Effects of age, psychological distress, and compassion on people’s decision-making during the COVID-19 pandemic

Joel Myerson¹², Michael J Strube¹, Leonard Green¹, Sandra Hale¹, Bridget Bernstein¹

1. Washington University in St. Louis, Saint Louis, USA
2. Department of Psychological & Brain Sciences, Washington University in St. Louis, Saint Louis, USA 63130
   [email protected]

OPEN ACCESS

PUBLISHED: 30 April 2025

CITATION: Myerson, J., et al., 2025. Effects of age, psychological distress, and compassion on people’s decision making during the COVID-19 pandemic. Medical Research Archives, [online] 13(4).
https://doi.org/10.18103/mra.v13i5.6479

COPYRIGHT: © 2025 European Society of Medicine. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

DOI https://doi.org/10.18103/mra.v13i5.6479

ISSN 2375-1924

ABSTRACT

If one thinks of the COVID-19 pandemic as a natural experiment, the most powerful psychological variable that differed in this natural experiment is age. Notably, however, adolescents and young adults appeared to be the most affected in this regard. Researchers were naturally quick to blame the pandemic for elevated distress levels based on studies that compared current distress with pre-pandemic levels, but they neglected the possibility that there is a secular trend that also affected younger adults more than older ones.

Additionally, national surveys in the U.S., Australia, and the U.K. by Hale and colleagues revealed that particular differences in distress levels were observed in all three countries over the decade preceding the pandemic, and the increases were age-related. This paper examines the roles other personality traits might play in decision-making during the pandemic, specifically focusing on the effects of age, psychological distress, and compassion on people’s decision-making.

Keywords

COVID-19, psychological distress, age, compassion, decision-making

Introduction

One of the first things that psychologists noted at the beginning of the COVID-19 pandemic was that psychological distress was reaching historically high levels¹, and that adolescents and young adults appeared to be the most affected in this regard. Researchers were naturally quick to blame the pandemic for elevated distress levels based on studies that compared current distress with pre-pandemic levels, but they neglected the possibility that the increase in distress levels was part of a secular trend that also affected younger adults more than older ones².

Indeed, subsequent analyses of national surveys in the U.S., Australia, and the U.K. by Hale and colleagues³ revealed that parallel increases occurred in all three countries over the decade preceding the COVID pandemic, and the increases were age-graded in all three countries: Young adults’ distress increased the most, middle-aged adults somewhat less so, with older adults showing the smallest increases. Analyses also revealed clear effects of lockdowns and quarantines, but the contributions to national or local distress levels wore off quickly when social isolation was no longer mandated. Notably, however, distress tended to return to the elevated levels predicted by the secular trends, rather than to the levels observed previously.

When Myerson and colleagues examined the determinants of psychological distress and pandemic mitigation behaviors at the individual level, they found that whereas adults’ distress levels decreased with age, the opposite pattern was observed in their mitigation behaviors⁴. Social distancing, for example, tended to increase with age. In fact, the inverse relation between distress and the tendency to engage in mitigation behaviors was observed in adults of all ages. Importantly, concern for oneself was not significantly correlated with mitigation, but concern for the effects of the pandemic on others was a significant predictor of decreases in social contact as well as decreases in actual physical contact with others.

Although we already suspected that differences in personality traits like Neuroticism might play a role in pandemic decisions (such as whether or not to engage in social distancing and/or get vaccinated), this finding inspired us to examine facets of Agreeableness that might be reflected in pandemic-related behaviors that could benefit others. In short, although one might expect the pandemic to cause people to focus on whether behaviors would increase or decrease the risks to themselves, the data suggested that whether one focused on the consequences for oneself or the consequences for others actually depended on one’s age⁵.

---

PANDEMIC DECISIONS fROM  AN  ETHICAL PERSPECTIVE

Many choices that people had to make during the COVID-19 pandemic (e.g., whether or not to engage in social distancing and whether or not to get vaccinated against the SARS-CoV-19 virus) may be viewed as fundamentally ethical ones; that is, they challenged people not to violate the negative injunction implicit in what has been termed the Golden Rule: Do not do unto others in ways that you would not want them to do unto you. More specifically, in the present case, the injunction is simply ‘do not infect others or put them at risk of infection.’

Of course, the kinds of behavioral choices just described may also benefit the individual making the decisions (e.g., by decreasing their own risk of infection), but it is an empirical question to what degree people’s choices in these matters were fundamentally self-serving or whether they were driven, at least in part, by their concern for others. Accordingly, the current study assessed the degree to which participants’ responses to survey items were consistent with personality traits like those reflected in specific facets of Agreeableness (i.e., Altruism, Sympathy, and Trust) from the IPIP-NEO (International Personality Item Pool) personality test as well as their responses to items from the Neuroticism (or Emotional Stability) scale of the IPIP-NEO. A secondary goal was to assess possible age differences in the contribution of altruism to pandemic decision-making. It should

be noted that while altruism is operationally defined here in terms of items from the Agreeableness scale of the IPIP-NEO, the conceptual definition implicit in our discussion is more closely related to that of reciprocal altruism in evolutionary biology⁶⁷ than it is to the selfless decisions and actions in traditional ethics.

