Can Psychological Traits Explain Mobility Behavior During the COVID-19 Pandemic?

Abstract

The current COVID-19 pandemic is a global, exogenous shock, impacting individuals’ decision making and behavior allowing researchers to test theories of personality by exploring how traits, in conjunction with individual and societal differences, affect compliance and cooperation. Study 1 used Google mobility data and nation-level personality data from 31 countries, both before and after region-specific legislative interventions, finding that agreeable nations are most consistently compliant with mobility restrictions. Study 2 (N = 105,857) replicated these findings using individual-level data, showing that several personality traits predict sheltering in place behavior, but extraverts are especially likely to remain mobile. Overall, our analyses reveal robust relationships between traits and regulatory compliance (mobility behavior), both before and after region-specific legislative interventions, and the global declaration of the pandemic. Further, we find significant effects on reasons for leaving home, as well as age and gender differences, particularly relating to female agreeableness for previous and future social mobility behaviors. These sex differences, however, are only visible for those living in households with two or more people, suggesting that such findings may be driven by division of labor.

Keywords

Big Five personality traits COVID-19 social mobility Google mobility data

A large body of work in personality psychology reveals two important points about personality: First, personality traits are stable, heritable, and consistently predict behavior even across cultural or environmental variation (Black et al., 2010); second, personality represents fundamental responses to biological challenges encountered across human evolution (Heine & Buchtel, 2009). It is not surprising, then, that different cultures, which face distinct structural and ecological challenges, vary in personality (Schmitt et al., 2007).

No personality trait is unconditionally optimal (Nettle, 2006); any universally beneficial trait would quickly become universal. Instead, each personality trait represents a fundamental behavioral trade-off. For example, extraversion seems to confer a swath of important benefits—more sexual and reproductive opportunities (Nettle, 2006; Whyte et al., 2017, 2019), greater social support (Franken et al., 1990), and (in some cultures) life satisfaction (Kim et al., 2018). Yet social interaction is not without risk, and extraversion is likely to be a liability when pathogen loads are high. Accordingly, exposure to pathogens is associated with reductions in both extraversion and agreeableness (Mortensen et al., 2010; Schaller & Murray, 2008). Similarly, agreeableness may facilitate cooperation and trust among group members (White et al., 2012)¹—yet excessive trust and regard for the interests of others can lead to exploitation or may cause individuals not to advocate for their own interests (Nettle, 2006).

These personality traits are thought to facilitate adaptive behavior that aids in cooperation or disease avoidance—but do extraverted people and nations actually fare worse when a pandemic arrives? Do agreeable and conscientious people actually cooperate more when the stakes are high? The COVID-19 pandemic offers a unique opportunity to test these theories of personality, disease threat, and cooperation hold using real-world data impacting people socially, emotionally, and cognitively (Kravitz et al., 2020).

In the present investigation, we examine whether the Big Five personality traits (Costa & McCrae, 1995; McCrae & Costa, 1999), at both national and individual levels, are associated with compliance to mobility restrictions during the COVID-19 pandemic. We argue that mobility is partly an issue of compliance or conformity but is also largely an act of cooperation. Given previous work on personality and pathogen threats (Mortensen et al., 2010; Schaller & Murray, 2008; Schaller & Park, 2011), we expect several broad trends. First, we expect extraverts to value social interaction more highly and thus to discount the risks faced by social mobility; thus, nations and people high in extraversion will be less likely to reduce their mobility than more introverted nations and people, particularly in domains associated with optional social behavior. Second, given that conscientiousness is associated with reduced risk tolerance and a future-oriented mindset, and agreeableness with empathy and cooperation, we expect conscientiousness and agreeableness to be associated with greater reductions in mobility, perhaps especially when restrictions are required by law.

In Study 1, we combine Google mobility data (Google LLC, 2020) and nation-level personality data to examine whether national differences in personality traits are associated with reductions in social mobility. Results vary depending on the domain of mobility (e.g., school vs. social gatherings), but we broadly find that agreeableness and (less consistently) conscientious regions show greater reductions in mobility, whereas regions with greater openness to experience show smaller reductions. Extroverted nations also show smaller reductions in mobility, but this effect disappears when controlling for other personality traits.

In Study 2, we seek to replicate these findings at the individual level, using data from the International Survey on Coronavirus (Fetzer et al., 2020), which includes personality and demographic variables for 113,083 individuals across 164 countries collected in March and April 2020. We again find that conscientiousness and agreeableness (but only for women) are associated with the likelihood of staying home; contrary to region-level results, individual-level openness to experience and neuroticism are both positively associated with staying home. The only personality trait that negatively predicted staying at home was extraversion.

General Method

In this study, we will use both dynamic mobility data at the regional or national level and cross-sectional individual mobility data based on large-scale international survey. We run random-effects models when exploring Google regional mobility data (Study 1) and employ logit or ordinary least squares (OLS) regression when working with the international survey data (Study 2). Analyses were performed in Stata 16 (Version 16.1 Stata/MP).

