Introduction

The world is facing an extraordinary crisis. The novel coronavirus that emerged and started to spread at the end of 2019 is now reported in nearly every country and territory in the world [1]. The response to control the spread has been equally extraordinary. During the acute phase of the pandemic, entire countries were put on lock down to slow the spread [2]. The United States restricted international travel [3] and most states issued varying degrees of “stay at home” orders [4]. Such measures are essential to slow the spread of the virus [5]. The concerns over the virus and the stress associated with the social restrictions may have acute psychological consequences [6].

Here we test whether there are acute changes in Five Factor Model (FFM; [7]) personality traits and facets in response to the emerging coronavirus pandemic. FFM traits are stable individual differences that tend to be resistant to normative stressful life events [8]. Although typically stable, evidence from the psychopathology literature [9] and intervention research [10] indicates that traits can and do change in response to distress and treatment for distress, respectively. The coronavirus outbreak and measures to control its spread have disrupted most aspects of life, including basic motives (e.g., relationships, work) and daily activities that have been fundamental to work on adult personality development [11].

Using a sample that was first assessed in late January and early February 2020 and then again in mid-March 2020 during the President’s 15 Days to Slow the Spread guidelines [12], we tested for acute personality change in response to the coronavirus pandemic. Under normal circumstances, there is no reason to expect that personality would change over such a short period of time. Given the extraordinary nature of the coronavirus pandemic, and the drastic measures that have been taken to control its spread, however, personality may be reactive to these rapidly changing events. We tested the preregistered hypothesis that Neuroticism, particularly the anxiety facet of this trait, increased between pre- and post-test because the collective worry and anxiety over the virus would increase a trait tendency toward worry and anxiety. We also hypothesized that Conscientiousness, particularly facets related to responsibility, would increase between pre- and post-test because the public health messaging on the importance of personal responsibility to control the spread (e.g., handwashing, social distancing) would consolidate into a greater tendency toward rule-following and responsibility (i.e., conscientiousness), particularly in relation to others. We did not make directional hypotheses for the other three traits. Exploratory analyses examined whether there were larger changes in adults older than 65, males, and those in isolation/quarantine status because of the greater risk in these groups.

Materials and methods

Results

Descriptive statistics are shown in Table 1. The test-retest correlation was high for all five traits (≥.80), supporting the reliability of the measure and data quality. We report the results by personality domain, starting with the two domains that were hypothesized to change. Note that the results of the preregistered analyses are presented in Table 2, the results of the preregistered exploratory analyses for age and gender are presented in S2 and S3 Tables, respectively, and the results of the non-preregistered exploratory analysis for isolation are presented in Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographic characteristics of the sample.

https://doi.org/10.1371/journal.pone.0237056.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Mean change in personality traits between pretest and posttest.

https://doi.org/10.1371/journal.pone.0237056.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Interaction between time and isolation on change in personality traits.

https://doi.org/10.1371/journal.pone.0237056.t003

Discussion

The present research suggests modest acute personality change during the initial stages of the coronavirus outbreak in the United States. Contrary to our hypothesis, there was a small decline in Neuroticism rather than the expected increase. This change in Neuroticism was only apparent among individuals who were not in quarantine/isolation. We likewise did not find the expected increase in Conscientiousness, and there was some evidence that the current social environment may have changed the meaning of an item. In exploratory analyses, there was modest evidence that isolation status moderated trait changes in Conscientiousness, Openness, and Agreeableness, as well as for Neuroticism.

Personality traits tend to be stable over time and resistant to normative life events that are stressful [8]. Of the five traits, there is the most evidence that Neuroticism may be the most reactive to stress. When individuals experience a great amount of distress, either through an extremely aversive event [17–19] or a depressive episode [9], Neuroticism tends to increase. A similar but weaker trend is found for long-term psychological responses to natural disasters, such as after the Christchurch Earthquake [20]. Likewise, interventions to improve mental health decrease Neuroticism [10]. Given the stress and anxiety over the coronavirus, we had expected Neuroticism to increase. Instead, the opposite pattern emerged. This decrease may be due to contrast effects. That is, reminders of the collective stress and anxiety that the world was under were everywhere: During the 10 days of the posttest data collection, there was significant volatility and losses in the stock market [21] (marker of economic anxiety), essential household products such as toilet paper were sold out across the country [22] (marker of consumer anxiety), and national polls indicated that 70% of American adults were concerned or very concerned about the virus in their community [23] (marker of individual anxiety). Feelings of personal stress and anxiety may be attributed less to the self when there is a tremendous amount of stress and anxiety experienced through the whole of society. In such a context, there might be an attenuated tendency to perceive and rate oneself as emotionally distressed as compared to other people. The stress and anxiety participants felt may have been ascribed to the external situation rather than their own personality. It is important to note that participants with pretest data but no posttest data scored higher in Neuroticism. This difference in attrition may have had an effect on the pattern of results. For example, as individuals higher in neuroticism were lost to follow-up, it is possible that this more emotionally vulnerable group responded differently to the pandemic. It is also of note, however, that the overall pattern that we found is consistent with anecdotal reports of decreases in anxiety among individuals who typically suffer from anxiety [24].