To understand the literature on decision-making during the COVID-19 pandemic, it is useful to divide that time into at least three periods: Period 1, before the pandemic; Period 2, from the onset of the pandemic until vaccines became available; Period 3, from the onset of vaccine availability in December 2020 until May 2023 when the Public Health Emergency ended in the U.S.

PERIOD 1.
To understand the psychological effects of the Covid-19 pandemic, one needs a baseline against which to compare them, and researchers turned first to the period immediately preceding the onset of the pandemic. They found that distress had increased in individuals ranging in age from adolescents to the elderly, an increase they naturally attributed to the pandemic, both its dangers and its economic consequences¹. As already noted, however, most researchers neglected the possibility that the increase in distress levels partly reflected an already increasing secular trend². However, analyses of periodic national surveys in the U.S., Australia, and the U.K. revealed parallel age-graded increases in distress in all three countries over the decade preceding the pandemic³⁸⁹. Although comparisons with other, less economically advanced and/or non-English-speaking nations would have made additional useful comparisons possible, data from the decade prior to 2019–2020 provide clear evidence of an increasing, international secular trend in psychological distress.

PERIOD 2.
The data from Period 2 provided clear evidence that although social deprivation was largely responsible for elevated distress levels during the pandemic, that elevation was temporary¹⁰. For example, although anxiety levels in the U.S. had more than doubled from 2019 levels by April, 2020 when stay-at-home orders were in effect in most states, anxiety levels immediately dropped when most states ended restrictions in May¹¹. Data from two sequential lockdown periods in the UK provided further evidence for the role of social isolation³. When the first lockdown period ended, distress levels rapidly returned to the level predicted by the secular trend observed in Period 1, rather than to the lower levels observed previously.

PERIOD 3.
Even after vaccines against SARS-CoV-19 became available, many people continued to engage in social distancing and to avoid enclosed public spaces, and these mitigation behaviors were positively correlated with each other but negatively correlated with psychological distress¹². In contrast, vaccination was positively correlated with distress. To explain the opposite effects of distress on mitigation and vaccination, Myerson et al. suggested that if social isolation led to distress, then distress might interfere with many mitigation behaviors (e.g., distancing, facemasks) because they might exacerbate isolation; on the other hand, distress might motivate vaccination because it would allow a return to normal levels of social interaction. Consistent with our hypothesis that social deprivation was responsible for much of the distress reported during the pandemic, Loneliness and self-reported chance of infection and/or hospitalization another pandemic stressor) accounted for nearly half of the variance in individual differences in distress.

Digging deeper, we proposed a form of diathesis-stress model in which Neuroticism captured individual differences in vulnerability to pandemic stressors¹³. More specifically, we hypothesized that Emotional Stability (the flip side of Neuroticism) would predict psychological distress during the pandemic because it measures both reactivity to negative events and how prone a person is to negative emotional states (e.g., anxiety, depression)¹³. Supporting evidence for the role of Emotional Stability in pandemic-

related distress comes from a study by Zacher and Rudolph¹⁰, who found that greater emotional stability (i.e., a lower level of Neuroticism) was associated with lower perceived stressfulness during the lockdown in Germany in 2020. When Emotional Stability was added to the two significant variables from our previous distress model (i.e., Loneliness and Chance of infection), the three measures together accounted for nearly two-thirds (64.9%) of the variance in psychological distress. Importantly, Age failed to account for any unique variance when added to our three-measure model, consistent with the hypothesis that age-related differences in Stability/Neuroticism underlay our previous finding of a significant association between Age and Distress.

Our findings regarding Emotional Stability led us to examine the roles other personality traits might play in pandemic decision-making. Recalling our early finding that concern for the effects of the COVID-19 pandemic on others was associated with mitigation behaviors whereas concern for its effects on oneself was not, we proceeded to examine several related facets of the Agreeableness component of the Big Five personality model¹⁴. Specifically, we assessed the contributions of Altruism, Sympathy, and Trust to pandemic decision-making, and the roles they played in the pronounced age-related differences we had observed in mitigation behaviors. Interestingly, Altruism and Sympathy were so strongly correlated (r = .768) that we combined them into a single variable (Compassion), and as we had suspected, Compassion proved to be a significant predictor of measures of social distancing, both decreases in physically close interactions and decreases in visits to enclosed public spaces. In fact, when Compassion was included in a logistic regression model, Distress was no longer a significant predictor. Trust, on the other hand, did not significantly predict distancing, although it did predict Vaccination intentions whereas Compassion did not.

