We collected data online in the Persian language between early November 2021 and late January 2022, a time period coinciding with the peak of the COVID-19 pandemic in Iran. During this time frame, the government enforced regulations related to the pandemic, such as social distancing, compliance with hygiene rules, and vaccination. Individuals were invited to participate using a variety of sources, including social media platforms and social networks. Exclusion criteria included age under 18 years, intellectual disability, and residing abroad. During the time period, 2,194 people received the survey and 800 fully completed surveys were returned. Of these, 54 surveys were removed because the participant was under 18 years old and 6 surveys were removed because the participant resided outside Iran. The final sample included 740 adult participants aged 18–74 years [72% female; mean age = 33.34 years (SD = 11.31)]. Participants provided basic sociodemographic information and completed measures of personality traits, psychological responses, and behavioral responses, and reported if they had experienced a confirmed COVID-19 infection based on positive polymerase chain reaction (PCR) test results or a physician’s diagnosis. COVID-19 infection was assessed as a dichotomous variable- Yes or No.

Hypotheses were tested with path modeling using Amos version 24.0 structural equation modeling (SEM) software package. The nonparametric asymptotically distribution-free (ADF) method was utilized to investigate model paths (Hancock and Mueller, 2013). Because of the sensitivity of the chi-square test to large samples, we followed convention and relied on multiple alternative fit indices to evaluate model fit (Hu and Bentler, 1999). The goodness of fit corresponds with chi-square fit statistics/degree of freedom (CMIN/DF) with values <5.0, goodness of fit index (GFI) with values >0.90, adjusted goodness of fit index (AGFI) with values >0.90, the comparative fit index (CFI) with values >0.90, incremental fit index (IFI) with values >0.90, and the root mean square error of approximation (RMSEA) with values <0.08 (Barrett, 2007). For all models, collinearity was examined by the variance inflation factor-VIF- (Vittinghoff et al., 2012) and by bivariate correlations. Heterogeneity in demographic characteristics including sex, history of medical disease and mental disorder, and COVID-19 vaccination was examined using independent samples t-tests.

All procedures followed were in accordance with the ethical standards of the Helsinki Declaration. The current project received ethical approval from the research ethics boards of school of medicine-Tehran University of Medical Sciences (Approval ID: IR.TUMS.MEDICINE.REC.1400.306). Informed consent was obtained from all participants.

We hypothesized that normative personality traits would be associated with COVID-19 infection and that these relationships would be mediated by psychological and behavioral responses to the pandemic situation. We tested all possible paths and associations (see supplemental Table S1). Fit indices suggested good model fit [Chi-square = 2.34; Degrees of Freedom (DF) = 5; CMIN/DF = 0.47; GFI = 0.99; AGFI = 0.99; CFI = 0.99; IFI = 0.98; RMSEA ~0.00]. Consistent with our hypothesis, all normative personality traits were significantly associated with COVID-19 infection via social distancing (SD) in the expected directions. Significant paths are presented in Figure 1. Social distancing (SD) was negatively associated with COVID-19 infection. There was a direct significant positive association between conscientiousness and SD. Openness and neuroticism had indirect positive associations with SD (through stress and anxiety), while extraversion and agreeableness had indirect negative associations with SD (through stress and anxiety).

In detail, conscientiousness was directly and positively associated with SD (β = 0.144; SE = 0.139; p < 0.001), and SD was negatively associated with COVID-19 infection (β = −0.167; SE = 0.003; p < 0.001). Openness was not associated directly with SD (β = −0.004; SE = 0.177; p = 0.930), but it was indirectly linked to it via a positive association with stress (β = 0.137; SE = 0.145; p < 0.001), and stress, in turn, was associated with anxiety (β = 0.754; SE = 0.022; p < 0.001); anxiety was associated with SD (β = 0.146; SE = 0.078; p = 0.013), that itself was negatively associated with the infection. Neuroticism also was not associated directly with SD (β = −0.021; SE = 0.138; p = 0.609), but it was indirectly linked to it via a positive association with stress (β = 0.404; SE = 0.104; p < 0.001), and then followed the same paths of stress, anxiety, SD, and infection. Extraversion was not associated directly with SD (β = −0.015; SE = 0.166; p = 0.752), but it was indirectly linked to it via a negative association with stress (β = −0.205; SE = 0.134; p < 0.001), and then followed the same paths of stress, anxiety, SD, and infection. Agreeableness also was not associated directly with SD (β = −0.007; SE = 0.150; p = 0.842), but it was indirectly linked to it via a negative association with stress (β = −0.066; SE = 0.124; p = 0.041), and then followed the same paths of stress, anxiety, SD, and infection. Finally, conscientiousness was directly and positively associated with compliance with hygiene rules (CHR; β = 0.138; SE = 0.072; p < 0.001). All other normal traits were indirectly associated with CHR via stress, which in turn was positively associated with anxiety (β = 0.754; SE = 0.022; p < 0.001). Anxiety was positively linked to CHR (β = 0.127; SE = 0.040; p = 0.029).