Study 1

To begin analyzing how personality traits influence social mobility, we look at regional and cross-country data over time. Daily data provide the opportunity to link the social mobility data with different stages of the pandemic and allow us to take into account the dynamic nature of the pandemic and how responses occurred at different periods. The different stages of responses also hold different implications for the expected level of norm compliance (e.g., recommended vs. mandatory social distancing).

Data

Mobility data

We use mobility measures on a country and regional level from the COVID-19 Community Mobility Reports (Google LLC, 2020; see also Chan et al., 2020). The data are anonymized and aggregated based on Google users who have opted in to their location history service. Mobility measure records the percentage change in total number of visitors to places classified as retail and recreation, grocery and pharmacy, parks, transit stations, workplaces, and percentage change to duration of stay at residential places within the geographic area, from February 15 to May 5, 2020, compared to the median value of the same day of the week between January 3 and February 6, 2020. We use principal component analysis to extract the first principal component from the six (standardized) mobility measures to form our outcome variable. The first principal component has an eigenvalue of 4.83, explaining 80.45% of the total variance of the partial mobility to each of the six localities.² The corresponding eigenvectors are .438 (retail and recreation), .404 (grocery and pharmacy), .31 (parks), .433 (transit stations), .420 (workplaces), and −.429 (residential), which serve as the component loadings of the mobility measure. The sample includes 586 geographical units from a total of 31 countries, with 24 countries at the subnational region level (579 regions) and seven countries at the national level (see Supplemental Material for details on region classification). For privacy reasons, Google censored values if the traffic volume is not high enough to ensure anonymity.

Personality traits

Big Five personality traits at the country level are based on Terracciano et al. (2005). The study explored whether national character is reflected in personality trait levels looking at 49 cultures. The five factors are standardized to unit variance.

COVID-19 response indicators

We explore an indicator variable denoting period before and after the World Health Organization (WHO) declared a worldwide pandemic on March 11, 2020. A set of government response indicators (recorded daily on a country level) on closures and containment relating to schools, workplaces, public events, private gatherings, public transport, residential confinement, and domestic travel is obtained from the Oxford COVID-19 Government Response Tracker (OxCGRT). Each indicator categorizes the level of strictness of the respective policy on an ordinal scale.³

Controls

From the OxCGRT database, we also obtain the daily record of the number of COVID-19-related deaths and confirmed cases, taken from the European Centre for Disease Prevention and Control and from the Johns Hopkins University (JHU) Center for Systems Science and Engineering (CSSE) data repository. We derived the number of days since or before the first confirmed death in the country and number of confirmed cases (in natural log) as controls. We also include country-level variables: population density (people per squared km of land area), urban population (%), share of population over 65, gross domestic product (GDP) per capita (constant 2010 USD, in log), and unemployment rate (percentage of total labor force) obtained from the latest World Development Indicators. In terms of household composition, we control for the average household size and the percentage of households whose head is aged between 20 and 64 years; these data are obtained from the United Nations Household Size and Composition (United Nations, 2019). We also include a binary variable denoting whether the day is a weekend according to the each country’s definition.⁴ Finally, we also control for the average temperature (tenths of degrees C), recorded on a daily basis at the regional level, obtained from the Global Historical Climate Network Daily database (see Supplemental Material for details).

Descriptive statistics for the variables used in Study 1 are reported in Table S2 in the Supplemental Material. Correlations between personality traits are reported in Table S3, showing high levels of correlations between the five personality traits (based on our sample countries).

Method

First, we model the relationship between mobility changes and personality traits with random-effects (variance components) model, given the panel structure of the data where mobility is measured daily across the entire sample period at the regional level. Table S1 lists the number of regions for each country in the sample with mobility measures.⁵ In each regression, we include all control variables to control for mobility changes due to heterogeneity in social and population structure, the severity of the situation, containment policies across countries, and temperature across geographic areas.⁶ Due to the high correlation between traits, we conduct our analysis examining one trait at a time. For robustness, we also present the result including all five traits in a single specification. Then, we examine whether such relationships are more salient after the WHO declares COVID-19 as a world pandemic by including an interaction term between trait and the declaration variable. To understand how different personalities respond to each confinement policy, we also show the changes to mobility with respect to each personality trait at different levels of strictness for each policy. We conduct all analyses with interaction terms for each trait separately.

Results

Our random-effects regression reveals that changes in mobility are significantly predicted by openness (β = .0725, standard error [SE] = .0313, 95% confidence interval [CI] = [.0112, .134], p = .0205) and agreeableness (β = −.0511, SE = .018, 95% CI = [−.0864, −.0158], p = .00454; see Figure 1). Regions with one standard deviation (SD) higher in openness are associated with a 0.1 SD increase in overall change in mobility; likewise, regions with 1 SD higher in agreeableness have 0.05 SD decrease in mobility change. Extraversion seems to be positively associated with mobility change (β = .0592, SE = .0306, 95% CI = [−.000829, .119], p = .0533), but the effect is not precisely estimated when other personality traits are controlled for (β = .0593, SE = .0583, 95% CI = [−.0549, .174], p = .309). On the other hand, conscientiousness is only negative and significant when all five traits are included in the model (β = −.0459, SE = .0239, 95% CI = [−.0928, .000927], p = .0547).

