two months on the following items: "I felt depressed;" "I felt that everything I did was an effort;" "I felt hopeful about the future;" "my sleep was restless;" "I was happy;" "I felt lonely;" "I enjoyed life;" "I felt sad;" and "I could not get 'going." Response options ranged from 1 (rarely or none of the time) to 4 (most or all of the time).Scores on the CESD-9 ranged from 0 to 27 and were recoded so that higher values denote higher levels of depressive symptoms.
Anxiety was measured with the 6-item form of the State-Trait Anxiety Inventory (Marteau and Bekker 1992).Respondents were asked how often have they felt calm / tense / relaxed / upset / content / worried in the past two months.Response options ranged from 1 (never) to 4 (always).Scores on the STAI-6 ranged from 2 to 16 and were recoded so that higher values denote higher levels of anxiety symptoms.
Stress was measured with the 4-item form of the Perceived Stress Scale (Cohen et al. 1983).Respondents were asked how often in the past two months have they felt: "that you were unable to control the important things in your life;" "confident about your ability to handle your personal problems;" "that things were going your way;" and "difficulties were piling up so high that you could not overcome them."Response options range from 1 (never) to 5 (all of the time).Scores on the PSS-4 ranged from 0 to 16 and were recoded so that higher values denote higher levels of stress symptoms.
To measure political support, we built on work on diffuse versus specific support (Norris 2011;Easton 1975) and obtained data on three of her five dimensions of political support, prioritizing those relating to the government and the political system.
Therefore, although we cannot speak to associations with 'national identities' and 'approval of core regime principles and values,' we included questions assessing 'evaluation of regime performance' (external efficacy), 'confidence in regime institutions' (trust in government), and 'approval of incumbent office-holders' (satisfaction with government).We measured external political efficacy with two questions ("Public officials don't care much about what people like me think" and "The political system allows people like me to influence what the government does") that, combined, yield a standard measure of external political efficacy with values from 2 to 10.We also asked a question on a 0-10 scale about trust in government (0=not at all, 10=completely) and a question about government performance on the pandemic ("How well or badly do you think the UK Government is handling the issue of the Coronavirus (COVID-19)?" where 1 "very well", 2 "fairly well", 3 "fairly badly", and 4 "very badly") which was reverse-scored.Descriptive statistics of the main variables are presented in Table S1 while rationale and coding of control variables is described in Section S3.
Finally, to facilitate comparisons across models, in the analyses we rescaled all our key dependent and independent variables ranging from 0 to 1 and used the weight variable suggested by YouGov as a fine-tuning measure to correct any discrepancies (our results do not change substantively using the unweighted data).

Results
We present findings of our linear regression analyses in Tables 1-2, which account for the inclusion of our control variables (analyses without controls are presented in Tables S2-S3 and analyses with controls displayed are presented in Tables S4-S9 of the Online Appendix).Table 1 reports the results of the key variables of interest (coefficients and standard errors in parentheses) based on the data from the August 2020 survey, which we replicate in Table 2 with the March 2021 data.The top portion of Table 1 examines whether there is an association between psychological stressors due to COVID-19 and the three measures of political support (H1).COVID-19 worry was negatively associated with both external efficacy (slightly less than 1 standard deviation) and government satisfaction, but not with trust in government.In turn, COVID-19 stress was associated only with satisfaction in the expected direction.The central part of Table 1 presents data evaluating our hypothesis about the negative association between COVID-19 stressors and mental/emotional distress (H2).We found strong support for this hypothesis: the coefficients of both the COVID-19 worry and stress variables are positive and significant (p<.01).Further, the effect sizes are substantive: psychological stressors predict about 1.5 standard deviations in symptoms of depression and stress, and 1 standard deviation in anxiety symptoms.Finally, the bottom portion of Table 1 presents data relevant to our third hypothesis, predicting a negative association between mental/emotional distress and political support outcomes.Again, we found strong and consistent support for the hypothesis across all of the measures of mental health.
Next, we examined whether later stages of the pandemic exacerbated or alleviated the associations detected in earlier stages.Table 2 indicates that all the relations observed in August 2020 hold in March 2021, with comparable effect sizes.
The main difference in the March data concerns trust in government, which was negatively associated with COVID-19 stress.Overall, we stress that the magnitude of the effects of psychological stressors and symptoms of mental/emotional distress on political outcomes was larger than the one of any of our socio-demographic factors and, in the context of external efficacy, larger than the ones of voting behavior.Above we presented analyses that tested bivariate associations among psychological stressors, mental distress, and political support using normal OLS models and controlling for possible confounding variables.We did so because of the difficulties involved in estimating causal mediation with observational data (Bullock et al. 2010;Imai et al. 2010).However, we have also estimated three sets of mediation analyses, one per mental health measure, using the Stata 14 'sem' package. 2 Given the limitations of our data, we report methodological discussions and analyses in the Online Appendix (Section S5 and Table S10 As we mentioned earlier, we cannot exclude the possibility that the association between mental/emotional distress and political support occurs through psychological stressors.Therefore, we estimated another set of SEMs where COVID-19 stressors are the mediators.Analyses of indirect effects provide some support this formulation (Tables S9-S10).In the August 2020 survey, COVID-19 worry mediated the effects of depression and anxiety on efficacy and satisfaction, and the effect of perceived stress on efficacy.In the March 2021 survey, COVID-19 worry mediated the effects of depression, anxiety, and perceived stress on external efficacy, and the effect of anxiety on satisfaction.Future research should assess the causal nature of the associations identified here.