We did not find evidence for change in Conscientiousness. We hypothesized that the ubiquitous public health messaging to be more attentive to personal behavior would translate into an overall increase in a trait tendency to be conscientious, particularly the facet of Responsibility. Rather than Responsibility, however, we found only modest evidence for an increase in the facet of Productiveness, which indicated that individuals saw themselves as more efficient and persistent in this crisis. There was, however, a fascinating pattern for Dutifulness. Dutifulness measures the tendency to adhere strictly to ethical principles [15]. This trait tendency decreased between pre- and post-test, a change that primarily occurred in participants younger than 65 (i.e., working-aged adults). This decrease was due entirely to declines on one item about going to work/school when not feeling well. In pre-pandemic times, this item was a fairly good marker of an individual’s willingness to follow through on their commitments. The swift changes in the social landscape, however, may have changed the meaning of this item. Now, rather than a marker of conscientiousness, going to work/school while sick may be a marker of recklessness or antagonism, whereas staying at home and protecting one’s community is conscientious. It is an example of how social context can (rapidly) change the meaning of an item and how it defines the trait it measures.

Approximately one-quarter of our sample reported being in isolation/quarantine within the last month. Our exploratory analysis suggested modest change in personality by isolation status. Of most note, isolation status moderated change in Neuroticism such that the decline in Neuroticism only occurred for those not in quarantine. Further, there was a cross-over interaction for the Depression facet: Individuals not in quarantine declined, whereas those in quarantine increased in a trait tendency toward depressed affect. Increases in depressed affect and other aspects of negative emotionality are common while in quarantine, and the effects may or may not be long lasting [25]. More generally, quarantine might provoke anxiety that is not assuaged by the stress and anxiety felt by the rest of the population. In addition to Neuroticism, isolation also moderated change in Openness, Agreeableness, and Conscientiousness. In all cases, these traits declined among individuals in isolation, specifically the facets of Curiosity, Trust, and Organization, respectively. The circumstances around isolation may lead to boredom and erode trust. There may also be less pressure to be organized because there is less that needs to get done in a timely manner. It is also of note that there were baseline differences in personality prior to quarantine. That is, individuals who go into quarantine had higher baseline levels of Neuroticism and lower Agreeableness and Conscientiousness. Individuals with these traits may be at greater risk of exposure through either who they interact with and/or they work jobs that put them at higher risk of exposure. Individuals higher in Neuroticism may also perceive more threat and go into quarantine to feel safer. There may also be bias associated with these traits in how quarantine/isolation is interpreted (e.g., safer at home may be interpreted as quarantine). We could not tease apart these different possibilities.

FFM personality traits are known to be stable [26] with normative changes across the lifespan [27] and are also known to be relatively resistant to change after normative life events [8]. As such, there would be no expectation that personality traits would change over just six weeks in normal circumstances. The coronavirus pandemic, however, is unprecedented in its disruption of daily life for most of the population. It was thus possible that it would also have an unprecedented effect on personality. As described above, extremely aversive and stressful events are associated with change in personality [17, 19], and the global scale of the current stressful event may have had a widely felt impact. And yet, even with the widespread fears over health consequences of complications of COVID-19, the economic uncertainty, and restrictions on daily life, personality traits have been mostly resistant to change. These findings support theoretical accounts of personality traits that argue for their stability [28], even in the face of acute environmental stressors. It may be the case that other aspects of psychological functioning, including state affect or mental health [29], may be more vulnerable to the impact of COVID-19 (but see [30]).

The present study had several strengths, including a pre-post design that captured trait psychological function just prior and during the acute phase of the coronavirus pandemic in the United States. The findings, however, need to be put in context. Although there was evidence of change, for example, the magnitude of change was small; in most cases, the change was less than one-tenth of a standard deviation. As such, overall there is more evidence of stability than substantial change. Still, personality would not be expected to change at all over such a short period of time in normal circumstances. The findings also need to be put in context of some limitations. First, the attrition analysis indicated that there were significant selection effects for who remained in the sample at Time 2 that may have had an effect on the results, particularly for Neuroticism (as discussed above). It is important to note that this study was not originally designed to be longitudinal, so participants in the pretest survey did not know that they would be asked to complete a second survey. With the pandemic, the study was reconceptualized to take advantage of the data collected on psychological functioning just prior to the pandemic. Fortunately, many participants were willing to fill out a second survey, but given that the original study was not meant to be longitudinal, there was no expectation that participants would continue to participate. Second, we tested for trait change in the acute phase of the pandemic. Although the purpose of this measurement was to address whether trait psychological functioning was responsive to an acute health-threatening crisis, it is also possible that the effects of the crisis could take longer to consolidate into substantial changes in personality. Future work will need to address whether there are long-term changes in personality in response to the coronavirus pandemic. Future work also needs to address personality change during the pandemic in other cultural contexts. Third, our measure of quarantine/isolation was broad and did not differentiate between quarantine or isolation and the situation for the participant during quarantine/isolation (e.g., whether the person was alone or with another person). As such, we could not disentangle the exact circumstance of the quarantine/isolation and whether such differences are important for personality change. Finally, as with all non-experimental research, there may be other explanations for the current set of results that we cannot rule out.

Overall, the results suggest more trait psychological resilience than harm during the acute phase of the coronavirus spread and response in the United States. Consistent with the notion that traits are stable and resistant to change, there were few changes in response to the spread of the coronavirus and the measures to control the spread in the United States. The results further suggest that the broader social environment may be modifying both how individuals see themselves (e.g., attributing less anxiety and depressed affect to themselves) and the meaning of specific items to how they measure a trait (e.g., items of Dutifulness). Future work will need to address whether these modest changes are long lasting and/or whether different patterns of change emerge if this crisis is protracted.

Supporting information