---

Objectives

Because Compassion and Neuroticism appeared to play opposite roles in pandemic mitigation decisions, and perhaps in ethical decisions more generally, we pitted them against each other in the current study. In order to better understand their contributions to pandemic decisions, we tested models of the associations of both Compassion and Psychological Distress with various other individual characteristics. These models also assessed the growing effects of political attitudes on pandemic decision-making¹⁵ and specifically examined their different roles in mitigation and vaccination decisions. In addition, we hypothesized that a participant’s community influenced their pandemic decisions. To test this hypothesis, we asked participants how many of their acquaintances had been vaccinated against COVID-19 as well as how many had been hospitalized because of a COVID-19 infection. Of interest was not just whether these measures would predict pandemic decisions such as whether or not to engage in social distancing or get vaccinated, but which measures would predict which specific decisions.

---

Methods

Participants

Volunteers were recruited online from March 31 to April 7, 2022. After indicating their consent online, which initiated the survey, 752 MTurk workers provided their responses and received $2.00 for their participation, which took, on average, 24.7 minutes. The submitted surveys were then screened for age, participation status in our previous vaccination study, valid IP addresses associated with internet providers in the United States of America, and survey completion time so as to exclude those whose times were less than the time a fast, expert reader would require to read the survey questions¹⁶. Based on these criteria, data from 210 individuals were excluded from our analyses, 5 based on invalid US IP addresses, 126 based on their previous participation in our preceding vaccination study, and 79 based on their completion times, leaving 542 participants ranging in age from 18 to 80 years: 252 females and 287 males, plus 3 participants who did not report their gender. These participants’ racial

ethnic breakdown was 89.7% White, 4.6% Black, 3.5% Asian, and 2.2% other races; 22.3% identified as Hispanic/Latinx.

The average age of participants in the sample was 40.14 years (SD = 14.18). The majority (55.19%) reported household incomes between $30K and $66K; relatively few reported incomes less than $30K (13.3%) or more than $120K (18.6%). Finally, the self-reported political ideology of the sample tended to be slightly conservative: More than half of the participants (53.3%) described themselves as conservative or very conservative, whereas about a quarter (26.0%) described themselves as liberal or very liberal. The study was approved by the Institutional Review Board of Washington University in St. Louis.

---

Table 1. Participant Characteristics as a Function of Age Group.

Note. HS and Uni indicate completion of high school and college or university, respectively.

---

PROCEDURE

The online survey consisted of four parts, beginning with questions about the frequency of three CDC-recommended mitigation behaviors and two mask-related behaviors. Specifically, the three mitigation behaviors included two Social Distancing behaviors and one Hygiene behavior: (1) being less than six feet from a person who was not a member of one’s household; (2) being in an enclosed public space with other individuals; and (3) cleaning one’s hands with either sanitizer or soap and water. We term these measures Proximity (Prox), Public Space (Public), and Hand Hygiene (Hand), respectively. For each mitigation behavior, participants were asked about its frequency in two separate time frames: on average this week and before the pandemic began. The two mask-related behaviors were (1) being less than six feet from a person who was not a member of one’s household while wearing a mask, and (2) being in an enclosed public space with other individuals while wearing a mask, although the mask data turned out to be confounded by mask mandates and were not analyzed further.

The second part of the survey consisted of the items from the Hospital Anxiety and Depression Scale (HADS¹⁷), questions as to participants’ degree of concern about the possible effects of the pandemic on themselves and others in their community, and questions regarding participants’ current levels of loneliness and social deprivation¹⁹. Participants then were asked questions regarding personal connections with COVID-19 cases (e.g., number of acquaintances hospitalized with COVID-19 infections) and subjective opinions about vaccines against COVID-19 (e.g., Do you believe vaccination against COVID-19 is safe?).

The third part of the survey asked participants about their overall health and the frequency of their attendance at religious gatherings, followed by inquiries concerning current employment status and household income. Participants then were asked about their level of formal education, age, gender, ethnicity, and race, and the number of children and adults in the home. Next, participants were asked to answer items from the Kessler Screening Scale for Psychological Distress, although our analyses

focused on the HADS for consistency with our previous studies.

The fourth part of the survey began with questions regarding participants’ current vaccination status (What is your current vaccination status?: Not Vaccinated and unlikely to be, Not Vaccinated but likely to be, Partially Vaccinated, Fully Vaccinated but not Boosted, Fully Vaccinated and Boosted), whether they would get vaccinated if it was recommended to them, and indicating, from a prepared list, what they felt was most important to them regarding COVID-19 vaccination. Participants were also asked to select up to three statements regarding their feelings about COVID-19 vaccination and life before the COVID-19 pandemic. Following this, participants were then asked to answer items from the IPIP-NEO personality test. Finally, participants were asked questions regarding their political ideology, who they voted for in the 2020 United States Presidential Election, and their home zip code.

