key: cord-1022869-swop1o17
authors: Rolon, Vania; Geher, Glenn; Link, Jennifer; Mackiel, Alexander
title: Personality Correlates of COVID-19 Infection Proclivity: Extraversion Kills()
date: 2021-05-14
journal: Pers Individ Dif
DOI: 10.1016/j.paid.2021.110994
sha: cf6cbd040af5a89f9d62d06d3623f62e5d92387b
doc_id: 1022869
cord_uid: swop1o17

The current research sought to shed light on the behavioral science that underlies the spread of SARS-CoV-2. We tested the extraversion hypothesis, which suggests that the sociability facet of extraversion may predispose people to becoming infected with the coronavirus via greater human-to-human contact. Since extraverts seek out social opportunities and seem less likely to follow containment measures related to social distancing, we hypothesized that people who have previously become infected would exhibit greater extraversion than would those who have not contracted the virus. We measured overall extraversion and three of its facets–sociability, assertiveness, and energy levels–as well as political orientation. We collected data from 217 adults, aged 40 and older, from the US and the UK, of whom 53 had had the virus at some point prior to the study, and 164 had not. Participants who had had COVID-19 were more dispositionally sociable and were also more conservative-leaning compared to participants who had never had COVID-19. Implications regarding the behavioral science underlying the current pandemic are discussed.

At the time of writing this paper, COVID-19 has killed more than a million people across the world and has infected more than 80 million, with the numbers still growing (WHO, 2020) .

Scientists from a broad range of fields, including biochemistry (Aloisio et al., 2020) , genetics (Carter-Timofte et al., 2020; Jacob, 2020) , and immunology (Brüssow, 2020) have been working to better understand this virus, largely with an eye towards stopping it in its tracks via the development of effective vaccines. Because every human is a possible SARS-CoV-2 carrier, and because humans are social animals, understanding the behavioral science behind the virus is essential in stopping its propagation. To date, some important advances regarding the behavioral science of COVID-19 have been made. For instance, we now know that the spread of SARS-CoV-2 within a population is in large part due to "superspreader events", where infected individuals contribute to an outsized number of cases (Frieden & Lee, 2020; Wong & Collins, 2020) . We also know that social and physical distancing significantly decrease the spread (Thu et al., 2020) .

While several recent studies have assessed the influence of personality traits on COVID-19-related measures (Han et al., 2021; Makhanova & Shepherd, 2020; Zajenkowski et al., 2020) , the current research focuses on group differences in extraversion between individuals who have contracted COVID-19 at some point versus individuals who have not. Because superspreader events are social gatherings, and extraverted people tend to be relatively outgoing and sociable, to the point Zajenkowski et al. (2020) found extraversion to be positively related to situational perceptions of social opportunities even during the pandemic, extraverts would be more likely to contribute to superspreader events and less likely to comply with social distancing and quarantine measures. Indeed, studies have found a negative association between extraversion and both overall and physical distance-related COVID-19 compliance measures (Han, 2021; Nofal et al., 2020) . Carvalho et al. (2020) assessed the association between extraversion-specifically the maladaptive variants of need for attention and intimacy avoidance-and engagement with COVID-19 containment measures in a sample of Brazilian adults. They, too, found that higher scores on extraversion were associated with lower scores on thinking that social distancing is an important measure. Interestingly, extraversion did not influence engagement with hand washing, suggesting that extraverts are not refusing to abide by containment measures across the board, but just the ones that prohibit sociability. These findings highlight the difficulty in reducing sociability and social proximity in extraverted people, as well as the importance of studying specific facets of extraversion, since the maladaptive variants underscored the tendency for extraverts to circumvent social distancing.

Extraversion and its subdimensions or facets have emerged as robust and reliable constructs since the inception of early models of personality (e.g., Cattell, 1943; Costa & McCrae, 1985; Eysenck & Wilson, 1991) . While earlier scales have several facets for all Big Five traits (e.g., the revised NEO personality inventory has six subdimensions for extraversion; Costa & McCrae, 1985) , new models, such as the more contemporary Big Five Inventory-2 (BFI-2; Soto & John, 2017a) , seek to enhance predictive power while decreasing the number of facets in the interest of parsimony. In the case of extraversion, this new model identifies three facets-sociability, assertiveness, and energy level-that closely align with extraversion facets proposed in previous research. Sociability is defined as the desire to approach and engage with others; assertiveness is the willingness to communicate thoughts and ideas in social situations; J o u r n a l P r e -p r o o f and energy level relates to experiencing positive affect and aroused states like enthusiasm and excitement.