Figure 2 presents the significant effects of maladaptive personality traits on COVID-19 infection via social distancing. We tested all the paths and associations (see Supplementary Table S2). Fit indices suggested good fit [Chi-square = 7.04; Degrees of Freedom (DF) = 6; CMIN/DF = 1.17; GFI = 0.98; AGFI = 0.99; CFI = 0.99; IFI = 1.17; RMSEA = 0.01]. SD was negatively associated with COVID-19 infection. Detachment was directly positively associated with SD, which in turn had a significantly negative association with COVID-19 infection. In contrast, antagonism and disinhibition were directly and negatively associated with SD, which was negatively associated with infection. Additionally, negative affectivity and psychoticism had indirect positive associations with SD (through stress and anxiety).

In detail, detachment was directly positively associated with SD (β = 0.087; SE = 0.093; p = 0.047), and SD was negatively associated with the infection (β = −0.167; SE = 0.003; p < 0.001). Antagonism was directly negatively associated with SD (β = −0.115; SE = 0.110; p = 0.006), which in turn was negatively associated with the infection. Disinhibition was directly and negatively associated with SD (β = −0.127; SE = 0.103; p = 0.008), which itself was negatively associated with the infection. Negative affectivity was not associated directly with SD (β = −0.015; SE = 0.096; p = 0.759), but it was indirectly linked to it via a positive association with stress (β = 0.358; SE = 0.070; p < 0.001), and stress, in turn, was associated with anxiety (β = 0.666; SE = 0.023; p < 0.001); anxiety was positively associated with SD (β = 0.196; SE = 0.078; p < 0.001), which itself had a negative association with the infection (β = −0.167; SE = 0.003; p < 0.001). Psychoticism also was not associated directly with SD (β = 0.015; SE = 0.088; p = 0.718), but it was indirectly linked to it via a positive association with stress (β = 0.175; SE = 0.081; p < 0.001), and then followed the same paths of stress, anxiety, SD, and infection. Finally, Antagonism (β = −0.107; SE = 0.056; p = 0.010) and Disinhibition (β = −0.115; SE = 0.052; p = 0.013) were directly and negatively associated with CHR. Other maladaptive traits were indirectly linked with CHR via stress, and stress in turn was positively associated with anxiety (β = 0.666; SE = 0.023; p < 0.001) and depression (β = 0.777; SE = 0.022; p < 0.001). Anxiety was positively associated with CHR (β = 0.182; SE = 0.040; p = 0.002), while depression was negatively associated with CHR (β = −0.158; SE = 0.042; p = 0.050).

Figure 3 presents the significant effects of Dark Triad traits on COVID-19 infection via social distancing. We tested all the paths and associations (see Supplementary Table S3). The model had a good fit [Chi-square = 4.42; Degrees of Freedom (DF) = 3; CMIN/DF = 1.47, GFI = 0.99, AGFI = 0.98, CFI = 0.99, IFI = 0.99, RMSEA = 0.02]. SD was negatively associated with COVID-19 infection. Psychopathy had a direct negative association with SD, while other Dark Triad traits were indirectly associated with SD (through stress and anxiety).

In detail, psychopathy was directly and negatively associated with SD (β = −0.130; SE = 0.056; p = 0.003), and SD was negatively associated with COVID-19 infection (β = −0.167; SE = 0.003; p < 0.001). Narcissism was not associated directly with SD (β = 0.010; SE = 0.053; p = 0.795), but it was indirectly linked to it via a negative association with stress (β = −0.179; SE = 0.047; p < 0.001), and stress was positively linked to anxiety (β = 0.738; SE = 0.020; p < 0.001). Anxiety in turn had a positive association with SD (β = 0.187; SE = 0.077; p = 0.001), which was negatively associated with the infection (β = −0.167; SE = 0.003; p < 0.001). Machiavellianism also was not associated directly with SD (β = −0.058; SE = 0.050; p = 0.175), but it was indirectly linked to it via a positive association with stress (β = 0.192; SE = 0.045; p < 0.001), and then followed the same paths of stress, anxiety, SD, and infection. Finally, psychopathy was directly and negatively associated with CHR (β = −0.166; SE = 0.029; p < 0.001), while Machiavellianism and narcissism were indirectly associated with CHR via stress and anxiety; Machiavellianism had a positive association with stress, while narcissism was negatively associated with stress, which in turn was linked to anxiety (β = 0.738; SE = 0.020; p < 0.001). Anxiety was positively linked to CHR (β = 0.164; SE = 0.040; p = 0.005).