Figure 1.

Change in human mobility during COVID-19 as predicted by regional-level personality traits. Note. Results are shown for models with a single trait (blue) and all five traits (red). All estimates are obtained from random-effects model with control variables (Table S4). Error bars represent 99.9%, 99%, 95%, and 90% confidence intervals, respectively.

The results in Figure 2 show that the effects of extraversion (χ² = 5.35, p = .021), openness (χ² = 11.86, p < .001), and agreeableness (χ² = 18.72, p < .001) on mobility change are more prominent in the period after COVID-19 was declared as a global pandemic (the predeclaration slope is not significantly different from 0, except for openness [at 10%]). Although the interaction term for conscientiousness is significant (χ² = 5.44, p = .02), the postdeclaration slope is significantly different from 0. To be clear, the results in Figure 2 represent regression coefficients, not absolute levels of mobility. Thus, the fact that openness and extraversion are more highly associated with mobility after pandemic declaration does not mean that people high on these traits increased the movement but only that these traits were more diagnostic of mobility in the postdeclaration period.

Figure 2.

Big Five before and after pandemic declaration (each bar represents the slope estimate predeclaration and postdeclaration period). Note. The signs show whether the two slopes (effect of personality on mobility) are significantly different from each other. n.s. = not significant. ^† p < .10. *p < .05. **p < .01. ***p < .001.

In Figure 3, we see regions that are higher in extraversion and openness are more likely to increase school-related mobility when only “recommended closures” are in place. Together with Figure 4, the results show that mobility incorporating public transport, workplaces, and residential confinement reduces significantly and uniformly for all traits (at all levels) once policies restricting movement are put in place. Interestingly, for mobility relating to domestic travel (internal movement) and public events (see Figures 5 and 6), regions high in openness and extraversion appear to behave in a counter-compliant way once movement restrictions are put in place. Environments high in openness and extraversion appear to see an increase in citizen mobility relating to domestic travel (internal movement) and public events. Overall, higher levels of agreeableness compared to other traits lead to stronger mobility responses when recommendations are in place or when restrictions are loose. On the other hand, more extraverts and more open societies are less likely to reduce their mobility due to enforcements.

Figure 3.

Big Five and mobility change with respect to school and public transport closure policies. Note. Estimates obtained from Tables S6 and S7.

Figure 4.

Big Five and mobility change with respect to workplace closure and stay-at-home policies. Note. Estimates obtained from Tables S8 and S9.

Figure 5.

Big Five and mobility change with respect to gatherings restrictions and public events cancellation policies. Note. Estimates obtained from Tables S10 and S11.

Figure 6.

Big Five and mobility change with respect to internal movement (domestic travel) restriction policies. Note. Estimates obtained from Tables S6 and S7.

Study 2

In Study 2, we complement Study 1 by utilizing individual-level data on psychological traits and social mobility, controlling for both individual and unique regional-based characteristics. Utilizing participant variables such as age, gender, marital status, years of schooling, household size, self-rated health, as well as country-specific data (population density, climate, GDP per capita, age distribution) and national COVID-19 information (total number of confirmed cases, number of days since first death, government response), we are able to provide a more nuanced analysis of the role of personality in norm compliance (e.g., recommended vs. mandatory restrictions on social mobility).

Data

Individual-level data

In Study 2, we utilize micro-level data from the International Survey on Coronavirus (Fetzer et al., 2020; N = 113,083) from 164 countries between March 20 and April 16, 2020. Table S13 shows the list of countries with more than 300 respondents;⁷ 48,894 (43.62%) male (M _age = 39.39; SD _age = 12.9), 63,193 (56.38%) female (M _age = 38.45; SD _age = 13.05), 0.88% (n = 996) selected other (M _age = 37.25; SD _age = 17.1; see Figure S1 and Table S14 for summary statistics on respondents’ characteristics).

Personality trait

Based on the 10-item Personality Inventory (TIPI; Gosling et al., 2003), the Cronbach’s αs (scale reliability coefficient) were .67, .3, .54, .65, and .4 for the extraversion, agreeableness, conscientiousness, emotional stability, and openness to experience scales, with inter-item correlation of .5, .18, .37, .49, and .25, respectively. The Cronbach’s α results are quite similar to the original version (.68, .40, .50, .73, and .45, respectively; Gosling et al., 2003). Personality trait measures are calculated as the average of the two items (and are reverse coded). The resulting scales are standardized across individuals; Table S15 shows the summary statistics and correlations of the five personality traits.