Conclusions
Our article makes three significant contributions.First, by examining psychological stressors and mental/emotional distress, we advance the understanding of the consequences of disasters and pandemics for democracy and political support.We found that people's worry due to COVID-19 is associated with lower perceived responsiveness of the political system and with lower satisfaction with government performance on the pandemic.In addition, we found that people's stressful feelings about anti-pandemic measures are associated with lower satisfaction with government performance on the pandemic and with lower trust in government.However, the trust finding only appears in 2021 but not in 2020, in line with previous research showing a rally-around-the-flag effect declining in a later stage of the pandemic.Thus, our findings suggest that pandemic-related worry is related not so much to confidence in regime institutions, but rather to evaluations of regime performance (Norris 2011).In addition, while the disaster literature has primarily pointed towards anti-incumbency effects, our study indicates that withdrawal effects may also play a role.
Second, our negative associations on external efficacy make sense in a context where governing parties would be more likely to listen to experts rather than to the public, and expand the scope of research on perceived responsiveness (Esaiasson and Wlezien 2017) to mental/emotional distress.Third, by expanding on the mental health measures, we extend research examining the effect of mental health problems on political attitudes, which to date has focused only mainly on depression (Bernardi 2020;Bernardi and Johns 2021;Bernardi et al. n.d.).
Although our study elucidates how psychological perceptions about the pandemic and mental health may affect political support, our analyses are limited to one country and, given that our surveys were conducted in the midst of the pandemic, we cannot determine whether the relations between mental/emotional distress and political support changed with the pandemic.Further, the relations we report above are correlational in nature and the lack of panel data prevents us from further exploring whether decreased levels of perceived responsiveness and government evaluations exacerbate COVID-19 stressors and symptoms of mental distress.Future research should examine how political perceptions may serve both to increase vulnerability to experience difficulties in mental health and to facilitate mental well-being.For instance, our finding of a negative association between COVID-19 stress and trust in government suggests that increasing institutional trust has a beneficial effect on people's feelings about anti-pandemic measures.Similarly, the negative association between mental/emotional distress and trust in government suggests that boosting institutional trust is beneficial for mental well-being (OECD 2020).Policymakers should consider these findings when drafting policies to improve mental health and well-being.

Table 2 : Analyses from March 2021 Survey H1: Negative Association between COVID-19 Stressors and Political Support
-S11) while briefly discussing them here because they provide useful insights for further testing causal relations in future studies.Symptoms of depression, anxiety, and stress all significantly mediated the effect of COVID-19 stressors on political support.Indeed, the indirect effect of distress is always statistically significant at at least p<.05, and the effect of COVID-19 stressors is