For normative personality traits, we found that openness, conscientiousness, and neuroticism were positively associated with social distancing, whereas extraversion and agreeableness were negatively associated. Subsequently, social distancing as a mediator in turn negatively predicted the COVID-19 infection. Recently, studies from the U.S., U.K., and Germany also examined personality traits as predictors of COVID-19 infection. Peters et al. (2023) indicated that in the early stages of COVID-19 in the US and Germany, the regional level of Openness acted as a risk factor, which is not aligned with our findings. This might be due to the regional level of assessment, different time periods in which data was gathered, or possible cultural differences. They also indicated that the regional level of Neuroticism acted as a protective factor, which is quite aligned with our results. The second study by Rolon et al. (2021) in the US and UK showed that extraversion predisposes people to become infected with the coronavirus, which is also in accordance with our findings.

According to the worldwide scale of the COVID-19 pandemic and the contagiousness of the virus, even small changes in people’s probability of social distancing behaviors such as shelter-in-place can substantially reduce the spread of COVID-19 both within and across countries (Dehning et al., 2020; Kissler et al., 2020). Our findings are wholly in line with conceptual definitions of the Big Five personality traits (DeYoung et al., 2007; Soto and John, 2017) and most prior research. A study conducted in 55 countries revealed that personality independently predicted protective behaviors including sheltering-in-place rates during the COVID-19 pandemic (Gotz et al., 2021). People high in Neuroticism are hyper-vigilant and experience anticipatory anxiety and threat sensitivity (Barlow et al., 2014), consistent with increased social distancing through increased stress and anxiety as we found here.

According to our study, openness positively predicted stress and anxiety, leading to increased social distancing. Although evidence has already shown that openness is related to risky behaviors (Schaller and Murray, 2008), it is also related to accurate risk perceptions- that could be associated with stress and anxiety (Fu and Wang, 2022)- through which openness is linked with increased social distancing behaviors (Trobst et al., 2000). Moreover, we found a direct positive association between conscientiousness and SD. Other studies similarly showed that more conscientious individuals are more likely to follow rules (John and Srivastava, 1999), are more cautious and socially responsible (Roberts et al., 2005), and more likely to consider others’ health (Roberts et al., 2009) which may lead them to comply with strict protective policies (Gotz et al., 2021). In our study, extraversion was a negative predictor of social distancing, which is consistent with its sub-factors of sociability and assertiveness (Soto and John, 2017), and previous research linking this trait to various risky health behaviors (Strickhouser et al., 2017) and decreased germ aversion (Gotz et al., 2021) that is also aligned with our results indicating Extraversion negatively predicted CHR. Further, agreeableness negatively predicted SD via negative association with stress and anxiety. The dimension of agreeableness is clearly represented in conceptions of dependency (Pincus, 2002). Thus, more agreeable people seek more frequent interpersonal contact (McCrae and Costa, 1989; Rollings et al., 2023)- although in contrast to extraverts, they would rather have smaller but more intimate social networks (Harada et al., 2023; Harari et al., 2020)- which can make it harder for them to comply with SD rules (Gotz et al., 2021).

In our study, maladaptive personality traits were negative predictors of social distancing and hygiene rules compliance behaviors, as well as positive predictors of stress, anxiety, and depression. Maladaptive personality traits are positively associated with emotional dysregulation, which in turn is related to symptoms of internalizing disorders such as depression and anxiety (Gratz et al., 2016). In detail, negative affectivity positively predicted SD via increased stress and anxiety, which was consistent with previous studies indicating the association between negative affectivity and SD (Srivastava and Coolidge, 2021) and indicating the relationship between negative affectivity and internalizing psychopathology, such as anxiety (Anderson et al., 2018). Negative affectivity also tends to associate with cold-submissive interpersonal problems, e.g., social avoidance (Wright et al., 2012), which is consistent with increased social distancing in our study.