Mobility

We took the responses (self-rated from 0 = does not apply at all to 100 = applies very much) to questions regarding near past behavior: “To what extent do the following statements describe your behavior for the past week?” (1) “I stayed at home” (M = 80.53, SD = 24.57) and (2) “I did not attend social gatherings” (M = 90.95, SD = 23.35) as indicating mobility in the past. Propensity to leave home in the near future is captured by the binary response (yes or no) to the question: “Do you need to leave your home in the next 5 days?” (59.82% indicated yes). The correlation between participants’ tendency to stay home in the past week and need to leave home in the next 5 days is −.274 (p < .001, N = 113,083). Participants were also asked the reasons for leaving home where respondents chose multiple response options. Summary statistics and correlations of mobility responses are presented in Table S16.

Reasons to leave home

We use principal component analysis (PCA) to reduce the dimensionality of reasons to leave home (excluding the option other) into three factors with an eigenvalue above 1 (see Table S17).⁸ The first factor (psychological needs) has high factor loadings on reasons such as getting bored (.73), getting tired of being inside of the house (.7), exercising my freedom (.66), and meeting friends or relatives (.5); the second factor (deemed basic necessity) has high factor loadings on procuring food for yourself or family (.76) and doing physical activity (.65; with loadings just below 0.5 for going to work [.43] and walking a pet [.48]); the third factor (deemed medical necessity) has high factor loadings on going to the pharmacy (.62), going to the hospital/receiving medical treatments (.69), and taking care of dependents (.5). The reason getting some adrenaline (from breaking the law) does not have high factor loading in any factors but is relatively high for Factor 1 (.36).

Method

We regress the two outcome mobility measures, past behavior and future propensity, on participants’ five personality traits using an OLS model and logit model, respectively. For mobility measures of past behavior and personality traits, we standardized the variables for ease of interpretation of effect size. For logistic models concerning future propensity mobility, we report the odds ratios (ORs). In each regression model, we control for respondents’ demographic characteristics such as age, gender (male, female, or other), marital status (married/cohabiting or single/divorced), years of schooling, and household size. A squared term for age is also included to assess nonlinearity of the age effect. We also control for participants’ self-rated health and self-reported number of comorbid conditions (e.g., cardiovascular diseases or diabetes). Additionally, we controlled for the total number of confirmed COVID-19 cases (in logs) in the country where the respondent resides on the day of survey completion, the number of days since the first nationwide COVID-19-related death, and a set of government response indicators relating to social movement restrictions (e.g., school, workplace, or public transport closures) recorded daily. Finally, we include country-fixed effects (i.e., inclusion of country dummies to control for country-level heterogeneity).⁹ We analyzed the data with three model specifications: (1) without sampling weights; (2) employing sampling weights constructed by Fetzer et al. (2020) according to representation of income, education, and age and gender structure in the population in each country; and (3) sampling weights, which also take into account the size of the population of each country. In addition, we present the findings without any controls in Table S18 with qualitatively similar results.

Results

Extroverts report a lower likelihood of staying home in the previous week (β = −.025, SE = .005, Model 1 in Table 1). Individuals with higher openness (β = .038, SE = .005) and or agreeableness (β = .012, SE = .004) report the opposite and are more likely to stay home in the previous week. Those lower in openness report being less likely to leave home in the next 5 days (OR = .97, SE = .006, Model 4 in Table 1). However, these effects of personality traits on mobility are quite small in general.¹⁰ Strong age and gender effects are evident in our control measures. Younger and older respondents are more likely to have stayed home and less likely to leave home in the near future. Females (compared to males) are also more likely to have stayed home previously (β = .037, SE = .015) and continue to stay at home in the future (OR = .79, SE = .044). Finally, the number of confirmed cases in the country is negatively correlated with the tendency to have stayed home in the past week (β = .19, SE = .079). For participants who said they need to leave home in the next 5 days, we do not find a strong link between personality traits and specific reasons to leave home (Table S19), with the exception that participants with higher neuroticism are more likely to leave home due to psychological needs (β = .088, SE = .045, Model 3, Table S19). Interestingly, males (as opposed to females; β = −.21, SE = .085) or those who rate themselves healthier (β = .13, SE = .06) are also more likely to report needing to leave home because of psychological needs. We also observe a strong nonlinear age effect (inverted U shape) on needing to leave home due to basic necessity.

Table 1.

Propensity to Stay Home During COVID-19 and Personality Traits.