Research has found extraversion to be a significant predictor of a wide range of outcomes, such as the formation of intimate relationships (Jain & Singh, 2019), the likelihood of divorce (Sayehmiri et al., 2010) , salary in one's career (Sutin et al., 2009) , and even one's lifespan (Turiano et al., 2015) . Several of these life outcomes (e.g., romantic satisfaction, peer status and peers' acceptance, occupational commitment, and financial security, etc.) are replicated when implementing the BFI-2 (Soto, 2019) , suggesting that past literature on extraversion and the Big Five is reasonably accurate.

Regarding the specific BFI-2 extraversion facets, evidence points to different facets having unique and stronger relationships with certain outcomes. For example, life satisfaction, health, and well-being are mostly driven specifically by energy level (Rammstedt et al., 2018a; Margolis et al., 2020) . The gradient between education and extraversion over the life cycle is accounted for by assertiveness and energy level, but less so by sociability (Gensowski et al., 2021) . Meanwhile, the negative associations between extraversion, and fluid and crystallized intelligence are uniquely driven by sociability and energy level, respectively (Rammstedt et al., 2018b) . Assertiveness drives the positive relationship between extraversion and ambiguity tolerance (Jach & Smillie, 2019), while health risk-taking is predicted more strongly by sociability compared to the other two facets (Joseph & Zhang, 2021). These findings suggest that studies should not only focus on overall Big Five constructs, but also examine if any specific facets drive certain relationships.

Given the positive association between extraversion and risk-taking (see Nettle, 2006) , Joseph & Zhang (2021) finding sociability to be a stronger predictor of health risk-taking, J o u r n a l P r e -p r o o f Carvalho et al. (2020) suggesting extraverts refuse to abide by containment measures related to social interactions, and the social nature of extraversion manifesting in an increased perception of sociality opportunities even in times of a pandemic (Zajenkowski et al., 2020), we believe sociability, more so than assertiveness and energy level, to be particularly important in the risk of COVID-19 infection, perhaps through a mediated effect of relatively reckless behavior.

We were also interested in the potential relationship between political ideology and COVID-19 infection. Makhanova and Shepherd (2020) found trait pathogen avoidance predicted greater compliance with social distancing measures and with taking the threat of COVID-19 more seriously, and pathogen avoidance has been shown to consistently be positively related to social conservatism (see Terrizzi et al., 2013) . As such, one could expect conservatives to be more cautious, and, thus, less likely to contract COVID-19.

Interestingly, Makhanova and Shepherd (2020) also found greater political conservatism to be associated with higher germ avoidance but lower perceived infectability. Perhaps because the pandemic has become highly politicized (Pennycook et al. 2020) , some research has found that American conservatives are actually less likely to be concerned about COVID-19 and to practice social distancing (Rothgerber et al., 2020) . Additionally, Gollwitzer et al. (2020) found that political partisanship at county-wide levels in the U.S. strongly predicted adherence to physical distancing, which was subsequently related to COVID-19 infection rates in pro-Trump counties. While, according to Pennycook et al. (2020) , the pandemic has not been as heavily politicized in the UK as it has been in the U.S., conservatism in both countries still correlated with greater COVID-19 misperceptions, and trust in liberal news outlets in both countries was associated with fewer misperceptions and higher COVID-19 risk perceptions, which one could J o u r n a l P r e -p r o o f argue would lead liberals to be more cautious of catching the virus. As such, any relationship between political ideology and COVID-19 infection could go in either direction.

Participants were recruited via a Psychology Today blog post and snowball sampling. We also emailed COVID-19 online support groups requesting they shared the survey on their newsletters and webpages. To be eligible, participants had to currently live in either the United States or the United Kingdom, be over 40 years of age, and provide consent after reading a Participant Information Sheet with details about the study. The age criterion was chosen because younger adults are more likely to be asymptomatic (Kronbichler et al., 2020) , and allowing them to participate could increase error by virtue of potentially asymptomatic individuals categorizing themselves as never having had COVID-19.

Four hundred thirty participants began the survey, of which 172 were excluded for either not providing their age or being younger than 40. Seventeen were excluded for not providing any data for the extraversion and political orientation items. Twenty-four additional participants were also discarded due to their response to the COVID-19 status question, which we explain in closer detail in the measures section. Our final sample consisted of 217 adults (160 women, 56 men, and 1 non-binary individual, M age = 52.66, SD age = 9.00), of which 184 were from the United States.