---

ANALYSIS

To minimize multicollinearity, five pairs of conceptually related and significantly correlated measures (all p < .001) were converted to mean Z-scores. For example, standard scores representing decreases in the frequencies of close proximity encounters and visits to enclosed public spaces were averaged into a single Distancing composite because the correlation between these measures (rho = .445) was much larger than their correlations with hand hygiene (rho = –.091 and .014). The new composite measures were labelled Distancing (decreases in close proximity interactions and in visits to enclosed public places compared to pre-pandemic levels), Distress (Anxiety and Depression), Compassion (Altruism and Sympathy), VaccAtt (Vaccine attitudes regarding safety and efficacy), and Chance (Estimated chance of infection with COVID-19 and/or hospitalization).

---

Results

MITIGATION BEHAVIORS

Were people still engaging in mitigation behaviors after two years of the COVID-19 pandemic? Indeed, they were: The frequencies of people’s close proximity interactions and of visits to enclosed public spaces were both significantly less than before the pandemic, and the frequency of hand hygiene was significantly greater than before (Table 2).

---

Table 2. Signed ranks tests of changes in mitigation behaviors

Note. Decreases in Proximity and Public Space measure Social Distancing.

---

PREDICTING DISTANCING

Replicating our previous studies (e.g., Myerson et al., 2021), Distress was negatively correlated both with Age (r = –.212) and with Distancing (r = –.195), whereas Age was positively correlated with Distancing (r = .099). Our primary interest, of course, was in the contributions of psychological traits and states to pandemic decisions. Consistent with the findings of Hale, Myerson, and colleagues³⁵, analysis of the present data revealed significant correlations between Distancing, Compassion, and Neuroticism: Distress and Compassion (r = –.597); Distress and Neuroticism (r = .578); Compassion and Neuroticism (r = –.578). Again, all p < .001. Notably, all the signs remained the same and the correlations all remained significant with Age statistically controlled.

Previous studies⁴⁵¹² found that whereas Distancing increased significantly with Age, it decreased with Distress, and the present study replicated that finding. When Compassion was added to the model, however, it was the only significant predictor of

Distancing, presumably because it accounted for the same variance as the other independent variables as well as some unique variance in individual Distancing scores (Table 3).

---

Table 3. Regression model predicting Distancing.

Note. R² = .061. S.E. represents Standard Error of the Mean.

---

When additional variables were added to the model to assess the contributions of political ideology, social influences, and demographic factors (i.e., income, education, and religiosity, measured as the frequency of attending religious services), Compassion remained significant. It may be noted that the contributions of social influences were selective: The more acquaintances one had who had been hospitalized with COVID, the lower the level of distancing, although the number of one’s acquaintances who had been vaccinated was not significant.

---

PREDICTING DISTRESS

Distress is of interest, of course, for multiple reasons, including its obvious effects on quality of life and the roles it plays motivating and/or interfering with health behaviors, not to mention its well-established effects on immune function¹⁸. Accordingly, we tested a model originally proposed by Hale et al.³, which posited that Neuroticism captured individual differences in vulnerability to pandemic stressors. The results strongly supported the model, which explained much of the distress observed during the pandemic as being the result of two stressors, social deprivation and anxiety about possible infection with COVID-19. These stressors were instantiated in the model by two measures, Loneliness and self-reported estimates of the likelihood of infection and/or hospitalization (Chance), although it should be noted that whereas Hale et al. measured Emotional Stability, here we simply changed the sign and called it Neuroticism, a term which we believe better reflects the role it plays in psychological distress.

Hale et al.’s model³ accounted for 44.1% of the variance in distress reported by the present sample (Table 4A) and provided a basis for assessing the role played by individual differences in Compassion. When Compassion was added to the model (Table 4B), it significantly increased the variance accounted for to 50.2% and provided more evidence of the opposing roles played by Compassion and Neuroticism, similar to the opposing roles these trait measures play in predicting Distancing. Again, Age and Compassion appear to work together as they did with respect to distancing and reveal again that while advancing Age is associated with increased Compassion, Age also accounts for its own unique variance.

---

PREDICTING COMPASSION

While our previous research had focused on the contributions that Compassion might make to mitigation and vaccination decisions, as with Distress it seemed useful to turn the scope around and see what it was that contributed to Compassion. This approach represented an opportunity to validate our measure of Compassion as well as to assess the possible sources of individual differences in Compassion. This analysis provided further evidence of the opposing roles played by Age, on the one

hand, and Distress and Neuroticism, on the one hand. It should be noted that, as predicted, concern for possible effects of the pandemic on oneself (Self) and concern for its effects on others one judged close to oneself (Close) played what appear to be analogous opposing roles here (Table 5).

---

Table 4. Two regression models predicting Psychological Distress.

A. Model from Hale et al. (2023).

Note. R² = .441.

---

B. Hale et al. model with Compassion added.

Note. R² = .502.

---

Table 5. Regression model predicting Compassion.

Note. R² = .390.