We found that detachment, in contrast to extraversion positively predicted SD, which is aligned with its features including withdrawal from interpersonal interactions (American Psychiatric Association, 2013), and with prior research during the COVID-19 pandemic (Srivastava and Coolidge, 2021). Consistent with research by Somma et al. (2020) conducted during the COVID-19 pandemic, our findings showed that detachment was associated with higher levels of stress, anxiety, and depression. Antagonism negatively predicted SD and CHR in our sample, which replicates previous findings showing that antagonistic people were less motivated to employ coping strategies, such as SD, during the COVID-19 pandemic (Sica et al., 2021a).

We found that disinhibition, in contrast to conscientiousness, was negatively associated with SD, which is consistent with its features including irresponsibility (American Psychiatric Association, 2013) and prior research showing that disinhibition was negatively related to perceptions of the COVID-19 threat (Sica et al., 2021a), which could lead to decreased motivation to engage in protective behaviors. Prior research also supports our results and indicated a positive direct link between disinhibition and stress (Sica et al., 2021a). Although psychoticism did not directly predict SD, it was linked to SD through psychological responses including stress and anxiety. Psychoticism is associated with feeling disconnected from the real world (Holden et al., 2015) and an increased perception of daily hassles (Compton et al., 2008), which may increase social distancing behaviors.

Prior research found that Dark Triad personality traits were strongly associated with a decreased willingness to comply (Starcevic and Janca, 2022) and non-compliance (Blagov, 2021) with recommendations to limit COVID-19 spread mainly due to resistance to accept personal restrictions. This is aligned with our results in which dark traits negatively predicted SD and CHR. Machiavellianism- which has a strong correlation with low conscientiousness- correlates positively with depression and anxiety symptoms (Jonason et al., 2015), which is consistent with our results. Low sensitivity to threat, affective-interpersonal features of psychopathy, rule-breaking, and disregard for others seems to be associated with decreased engagement in protective and policy-compliant behaviors during the pandemic (Harper et al., 2021; Patrick et al., 2009). Our findings linking psychopathy to anxiety and depression are also consistent with numerous prior investigations (Jonason et al., 2015). We also found that narcissism was negatively associated with stress thereby negatively associated with SD. Similarly, several studies showed that narcissistic people did not experience excessive stress, anxiety, and depression during the COVID-19 pandemic and resisted engaging in protective behaviors (Gogola et al., 2021; Hatemi and Fazekas, 2022), even endorsing conspiracy theories regarding the COVID-19 virus and its vaccines (Hughes and Machan, 2021).

Since personality traits capture individual differences related to social and compliance behaviors through which the virus is transmitted, they can help explain differential transmission of COVID-19, even after controlling for important sociodemographic, economic, and pandemic-related factors (Peters et al., 2020). However, as all countries have suffered the impacts of the COVID-19 pandemic, regional and cultural differences should also be considered. Different countries have diverse cultures that could influence personality and social behaviors, and this could guide government policies encountering public health events, such as COVID-19 pandemic. For example, Gelfand et al. (2021) argued that the US, Canada, and European countries are loose cultures, whereas Asian cultures, as tight cultures, have imposed strict measures and punishments for deviance. Recent research found that compared with nations with high levels of cultural tightness, nations with high levels of cultural looseness had more confirmed COVID-19 cases and deaths (Gelfand et al., 2021). Moreover, cultural collectivism correlates positively with cultural tightness (Gelfand et al., 2011). Recent studies indicated that people from individualistic and collectivistic countries follow governments’ preventive measures (e.g., lockdown, social distancing, using a face mask, etc.) differently to contain the transmission of COVID-19 (Chen and Biswas, 2022; Lee et al., 2021). The studies showed hygiene behaviors such as using masks were higher in more collectivistic countries during the COVID-19 pandemic (Liu et al., 2023). In collectivistic cultures, the importance of “We” surpasses that of “I.” Compared to individual interest, common interest is viewed as ‘in-group,’ rooted in the tightly integrated relationships among families and close friends. As a collectivistic culture, Iranians might adhere to COVID-19 protective behaviors and health guidelines more in the face of the pandemic as a new environment, compared to other countries with individualistic cultures. Consequently, further related studies in different nations, like our study in Iran, can inform public health officials about the important role of individual differences in personality in mitigation behaviors which is critical to hinder infection and also is an influence that is not simply minimized by governmental policy. Public health policymakers can benefit from personality-tailored interventions for maximizing preventive health behaviors (Abdullahi et al., 2020; Boersma et al., 2011; Hagger, 2025; Hirsh et al., 2012; Juchem et al., 2024) and slowing the spread of infection. This knowledge also could contribute to more effective measures to prepare for public health emergencies in the future.