Variable	Stayed Home Past Week			Need to Leave Home in the Next 5 Days
	(OLS)			(Logit)
	(1)	(2)	(3)	(4)	(5)	(6)
Personality traits
Neuroticism	.017**	−.0033	−.011	0.99	0.97	0.92
Neuroticism	(.00638)	(.0131)	(.0261)	(0.0119)	(0.0464)	(0.0966)
Extraversion	−.025***	−.042**	−.083*	1.00	0.95	0.98
Extraversion	(.00504)	(.0129)	(.0320)	(0.0107)	(0.0367)	(0.0751)
Openness	.038***	.046*	.043^†	0.97***	0.91	0.89
Openness	(.00511)	(.0181)	(.0228)	(0.00608)	(0.0593)	(0.0797)
Agreeableness	.012**	.041***	.033	0.99	0.93	0.83*
Agreeableness	(.00404)	(.0122)	(.0240)	(0.00834)	(0.0454)	(0.0651)
Conscientiousness	.014***	−.0033	.0076	0.96***	0.96	1.02
Conscientiousness	(.00281)	(.00912)	(.0268)	(0.0103)	(0.0467)	(0.116)
Controls
Age	−.018***	.0051	−.0051	1.11***	1.09***	1.11**
Age	(.00305)	(.00742)	(.00970)	(0.0120)	(0.0208)	(0.0365)
Age*age	.00017***	−.000064	.000065	1.00***	1.00***	1.00**
Age*age	(.0000366)	(.0000816)	(.000105)	(0.000121)	(0.000207)	(0.000330)
Gender
Male	(ref.)	(ref.)	(ref.)	(ref.)	(ref.)	(ref.)
Female	.037*	.055*	.027	0.79***	0.83*	0.70*
Female	(.0149)	(.0247)	(.0524)	(0.0419)	(0.0779)	(0.125)
Other	−.033	−.23	−.36	0.84*	0.62	5.24
Other	(.0216)	(.140)	(.253)	(0.0613)	(0.208)	(10.72)
Marital status
Married/cohabiting	(ref.)	(ref.)	(ref.)	(ref.)	(ref.)	(ref.)
Single/divorced	−.029*	.013	.098	1.00	0.84	0.87
Single/divorced	(.0133)	(.0372)	(.0637)	(0.0295)	(0.129)	(0.235)
Years of schooling	.0089***	.0087*	.012	1.00	1.01*	1.03
Years of schooling	(.00131)	(.00423)	(.00706)	(0.00294)	(0.00663)	(0.0192)
Household size
One	(ref.)	(ref.)	(ref.)	(ref.)	(ref.)	(ref.)
Two	.018	.012	.13	0.88***	0.76	0.97
Two	(.0159)	(.0505)	(.0978)	(0.0331)	(0.157)	(0.333)
Three	.023	−.095	.11	0.82***	0.73*	0.89
Three	(.0155)	(.0624)	(.105)	(0.0425)	(0.101)	(0.305)
Four to five	.030	−.017	.16	0.74***	0.71*	0.89
Four to five	(.0198)	(.0635)	(.101)	(0.0387)	(0.108)	(0.316)
Five+	.023	−.039	.16	0.72***	0.67**	0.92
Five+	(.0200)	(.0768)	(.117)	(0.0427)	(0.0963)	(0.387)
Self-rated health	−.016	−.039	−.055	1.00	1.00	0.92
Self-rated health	(.0140)	(.0293)	(.0353)	(0.0136)	(0.0448)	(0.113)
Comorbidities	.024**	−.011	−.19*	0.97	1.04	0.95
Comorbidities	(.00857)	(.0412)	(.0770)	(0.0231)	(0.0966)	(0.108)
Day since the first death	−.013*	−.024*	−.0077	1.00	1.00	1.04
Day since the first death	(.00663)	(.0109)	(.0176)	(0.00810)	(0.0415)	(0.0849)
ln(# confirmed cases +1)	.19*	.26***	.14	1.18**	1.15	0.63
ln(# confirmed cases +1)	(.0787)	(.0729)	(.134)	0.88***	0.76	0.97
Stayed home				0.43***	0.37***	0.31**
Stayed home				(0.0715)	(0.0902)	(0.122)
Government response indicators	Yes	Yes	Yes	Yes	Yes	Yes
Country-fixed effects	Yes	Yes	Yes	Yes	Yes	Yes
Weights	Unweighted	Weights 1	Weights 2	Unweighted	Weights 1	Weights 2
Observations	105,857	105,739	105,739	105,857	105,739	105,739
Number of clusters	130	128	128	130	128	128
R ² or McFadden’s R ²	.122	.137	.309	.106	.124	.238

Note. For OLS regression models, coefficients are presented. For logistic regression, odds ratios are presented. Standard errors (clustered at country level) are in parentheses. Reference categories are male, married/cohabiting, and one member in household. Weights 1 = sampling weights calculated to reflect sample representation in terms of age, gender, income, and education in the country’s population; Weights 2 = sampling weights also accounting for the different population sizes across countries; OLS = ordinary least squares.

^† p < .10. *p < .05. **p < .01. ***p < .001.

Exploring the sex-differentiated age distribution of participant responses for previous and future mobility variables shows that (in general) males are more likely to need to leave home in the next 5 days (by 6 percentage point, on average, z = 20.34, p < .001), with this pattern holding across all age groups, while gender differences in past behavior (stayed home in the last week) are only observed for 30 to 45 and 65+, with females reporting a higher tendency toward staying home during the past week than their male counterparts (Figure 7). For both mobility variables, we see a clear pattern that younger and older people are more likely to remain at home. Interacting the Big Five traits with sex shows a strong statistically significant gender effect difference for agreeableness (Figure 8) but not on other traits. More agreeable women respond substantially stronger with a reduction in mobility than more agreeable men. This effect holds for both mobility variables (stayed home past week: β = .019, p = .004; need to leave home in the next 5 days: OR = .955, p = .047). We further explore whether the sex-differentiated effect of agreeableness on mobility can be explained by division of labor in the households. Specifically, we find that the sex difference in the effect of agreeableness is only visible for participants living in households with two or more members but not for single-member households where division of labor is not possible (Table S21).