---

Interestingly, when Education and Religiosity were added to the regression model, neither was significant. Indeed, correlational analyses revealed that while Religiosity was weakly correlated with Compassion, the correlation was negative – more frequent attendance at religious services was associated with lower scores on Compassion (r = –.113, p = .009). Moreover, neither education nor income was significantly correlated with Compassion, so that while Compassion increases with Age, our analyses have as yet yielded little insight into why it increases, although we have been able to assess some of the positive consequences of that increase.

---

PREDICTING VACCINATION

It was not until April 19, 2021, that all U.S. states had opened vaccine eligibility to residents aged 16 and older, and when the present sample was recruited approximately one year later almost all of them (92.6%) had been vaccinated. In fact, a little

over half (56.3%) reported having received a second, ‘booster’ shot as recommended by the CDC. As a result, we were able to assess what measures predicted who had chosen to be boosted and to compare and contrast the determinants of that decision with what public health researchers refer to as ‘vaccination intentions’ – self-reports of the likelihood of getting (re)vaccinated in the future.

Logistic regression was used to determine what measures were significantly associated with having received a Booster, beginning with a model that we previously used to predict vaccination¹². However, whereas all six predictors in the model had previously been significant, now only half of them were: Age, VaccAtt, and OtherHosp. Neither Distress, Chance, nor Religiosity were significant (Table 6). It should be noted, however, that not only were we now trying to predict decisions by a different sample made at a different point in the pandemic, but we now were also trying to predict who had been Boosted, whereas before we had been predicting first-time vaccinations.

---

Table 6. Logistic regression model predicting Boosted.

Note. X² = 102.351, df = 491, p < .001; McFadden R² = .150.

---

Of course, another thing that had changed was that vaccination had become increasingly politicized, and indeed, when Ideology was added to the model, it was significant, although the results had not changed in other respects. Notably, a more liberal ideology was associated with an increased likelihood of having received a booster vaccination, although the strongest predictor was VaccAtt, which assessed how positively participants felt about the safety and efficacy of COVID-19 vaccines. Also, those who had more acquaintances hospitalized with COVID-19, which we suspect meant that more of their acquaintances had not been vaccinated, were also less likely to have gotten a booster shot themselves. Finally, Trust was a significant predictor, whereas Compassion was not.

Would the same variables that predicted being Boosted predict whether people said that they would get revaccinated if it were recommended? A comparison of Tables 6 and 7 reveals multiple differences, perhaps most surprising (and perhaps even hopefully), Ideology, which had been significantly associated with being boosted, was not predictive of whether people were likely to be revaccinated in the future. In addition, Age was no longer significant when participants were predicting their future vaccination decisions. Another difference was that whereas Compassion was not a significant predictor of vaccination intentions when predicting Boosted, Compassion was a significant predictor of Likely, our measure of participants’ intentions regarding future revaccination.

Taken together, these differences highlight ways in which people actual vaccination decisions differed from decisions regarding hypothetical future vaccinations. Perhaps more importantly, however, the strongest predictor of vaccination in both cases was the same (i.e., VaccAtt, the extent to which people considered vaccination safe and effective). This finding, in turn, suggests that whether one got Boosted and whether one would get vaccinated yet again in the future should be relatively strongly correlated.

---

Table 7. Linear regression model predicting Likely.

Note. R² = .310.

---

Our final analyses focused on the correlations among our principal measures (Table 8). Regression models, like those presented above, identify sources of variance that are uniquely associated with the dependent variable after controlling for all the other variables in the model. In contrast, correlation measures address the question of associations between a pair of variables regardless of their relations to other measures. Moreover, correlation can be used to quantify associations between binary and continuous measures (e.g., whether one got a booster vaccination and one’s estimate of how likely it is one will get another if it is medically recommended). As predicted based on the importance of vaccine safety and efficacy for people’s decisions on both of our vaccination measures (Boosted and Likely), the correlation between them was relatively strong (r = .455). Perhaps the other finding that stands out is that our two vaccination measures were not significantly correlated with either Age, Distress, or Neuroticism, whereas both vaccination measures were positively correlated with Compassion, Trust, and having a more liberal Ideology, as well as with Distancing.

---

Discussion

One striking revelation from the data reported early in the pandemic was the finding that although distress levels were high overall, degree of distress decreased with age, revealing a perhaps unexpected resiliency in older adults²⁰. Notably, for some years prior to the pandemic, older adults had been reporting less distress than younger adults in a number of countries that conducted regular national surveys³. What was surprising was that during the pandemic, older adults were consistently reporting lower levels of distress despite the fact that they were probably aware early on that they were at much greater risk of serious illness and even death than younger adults²¹. To understand these findings and their implications, we began a series of studies of pandemic decision-making initially focused on just age, distress, and social distancing, and subsequently examined vaccination decisions as well⁴⁵¹².

Table 8. Correlations among personal characteristics, personality measures, and pandemic decision measures (distancing and vaccinations).

