An existential perspective on interpersonal closeness and social distancing during the COVID-19 pandemic

Abstract

Objective

The terror management health model proposes that COVID-19’s association with death should increase a desire for closeness, which could undermine social distancing intentions. Alternatively, social distancing intentions may increase if it has become culturally valued. The present research assessed these claims.

Design

We conducted three pre-registered online experiments (N = 409) where we manipulated COVID-19 thoughts, and examined either proximal (i.e. immediate) or distal (i.e. delayed) reactions.

Main outcome measures

Death-thought accessibility, desire for closeness, social distancing intentions, and measures concerning the perceived cultural value of social distancing.

Results

COVID-19 thoughts increased death-thought accessibility (Study 1). COVID-19-induced death thoughts increased a distal desire for closeness, particularly in securely attached individuals, but also increased the perceived value of, and intentions to engage in, social distancing (Studies 2 and 3).

Conclusion

Our findings demonstrate how proximal strategies may progress to distal strategies when they become embedded within cultural worldviews. These findings can aid in sustained efforts to encourage social distancing behavior.

Keywords:

• COVID-19
• social distancing
• terror management
• interpersonal relationships
• closeness

COVID-19 poses an unprecedented global crisis. Citizens worldwide have been limiting close physical contact with others (social distancing), which was recommended by health agencies as a means of mitigating the spread of the disease (Centers for Disease Control and Prevention (CDC), 2020). Achieving levels of social distancing stringent enough to be effective (Lewnard & Lo, 2020) can prove challenging. The virus’ association with death should, according to the terror management health model (TMHM; Arndt & Goldenberg, 2017), instigate desires to affirm interpersonal relationships. These existential concerns could undermine effective social distancing practices, reducing the ability to manage COVID-19. Utilizing the TMHM as a framework, we conducted three studies to elucidate how COVID-19-induced death thoughts influence competing desires to seek closeness with others and intentions to engage in virus mitigating behaviors.

The TMHM (Arndt & Goldenberg, 2017) starts with the assumption that the human awareness that death is inevitable has the potential to be psychologically debilitating. The model stipulates that people manage the awareness of their mortality through defensive tactics that operate at a conscious (proximal defenses) and unconscious level (distal defenses).

Proximal defenses are engaged after people are consciously reminded of their mortality; for example, by threats to one’s health. These defenses entail attempts to push death thoughts outside of conscious awareness, whether by effortful suppression (Arndt et al., 1997), denying one’s vulnerability to the health threat (Cooper et al., 2010), or engaging in health preventive action (Arndt et al., 2006). These proximal means of addressing death-related concerns can therefore produce both effective and ineffective health responses.

When thoughts of death are no longer conscious, albeit still accessible, people engage in distal defenses that manage death anxiety on a symbolic level through cultural worldviews and self-esteem (Arndt & Goldenberg, 2017). Distal defenses typically occur after a short delay after thoughts of death have left focal attention (Pyszczynski et al., 1999). As with proximal defenses, distal defenses can produce health-facilitating or defeating outcomes. This is because distal defenses are driven by concerns related to self-esteem rather than health. For example, after being reminded of death, those who’s self-esteem is based on fitness reported increased intentions to exercise (Arndt et al., 2003), while those who base their self-esteem on smoking reported increased intentions to smoke (Hansen et al., 2010). Because distal defenses concern symbolic ways of managing the problem of death, these defenses may not have any logical connection to the health threat itself (Courtney et al., 2020).

Thoughts of death are often intimately connected to health-related issues. For example, thinking about contracting the Ebola virus increases death-thought accessibility (DTA; Arrowood et al., 2017). Due to the COVID-19 virus’ association with death (Fairlamb, 2021; Jimenez et al., 2020), the TMHM serves as a useful framework to illuminate some of the ways that people psychologically manage the threat of the virus.

First, if reminders of COVID-19 arouse conscious death thoughts, this should motivate proximal attempts to reduce one’s perceived susceptibility to the virus. For example, people might engage in increased handwashing or face mask usage (Courtney et al., 2020). Social distancing may also act as a proximal defense, whereby staying home and maintaining a safe distance from others reduces perceived risk.

However, when death thoughts are no longer conscious but still non-consciously accessible, this should motivate distal defenses. Although typically a proximal defense, social distancing may also act as a distal defense to death awareness. This is because social distancing has been conveyed as a way for people to play their part in the pandemic, as social distancing not only protects oneself, but also others by minimizing the spread of the virus. Such discourse implies that people can contribute to something greater than their mortal selves. Thus, although social distancing initially serves as a proximal defense, it may also become embedded within one’s worldviews, and be adopted as a distal defense (Courtney et al., 2020). This perspective would suggest that proximal and distal defenses can produce the same behavioral response (i.e. increased social distancing), although these responses would have different underlying motivations. Demonstrating that proximal defenses may progress to a distal status when they become embedded within one’s cultural worldview would be a novel contribution to TMHM research.