Figure 7.

Propensity to stay home across participants’ age by gender.

Figure 8.

Interaction between Big Five and gender on stay-at-home. Note. Estimates obtained from Table S20.

Discussion

Living in modern developed economies—particularly in dense urbanized population—provides significant advantages to the human species; for example, legal and economic advantages such as assigned property rights, economies of scale for commerce, health care and education, public infrastructure, and many different forms of government support services. There are also personal, interpersonal and social advantages, such as increased marital and reproductive opportunities, exposure and access to science and the arts, the luxury of sporting endeavors, and religious and political freedoms. However, these potential gains come with risks and costs. Societies are often structured in ways that reflect these particular trade-offs—some aspects of culture may, for example, prioritize order over creativity (Jackson et al., 2019). Rule of law dictates that populations must behave in a rational, organized, and uniform way. A “herd mentality,” whereby each individual’s reciprocal agreement and compliance with not just the law, but also the social norm, results in both individual and socially optimal outcomes. A health epidemic presents a unique problem for society in that a potentially life-threatening, exogenous shock may cause some individuals to question whether they should continue to make decisions in line with what is best for themselves, or for society, especially when the outcomes for each may differ. A global pandemic is even more unique in that the micro and macro behaviors of specific regions or countries can systematically impact other societies and environments. Exploring whether country-level psychological trait analysis can explain regulatory compliance (mobility behavior) is not just important for the current global COVID-19 crisis but may also provide vital insight into a host of current and ongoing transnational issues and behaviors such as global trade, public health and education, armed conflict, climate change, and human rights or disasters and high-stress life events in general (Gomez et al., 1999; Kopala-Sibley et al., 2016; Savage et al., 2020). The degree to which societies and individuals are impacted by their personality type, how they cope with stress (Blatt & Zuroff, 1992; Ingram & Price, 2010; Monroe & Simons, 1991), and the coping mechanisms they choose to employ based on those types (see Chung et al., 2005) is less clear.

The current study explores how psychological traits can be utilized to understand and explain domestic mobility behavior during the COVID-19 pandemic. Our study repeatedly finds statistically significant relationships between particular traits and regulatory compliance (mobility behavior) both before and after region-specific legislative interventions, and the global announcement of the pandemic. Our random-effects analysis indicates that societies higher in openness experience less mobility decrease compared to the average, both pre and post the pandemic announcement, while regions higher in agreeableness showed a stronger decreased mobility compared to the average. When exploring specific reasons for personal movement changes such as work attendance, public transport, and residential isolation, our study reports somewhat uniform behaviors across all traits. More specific analysis of individuals’ willingness for domestic movement (internal travel) after the implementation of movement restrictions indicates that regions high in openness and extraversion appear to behave in a counter-compliant way. Societies with higher levels of agreeableness show already cooperative behavior, when recommendations are lighter enforcements are in place.

Our micro-level analysis in Study 2 reiterates our Study 1 findings; that is, higher extraversion is shown as a predictor of increased mobility, both through previous behavior and future intended movement. Interestingly, our age and gender analysis indicate that younger and older individuals (compared to middle-aged [30–60 years old] and females [compared to males]) are more likely to reduce mobility. Higher agreeableness in females (compared to males) results in an increased likelihood of staying home previously and an increased likelihood to stay home in the future. However, as these sex differences are only visible for those living in households with two or more people (not for single-person households), such findings may allude to a simpler explanation of division of labor. Such findings also mirror a body of scientific research findings, identifying women as less risk-seeking, more altruistic, and more cooperative or compliant (Andreoni & Vesterlund, 2001; Brañas-Garza et al., 2018; Croson & Gneezy, 2009; Hasseldine, 2002; Jianakoplos & Bernasek, 1998; Kastlunger et al., 2010; Seguino et al., 1996; Torgler & Valev, 2010; Vugt et al., 2007).

The current study is not without limitations. First, the Big Five personality traits, while shown to be globally present across cultures, are not a measure of the level of individuality or collectivism of a nation or group. As several traits directly impact a person’s perception of their own individuality, the relevance of the Big Five is arguably diminished in more collectivistic cultures. This may be because people in collectivistic countries may rely less on traits when understanding themselves and others, compared to those from more individualistic societies (Heine & Buchtel, 2009, p. 369).

Second, personality is not uniform inside any population. Arguably, and for future research, a country’s sense of nationalism may be more homogenous across a population and may be extremely influential on human behavior, particularly in developed economies. But again, this becomes scientifically problematic as both domestic and international perceptions of national character are more likely generalized stereotypes that serve as a function of nationalistic identity (Terraciano et al., 2005), which is why it is important to note that an abundance of symmetry in the personality profiles of a particular population may not equate to casual findings of any particular behavior, rather, just provide a reflection of similar cultural standards for a particular in-group (Heine & Buchtel, 2009).