---

Importantly, our first study found that participants’ concerns regarding the effects of the pandemic on close acquaintances were significantly associated with social distancing, whereas their concerns for the effects of the pandemic on themselves were not. In addition, we were surprised to find that distress was repeatedly associated with lower levels of distancing, while age was associated with higher levels of distancing, a finding we subsequently discovered was linked to lower levels of Neuroticism in older adults³. That finding inspired us to take a closer look at personality measures and assess both Neuroticism and three of the most prosocial facets of Agreeableness from the NEO (i.e., Trust plus Altruism and Sympathy, which we combined into a single measure of Compassion because of their strong correlation).

Accordingly, the present investigation directly compared the relations among Compassion, Neuroticism, and pandemic decisions. In a previous study⁵, we had tried adding Compassion to a model that predicted Distancing from Age and Distress and found that with Compassion in the model, it accounted for all the variance in Distancing previously attributable to the other two variables. For the present study, we added Neuroticism to the model as well and found that Compassion still accounted for all the variance otherwise attributable to the other measures.

Given the multicollinearity among our measures, more detailed analyses of individual characteristics associated with both Distress and Compassion were called for. Previously, we had hypothesized that Neuroticism assessed one’s vulnerability to stressors, and posited that it modulated their effects. Accordingly, we modelled the Distress observed during the COVID-19 pandemic as the combined effects of Age, Loneliness, Chance, and Neuroticism, all of which were significant (R² = .441). When Compassion was added to the model (R² = .502), it, too, was significant, indicating that each of these variables accounted for a significant proportion of unique variance. Importantly, it suggested further that when Distress was used to predict other measures (e.g., Distancing), Distress likely represented effects of all five of these measures.

A more detailed analysis of Compassion revealed that it, too, was associated with Distress, Age, and Neuroticism. In addition, this analysis validated our Compassion measure: The association of Compassion with concerns regarding the effects of the pandemic on others with whom one was close was positive, whereas the relation to concerns regarding the effects of the pandemic on oneself was negative.

effects of the pandemic on oneself was negative. Taken together, these findings are consistent with the view that the age-related decline in Neuroticism is responsible, at least in part, for the observed increase in Compassion with age.

---

A cascade model of age-related changes in Neuroticism, Distress, and Compassion

The decline in Neuroticism as people get older is well established, as is the increase in the trait of Agreeableness¹⁴ on which our Compassion measure is based. Consistent with these findings, a meta-analysis found that older adults were significantly more altruistic than younger adults, although the authors suggested that the level of altruism was limited by one’s resources, including health, finances, and social networks, as well as cognitive resources²²²³. Taken together with the present results, these findings suggest that distress leaves people with less resources for Compassion. Thus, the causal chain may begin with higher levels of Neuroticism in younger adults, which leads to greater vulnerability to stressors, and thus to greater psychological distress¹³ that, in turn, leaves one with fewer resources needed for compassionate inclinations and decisions. Taken together, this chain leads to the observed pattern of increases and decreases in various psychological measures (Neuroticism, Distress, and Compassion) and their covariation with Age. It should be noted, however, that the correlations among these measures remain strong (all r > .500) when Age is statistically controlled, suggesting that although Age affects all three measures, the mechanisms underlying the interrelations among them are quite general and not age-specific.

As for why Neuroticism declines with age, initiating this causal chain, Costa and McCrae¹⁴ have long interpreted such trends as representing normative developmental changes in adults’ personality traits, and as such to be the result of biological mechanisms presumably associated with evolutionary advantages. This attribution is not really an explanation, but it does suggest that in search of a deeper understanding, it may be useful to borrow aspects of life course theory from evolutionary biologists²⁴. Taking such an evolutionary perspective suggests a role for mechanisms like reciprocal altruism (i.e., behaviors with mutual benefits that emphasize the possible delayed benefits of prosocial behaviors⁶⁷).

Indeed, the current results exemplify the kind of age differences and their consequences that are the focus of life course theory. We refer to the intergenerational conflict inherent in the differential timing of changes in Neuroticism and Compassion pointed out by Myerson and colleagues⁵. Neuroticism tends to be associated with individualism²⁵ and is more likely to characterize younger adults, whereas Compassion tends to be associated with a more prosocial, communitarian perspective that tends to characterize older adults. Thus, younger adults, perhaps because they were less susceptible to COVID-19’s most serious consequences (as well as their being more likely to be in search of productive partners), were also less likely to engage in some behaviors (e.g., social distancing) that could affect the spread of the disease. Older adults, in contrast, were more likely to suffer severe consequences from the disease that younger adults may have been more likely to infect them with.