Participants were asked if they had ever been confirmed to have had COVID-19, either through a formal viral or antibody test. We gave a third option for participants who felt neither a yes or no statement accurately described them and asked them to elaborate on their response. A total of 164 participants reported never having had COVID-19 to their knowledge; 53 reported having had a positive diagnosis at some point either through a viral or antibody test; and 24 selected the third option. Of those, 10 mentioned being clinically diagnosed by doctors based on symptoms when viral testing was not readily available. While we initially considered those 10 individuals as part of the positive group, Struyf et al. (2020) shows that diagnoses based on symptoms are not accurate enough to either rule in or rule out the disease. As a result, all 24 participants who selected the third option were excluded.

We used the Big Five Inventory-2 Short Form (BFI-2-S; Soto & John, 2017a), which divides extraversion into 3 facets: sociability, energy level, and assertiveness. Both the BFI-2 and the BFI-2-S have been shown to possess good validity (Soto & John, 2017a , 2017b , even amongst Dutch (Denissen et al., 2020; Vedel et al., 2020) , German (Rammstedt et al., 2018a) , Slovak (Halama et al., 2020; Kohút, Kohútová, et al., 2020) , and Russian samples (Shchebetenko et al., 2020) . While Soto and John (2017a, b) caution against the use of the shortened BFI-2 versions on smaller samples, Rammstedt et al. (2020) found that these perform sufficiently well in terms of their psychometric properties and their representation of the Big Five nomological network. Further, we found good internal reliability for all three extraversion facets (α range = .71 -.75) and for the overall construct (α = .74). Note that participants completed the full BFI-2-S, but we kept our analyses to only the 6 items pertaining extraversion.

Journal Pre-proof

For exploratory analyses on political ideology, we asked participants to rate how they would describe their political orientation with regards to social issues and to economic issues. Both items were on a 7-point Likert scale, with 1 being "very liberal" and 7 "very conservative."

Participants from the U.S. were also asked to rate which party they agreed with more often, with 1 being "much more with Democrats" and 7 being "much more with Republicans."

While we initially intended to run analyses separately for each nation, we were not able to recruit enough participants, particularly ones who had had COVID-19. As such, we ran several preliminary t-tests comparing both countries in terms of age, extraversion and its facets, and political attitudes. We found no significant country differences in these constructs. Thus, to increase power, both samples were combined for analyses going forward.

We ran four one-tailed independent groups t-tests with a Bonferroni adjusted alpha level of .012 (.05/4) and ran post-hoc power analyses with G*Power (Faul et al., 2007) . We found no We then conducted a post hoc hierarchical regression to better focus on the relationship between COVID-19 status and sociability while accounting for the other facets of extraversion and for country of origin, which allowed us to observe the relationship of our variables within a single analysis in order to address issues of Type-I error and collinearity. In step 1, we entered country of origin as a control variable. In step 2, assertiveness and energy level were entered into the step 1 equation. Finally, in step 3, we entered COVID-19 status. All variance inflation factors were less than .20 and all collinearity tolerance values were greater than .90, suggesting that the estimated βs in our model were well-established.

The results of step 1 indicated that country of origin did not contribute to the regression model, F(1, 214) = .01, p = .92. In step 2, the addition of assertiveness and energy level resulted in a significant change, ΔR 2 = .18, F(2, 212) = 23.25, p < .001. Finally, adding COVID-19 status in step 3 resulted in an additional significant change, ΔR 2 = .02, F(1, 211) = 5.20, p = .02.

Together, all variables accounted for 20% of the variability in sociability. Table 1 . provides a summary of the regression statistics. 

To explore any differences in political ideology between the COVID-19 positive and negative groups, we ran three two-tailed independent samples t-tests. Participants in the COVID- , and sociability and party affiliation (r = .08, n = 172, p = .32) were all nonsignificant. Thus, while there seems to be a relationship between contracting COVID-19, and both greater sociability and slightly more conservative political ideologies, the latter two do not seem to be inter-related.