Finally, it is important to acknowledge some differences in results between Studies 1 and 2, primarily that Study 1’s findings for openness are reversed for Study 2. And that, further, standardized β coefficients and effect sizes are small for both agreeableness and openness. Such differences may stem from the self-selected sample used in Study 2.

Regional-level personality trait analysis is an innovative way to understand human behavior in the context of large-scale, exogenous health shocks such as the current COVID-19 pandemic. And while the specific life-threatening health issue of the novel coronavirus may be overcome by medical science in the not too distant future (in the form of a vaccine), scientific research that seeks to understand how individuals and groups react to societal-level emergencies and the ensuing government regulatory responses is of critical importance. In fact, in the absence of such vaccinating silver bullets for this or future epidemics and pandemics, behavioral interventions may be humanity’s best and only option.

Supplemental Material

Supplemental Material, BIG_5_COVID19_24200723_supplementary_material_clean - Can Psychological Traits Explain Mobility Behavior During the COVID-19 Pandemic?

Supplemental Material, BIG_5_COVID19_24200723_supplementary_material_clean for Can Psychological Traits Explain Mobility Behavior During the COVID-19 Pandemic? by Ho Fai Chan, Jordan W. Moon, David A. Savage, Ahmed Skali, Benno Torgler and Stephen Whyte in Social Psychological and Personality Science

Footnotes

Authors’ Note

All the authors contributed to all parts of the analysis and write-up of this study.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Jordan Moon

Stephen Whyte

Supplemental Material

The supplemental material is available in the online version of the article.

Notes

References

Andreoni

Vesterlund

(2001). Which is the fair sex? Gender differences in altruism. Quarterly Journal of Economics, 116(1), 293–312.

Black

I. R.

Organ

G. C.

Morton

(2010). The effect of personality on response to sexual appeals. European Journal of Marketing, 44(9), 1453–1477.

Blatt

S. J.

Zuroff

D. C.

(1992). Interpersonal relatedness and self-definition: Two prototypes for depression. Clinical Psychology Review, 12, 527–562.

Brañas-Garza

Capraro

Rascon-Ramirez

(2018). Gender differences in altruism on Mechanical Turk: Expectations and actual behaviour. Economics Letters, 170, 19–23.

Chan

H. F.

Skali

Savage

Stadelmann

Torgler

(2020). Risk attitudes and human mobility during the COVID-19 pandemic (No. 2020-06) . Center for Research in Economics, Management and the Arts (CREMA).

Chung

M. C.

Dennis

Easthope

Werrett

Farmer

(2005). Multiple-indicator multiple-cause model for posttraumatic stress reactions: Personality, coping, and maladjustment. Psychosomatic Medicine, 67, 251–259.

Costa

P. T.

Jr McCrae

R. R.

(1995). Domains and facets: Hierarchical personality assessment using the revised NEO personality inventory. Journal of Personality Assessment, 64(1), 21–50.

Croson

Gneezy

(2009). Gender differences in preferences. Journal of Economic Literature, 47(2), 448–474.

Fetzer

Witte

Hensel

Jachimowicz

J. M.

Haushofer

Ivchenko

Caria

Reutskaja

Roth

Fiorin

Gomez

Kraft-Todd

Goetz

Yoeli

(2020). Global behaviors and perceptions in the COVID-19 pandemic. https://doi.org/10.31234/osf.io/3kfmh

10.

Franken

R. E.

Gibson

K. J.

Mohan

(1990). Sensation seeking and disclosure to close and casual friends. Personality and Individual Differences, 11(8), 829–832.

11.

Gomez

Holmberg

Bounds

Fullarton

Gomez

(1999). Neuroticism and extraversion as predictors of coping styles during early adolescence. Personality and Individual Differences, 27, 3–17.

12.

Gosling

S. D.

Rentfrow

P. J.

Swann

W. B.

Jr . (2003). A very brief measure of the Big-Five personality domains. Journal of Research in Personality, 37(6), 504–528.

13.

Google LLC. (2020, May 12). Google COVID-19 community mobility reports. https://www.google.com/covid19/mobility/

14.

Hasseldine

(2002). Gender differences in tax compliance. Taxation: Critical Perspectives on the World Economy, 3, 125.

15.

Heine

S. J.

Buchtel

E. E.

(2009). Personality: The universal and the culturally specific. Annual Review of Psychology, 60, 369–394.

16.

Ingram

R. E.

Price

J. M.

(2010). Understanding psychopathology: The role of vulnerability. In Ingram

R. E.

Price

J. M.

(Eds.), Vulnerability to psychopathology: Risk across the lifespan (2nd ed., pp. 3–17). Guilford Press.

17.

Jackson

J. C.

Gelfand

Fox

(2019). The loosening of American culture over 200 years is associated with a creativity-order trade-off. Nature Human Behaviour, 3, 244–250.

18.

Jianakoplos

N. A.

Bernasek

(1998). Are women more risk averse? Economic Inquiry, 36(4), 620–630.