We are not suggesting that people were actually thinking about infection in these terms during the height of the pandemic. Rather, what we are suggesting is that people’s age-specific behavioral tendencies were perhaps shaped by these contingencies, which may have characterized age differences in susceptibility to infectious diseases and age-related declines in reproductive potential for many generations. It may be noted that life course theory, with its many examples from animal behavior, offers explanations in terms of delayed benefits that are independent of individuals’ conscious intentions. As a result, it represents a fundamental contrast with traditional ethical views that emphasize intentions and selfless acts. However, our focus here is on what evolutionary biologists

might call proximate causes rather than ultimate causes like a behavior’s evolutionary advantages²⁶.

---

Conclusions

We believe the present findings have important implications for public health communication strategies although, unfortunately, perhaps the most general implication is that the communication problem is a complicated one. To begin with, the characteristics that best predicted a person’s decisions, and that one would therefore want to influence, depended on the specific nature of the decision. For example, measures were differently predictive depending on whether decisions involved distancing or vaccination: Compassion was only weakly (albeit significantly) associated with past and future vaccination decisions, but it was a reliably good predictor of distancing. Even with vaccination, it made a difference as to whether the decisions had actual or hypothetical (future) consequences. The final straw, however, is that not only do people and kinds of decisions differ, but so do pandemics²⁷²⁸. Pandemics differ in which segments of the population are most susceptible to infection as well as in how those infections are transmitted. The good news is that despite all this specificity, the internet has greatly facilitated the ability to take advantage of market segmentation in public health²⁹, making it possible to target different messages at different groups in different situations.

Nevertheless, some discoveries provide information that is likely to be of general use, and we hope that is the case for our cascade model. For example, our findings concerning age and distress and its effects on mitigation behaviors appear to be of general applicability. While social deprivation produces psychological distress³⁰, distress also appears to interfere with hygienic practices that either involve some degree of social deprivation themselves or that are reminiscent of such deprivation⁴. It is possible that during the COVID-19 pandemic this occurred because these practices exacerbated the distress-inducing effects of quarantines and social distancing. Regardless of the mechanism, however, what is clear is that young adults are particularly susceptible, and that as a result, the use of so-called ‘fear messages’ is problematic, and especially so for that audience. Although creating anxiety about not engaging in a particular hygienic practice is a common approach to motivating that practice³¹, it may be self-defeating in light of the present findings that suggest that fear messages have the potential to create resistance to the very behavior they are promulgating. Moreover, because regardless of age, psychological distress appears to interfere with compassion, reducing distress would appear to be desirable not just during pandemics, but whenever one wants people to make compassionate decisions.

---

Conflicts of Interest Statement:

The authors have no conflicts of interest to declare.

---

Funding:

This research was supported by the National Institute on Aging of the National Institutes of Health under award number R01AG058885.

---

Acknowledgement:

The authors would like to thank Yu-Hua Yeh and Julia Myerson for pointing us in this direction.

 

References

 

1. Aknin LB, De Neve JE, Dunn EW, et al. Mental health during the first year of the COVID-19 pandemic: A review and recommendations for moving forward. Perspect Psychol Sci 2022;17(4): 915–936. https://doi.org/10.1177/174569162110299.

2. Twenge JM, Joiner TE, Rogers ML, Martin GN. Increases in depressive symptoms, suicide-related outcomes, and suicide rates among U.S. adolescents after 2010 and links to increased new media screen time. Clin Psychol Sci 2018;6(1):3–17. https://doi.org/10.1177/2167702617723376.

3. Hale S, Myerson J, Strube MJ, Green L, Lewandowski AB. Distress signals: Age differences in psychological distress before and during the COVID-19 pandemic. Int J Environ Res Public Health 2023;20(4): 3549.
https://doi.org/10.3390/ijerph20043549

4. Myerson J, Strube MJ, Green L, Hale S. Individual differences in COVID-19 mitigation behaviors: The roles of age, gender, psychological state, and financial status. PLoS One 2021;16(9):e0 257658. https://doi.org/10.1371/journal.pone.0257658

5. Myerson J, Strube MJ, Green L, Hale S, Bernstein B. Decision-making during the pandemic: Compassion, trust, and the altruistic paradox. COVID 2023;3(5):744-756.
https://doi.org/10.3390/covid3050055

6. Trivers RL (1971). The evolution of reciprocal altruism. The Quarterly review of biology 1971;46 (1):35-57.

7. Trivers R. Reciprocal altruism: 30 years later. Cooperation in primates and humans: Mechanisms and evolution 2006; 67-83.