The current research sought to add to the efforts to understand the COVID-19 pandemic by exploring the behavioral science surrounding the spread of SARS-CoV-2. People who contracted COVID-19 reported marginally greater sociability scores than did people who never contracted COVID-19, whereas differences in overall extraversion, assertiveness, and energy level did not differ significantly between groups. In terms of political orientation, exploratory analyses suggest that participants who had been infected with SARS-CoV-2 leaned relatively more conservative, especially regarding social issues, and, in the case of American participants, and potentially agreed more with the Republican Party than did participants who had never been infected.

While our sample size may not have been ideal, we believe our study to be a first step into looking at personality differences between people who have become infected with COVID- Nettle, 2006) argues that traits with heritable components that show high degrees of variability might have been selected because both high and low ends of such traits might have benefits and costs from an evolutionary perspective. High extraversion may provide increased social status, but also greater fitness-related costs in the form of more risk-taking behavior. Perhaps across many generations of human life, pandemics that thrive on people interacting with others wipe out sociable extraverts disproportionately.

The directionality of the extraversion-COVID-19 relationship also merits further research. Are extraverts more likely to get infected; can the virus impact hosts' social behavior;

or both? The nervous-system hijacking model (see Reiber et al., 2010 , Seitz, et al., 2020 suggests viruses may temporarily increase sociable activities in hosts to benefit their spread and replication. Given organisms like toxoplasma gondii can alter our behavior (da Silva & Langoni, J o u r n a l P r e -p r o o f 2009), and that SARS-CoV-2 shows up in neural tissue (Mao et al., 2020) and in spinal fluid (Wu et al., 2020) in severe cases-suggesting the virus is making contact with the nervous system in some way-the possibility of viruses affecting human host behavior should not go unresearched.

Because the SARS-CoV-2 spreads via human-to-human contact, the pandemic is largely an issue of human social behavior. We documented two important personality-based variables that potentially significantly relate to the spread of the virus: the sociability facet of extraversion and self-identified conservative tendencies. We hope that these findings can help inform policy and processes moving forward as we work together as a global community to stop this pandemic in its tracks and get back to life as we knew it.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

A Comprehensive Appraisal of Laboratory Biochemistry Tests as Major Predictors of COVID-19 Severity

PSYCHOMETRIC PROPERTIES OF SLOVAK SHORT AND EXTRA-SHORT FORMS OF BIG FIVE INVENTORY-2 (BFI-2)

Further validation of Slovak Big Five Inventory 2: 6-months test-retest stability and predictive power

Asymptomatic patients as a source of COVID-19 infections: A systematic review and metaanalysis

Opinión: El gobierno mexicano, entre la incompetencia y laaddición a la mentira

Behavioral immune system linked to responses to the threat of COVID-19

Neurologic Manifestations of Hospitalized Patients With Coronavirus Disease

The association between Extraversion and well-being is limited to one facet

The evolution of personality variation in humans and other animals

Who complies with COVID-19 transmission mitigation behavioral guidelines

Validation of the short and extrashort forms of the Big Five Inventory-2 (BFI-2) and their German adaptations

Relationships between personality and cognitive ability: A facet-level analysis

Short forms do not fall short: A comparison of three (extra-)short forms of the Big Five

Change in Human Social Behavior in Response to a Common Vaccine

Politicizing the COVID-19 pandemic: ideological differences in adherence to social distancing

The problem of multiple testing

The relationship between personality traits and marital satisfaction: a systematic review and meta-analysis

The pandemic exposes human nature: 10 evolutionary insights

Measurement Invariance and Sex and Age Differences of the Big Five Inventory-2: Evidence From the Russian Version

How replicable are links between personality traits and consequential life outcomes? The life outcomes of personality replication project

The next Big Five Inventory (BFI-2): Developing and assessing a hierarchical model with 15 facets to enhance bandwidth, fidelity, and predictive power

Short and extra-short forms of the Big Five Inventory-2: The BFI-2-S and BFI-2-XS

Signs and symptoms to determine if a patient presenting in primary care or

Personality and Career Success: Concurrent and Longitudinal Relations

The behavioral immune system and social conservatism: A meta-analysis

Effect of the social distancing measures on the spread of COVID-19 in 10 highly infected countries

Personality and the leading behavioral contributors of mortality. Health psychology : official journal of the Division of Health Psychology

Development and validation of the Danish Big Five Inventory-2: Domain-and facet-level structure, construct validity, and reliability

WHO Coronavirus Disease (COVID-19) Dashboard

Evidence that coronavirus superspreading is fat-tailed

Nervous system involvement after infection with COVID-19 and other coronaviruses