19.

Kagel

McGee

(2014). Personality and cooperation in finitely repeated prisoner’s dilemma games. Economics Letters, 124(2), 274–277.

20.

Kastlunger

Dressler

S. G.

Kirchler

Mittone

Voracek

(2010). Sex differences in tax compliance: Differentiating between demographic sex, gender-role orientation, and prenatal masculinization (2D: 4D). Journal of Economic Psychology, 31(4), 542–552.

21.

Kim

Schimmack

Oishi

Tsutsui

(2018). Extraversion and life satisfaction: A cross-cultural examination of student and nationally representative samples. Journal of Personality, 86(4), 604–618.

22.

Kopala-Sibley

D. C.

Kotov

Bromet

E. J.

Carlson

G. A.

Danzig

A. P.

Black

S. R.

Klein

D. N.

(2016). Personality diatheses and hurricane sandy: Effects on post-disaster depression. Psychological Medicine, 46, 865–875.

23.

Kravitz

D. J.

Mitroff

S. R.

Bauer

P. J.

(2020). Practicing Good Laboratory Hygiene, Even in a Pandemic. Psychological Science. https://doi.org/10.1177/0956797620920547

24.

McCrae

R. R.

Costa

P. T.

Jr . (1999). A five-factor theory of personality. In John

O. P.

Robins

R. W.

(Eds.), Handbook of personality: Theory and research (pp. 159–181). Guilford Press.

25.

Monroe

S. M.

Simons

A. D.

(1991). Diathesis–stress theories in the context of life stress research: Implications for the depressive disorders. Psychological Bulletin, 110, 406–425.

26.

Mortensen

C. R.

Becker

D. V.

Ackerman

J. M.

Neuberg

S. L.

Kenrick

D. T.

(2010). Infection breeds reticence: The effects of disease salience on self-perceptions of personality and behavioral avoidance tendencies. Psychological Science, 21(3), 440–447.

27.

Nettle

(2006). The evolution of personality variation in humans and other animals. American Psychologist, 61(6), 622.

28.

Savage

Chan

H. F.

Moy

Schaffner

Torgler

(2020). Personality and individual characteristics as indicators of lifetime climbing success among Everest mountaineers. Personality and Individual Differences, 162, 110044.

29.

Schaller

Murray

D. R.

(2008). Pathogens, personality, and culture: Disease prevalence predicts worldwide variability in sociosexuality, extraversion, and openness to experience. Journal of Personality and Social Psychology, 95(1), 212.

30.

Schaller

Park

J. H.

(2011). The behavioral immune system (and why it matters). Current Directions in Psychological Science, 20(2), 99–103.

31.

Schmitt

D. P.

Allik

McCrae

R. R.

Benet-Martínez

(2007). The geographic distribution of Big Five personality traits: Patterns and profiles of human self-description across 56 nations. Journal of Cross-Cultural Psychology, 38(2), 173–212.

32.

Seguino

Stevens

Lutz

(1996). Gender and cooperative behavior: Economic man rides alone. Feminist Economics, 2(1), 1–21.

33.

Terracciano

Abdel-Khalek

A. M.

Adam

Adamovová

Ahn

C.-k.

Ahn

H. -n.

Alansari

B. M.

Alcalay

Allik

Angleitner

Avia

M. D.

Ayearst

L. E.

Barbaranelli

Beer

Borg-Cunen

M. A.

Bratko

Brunner-Sciarra

Budzinski

Camart

McCrae

R. R.

(2005). National character does not reflect mean personality trait levels in 49 cultures. Science, 310(5745), 96–100.

34.

Thielmann

Spadaro

Balliet

(2020). Personality and prosocial behavior: A theoretical framework and meta-analysis. Psychological Bulletin, 146(1), 30–90.

35.

Torgler

Valev

N. T.

(2010). Gender and public attitudes toward corruption and tax evasion. Contemporary Economic Policy, 28(4), 554–568.

36.

United Nations, Department of Economic and Social Affairs, Population Division (2019). Database on Household Size and Composition 2019.

37.

Vugt

M. V.

Cremer

D. D.

Janssen

D. P.

(2007). Gender differences in cooperation and competition: The male-warrior hypothesis. Psychological Science, 18(1), 19–23.

38.

White

A. E.

Kenrick

D. T.

Y. J.

Mortensen

C. R.

Neuberg

S. L.

Cohen

A. B.

(2012). When nasty breeds nice: Threats of violence amplify agreeableness at national, individual, and situational levels. Journal of Personality and Social Psychology, 103(4), 622.

39.

Whyte

Brooks

R. C.

Chan

H. F.

Torgler

(2019). Do certain personality traits provide a mating market competitive advantage? Sex, offspring & the big 5. Personality and Individual Differences, 139, 158–169.

40.

Whyte

Savage

D. A.

Torgler

(2017). Online sperm donors: The impact of family, friends, personality and risk perception on behaviour. Reproductive Biomedicine Online, 35(6), 723–732.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.60 MB