8. Enticott J, Dawadi S, Shawyer F, et al. Mental health in Australia: Psychological distress reported in six consecutive cross-sectional national surveys from 2001 to 2018. Front Psychiatry 2022; 13:815904. https://doi: 10.3389/fpsyt.2022.815904

9. Pierce M, Hope H, Ford T, et al. Mental health before and during the COVID-19 pandemic: A longitudinal probability sample survey of the UK population. Lancet Psychiatry 2020;7(10):883-892. https://doi.org/10.1016/ S2215-0366(20)30308-4

10. Zacher H, Rudolph CW. Individual differences and changes in subjective wellbeing during the early stages of the COVID-19 pandemic. Am Psychol 2021;76(1):50. https://doi.org/10.1016/j.paid.2021.110694

11. Daly M, Robinson E. Longitudinal changes in psychological distress in the UK from 2019 to September 2020 during the COVID-19 pandemic: Evidence from a large nationally representative study. Psychiatry Res 2021;300:113920. https://doi.org/10.1016/j.psychres.2021.113920

12. Myerson J, Strube MJ, Green L, Hale S, Bernstein B. Differential effects of psychological distress on mitigation and vaccination: A public health conundrum. Frontiers Psychol 2022;13:9230 56. doi: 10.3389/fpsyg.2022.923056

13. Luo J, Zhang B, Cao M, et al. The stressful personality: A meta-analytical review of the relation between personality and stress. Pers Soc Psychol Rev 2023;27(2):128-194.

14. Costa PT Jr, McCrae RR, Löckenhoff CE. Personality across the life span. Annu Rev Psychol 2019;70:423–448.

15. Van Bavel JJ, Gadarian SK, Knowles E, Ruggeri K. (2024). Political polarization and health. Nat Med 2024;30; 1-9. https://doi.org/10.1038/s41591-024-03307-w

16. Rayner K, Schotter ER, Masson ME, Potter MC, Treiman R. So much to read, so little time: How do we read, and can speed reading help? Psychol Sci Public Interest 2016;17:4–34.
https://doi.org/10.1177/1529100615623267.

17. Zigmond RP, Snaith AS. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983;67:361–70. https://doi.org/10.1111/j.1600-0447.

18. Morey JN, Boggero IA, Scott AB, Segerstrom SC. Current directions in stress and human immune function. Curr Opin in Psychol 2015;5:13-17.
https://doi:10.1016/j.copsyc.2015.03.007

19. Hughes ME, Waite LJ, Hawkley LC, Cacioppo JT. (2004). A short scale for measuring loneliness in large surveys: Results from two population-based studies. Res Aging 2004;26(6):655-672.
https://doi:10.1177/0164027504268574

20. Breslau J, Finucane ML, Locker AR, Baird MD, Roth EA, Collins RL. (2021). A longitudinal study of psychological distress in the United States before and during the COVID-19 pandemic. Prev Med 2021;143:106362.
https://doi.org/10.1016/j.ypmed.2020.106362

21. Bruine de Bruin W. Age differences in COVID-19 risk perceptions and mental health: Evidence from a national US survey conducted in March 2020. J Gerontol B Psychol Sci Soc Sci 2021; 76(2):e24-e29. https://doi:10.1093/geronb/gbaa074

22. Sparrow EP, Swirsky LT, Kudus F, Spaniol J. Aging and altruism: A meta-analysis. Psychol Aging 2021;36(1):49-56. http://dx.doi.org/10.1037/pag0000447

23. Wiepking P, James RN 3rd. Why are the oldest old less generous? Explanations for the unexpected age-related drop in charitable giving. Ageing Soc 2013;33(3):486–510.

24. Del Giudice M, Gangestad SW, Kaplan HS. Life history theory and evolutionary psychology. In: Buss, DM, editor. The handbook of evolutionary psychology – Vol 1: Foundations (2nd ed.), John Wiley & Sons Inc. 2015. pp. 88-114.

25. Uziel L, Seemann M, Schmidt‐Barad T. From being alone to being the only one: Neuroticism is associated with an egocentric shift in an alone context. J Pers 2020;88(2):339-355.
https://doi.org/10.1111/jopy.12494.

26. Alessi G. (1992). Models of proximate and ultimate causation in psychology. Am Psychol 1992; 47(11):1359–1370. https://doi.org/10.1037/0003-066X.47.11.1359.

27. Morens DM, Taubenberger JK, Fauci AS. A centenary tale of two pandemics: The 1918 influenza pandemic and COVID-19, Part I. Am J Public Health 2021;111(6):1086–1094.

28. Morens DM, Taubenberger JK, Fauci AS. A centenary tale of two pandemics: The 1918 influenza pandemic and COVID-19, Part II. Am J Public Health 2021;111(7):1267-1272. https://doi.org/10.2105/AJPH.2021.306326

29. Evans WD, Thomas CN, Favatas D, Smyser J, Briggs J. Digital segmentation of priority populations in public health. Health Educ Behav 2019;46:81–89. https://doi.10.1177/1090198119871246

30. Brooks SK, Webster RK, Smith LE, Woodland L, Wessely S, Greenberg N, et al. The psychological impact of quarantine and how to reduce it: rapid review of the evidence. Lancet 2020;395:912–20. https://doi.org/10.1016/S0140-6736.

31. Witte K, Allen M. A meta-analysis of fear appeals: Implications for effective public health campaigns. Health Educ Behav 2000;27(5):591-615.