However, interpersonal relationships also constitute a distal defense, as they can increase chances of reproduction, provide a sense of belonging, and constitute an important source of self-esteem (Florian et al., 2002). For example, death reminders increase desire for closeness (Mikulincer & Florian, 2000; Taubman-Ben-Ari et al., 2002), and threats to close relationships increase death anxiety (Vail et al., 2019) and DTA (Mikulincer et al., 2002). This desire for interpersonal connection might drive people towards closeness at the expense of social distancing.

This reliance on interpersonal relationships to manage death anxiety may, in part, depend on attachment style. Securely attached individuals, in comparison to insecurely (i.e. high avoidance or anxiety) attached individuals, tend to rely on others to manage distressing events (Shaver & Hazan, 1993). Therefore, death reminders heighten desire for closeness among those with secure attachments (Mikulincer & Florian, 2000; Taubman-Ben-Ari et al., 2002). In contrast, insecurely attached individuals respond to distressing events with increased defensiveness (Mikulincer & Florian, 1998, 2000).

Because both close relationships and social distancing behaviors have the potential to serve as distal defenses, particularly for securely attached individuals, it is difficult to predict how one might respond in the pandemic. More generally, the contradictory nature of these intentions highlights that existential concerns may promote ideological dilemmas (Billig et al., 1988). Existential concerns heighten a desire to be close to others, which, in the context of the pandemic, may require transgressing norms regarding social distancing. Conversely, adherence to social distancing-based norms may leave closeness desires unfulfilled. To our knowledge, no prior research has shown that death concerns may motivate two contradictory tendencies.

The goal of this research was to examine how COVID-19-induced death thoughts affect desire for closeness and social distancing intentions. We report three pre-registered studies (https://osf.io/ua3bh), but also conducted an initial pilot study that is reported in the Supplementary Analyses which showed that thoughts of COVID-19 significantly increased a desire for closeness providing initial support for our assertions. The present research was conducted March–April 2020 as the U.S was approaching their first wave of the pandemic. All studies obtained ethical approval from the university. The datasets for all studies can be found online (https://osf.io/ua3bh).

In Study 1, we first examined whether COVID-19 would heighten existential concerns by measuring DTA. Participants answered two-open ended questions about COVID-19 or a control topic (dental pain), and then immediately completed a DTA measure. We predicted that COVID-19 would increase DTA.

Study 1

Participants and procedure

We recruited 102 U.S. participants via MTurk for a study ostensibly concerning personality styles and health. While Amazon MTurk samples might tend to be slightly younger and less religious than the average US population (e.g. Berinsky et al., 2012), MTurk provides a more diverse and nationally representative sample than other recruitment methods (e.g. college students) that are typically used in social science (e.g. Behrend et al., 2011), and can provide high quality data providing researchers use quality control checks (e.g. attention checks; Cheung et al., 2017). We excluded 6 who failed an attention check and 10 who had incomplete data, leaving a final sample of 86 (M_age = 40.8, SD_age = 11.7), with 43 men and 43 women. Participants were randomly assigned to the COVID-19 (n = 40) or control (n = 46) condition.

Materials

Participants were randomly assigned to answer two open-ended questions about contracting COVID-19 (e.g. ‘please briefly describe the emotions that the thought of contracting the coronavirus arouses in you’; ‘jot down, as specifically as you can, what you think will happen to you when you experience the coronavirus’), similar to the traditional manipulation used in terror management research (Pyszczynski et al., 2015). Control participants answered two parallel questions about dental pain. Participants on average spent 122.5 seconds (SD = 97.5) on the task and engagement did not differ by condition.

Immediately after the COVID-19/dental pain manipulation, participants completed a 33-item word-stem DTA task (e.g. Greenberg et al., 1994) whereby 8 words could be completed in a death or neutral-related manner (e.g. KI _ _ E D could be killed or kissed). More death word completions reflect higher levels of DTA (score from 0–8).¹

Study 1 results

The data were normally distributed (skewness and kurtosis <.4) and the assumption for homogeneity of variance was met, F = 1.23, p = .270. Thinking about COVID-19 significantly heightened DTA, t (84) = 2.23, p = .028, d =.48, 95%CI [.05, .91]. Those who wrote about COVID-19 completed more death words (M = 3.08, SD = 1.23), than those who wrote about dental pain (M = 2.54, SD = .98).

Study 1 discussion

Our findings provide a basis for the notion that thinking about COVID-19 might function similarly to other death reminders by increasing DTA, suggesting that COVID-19 arouses existential concerns. In our next study, we set out to test our assertion that COVID-19-induced death thoughts can produce proximal and distal reactions.

Study 2

Our key goal of Study 2 was to examine the proximal and distal effects of COVID-19 reminders on desire for closeness and social distancing intentions. Our central prediction was that COVID-19 reminders would decrease a desire for closeness immediately, but that desire for closeness would increase after a delay. We also examined desire for closeness with significant others (i.e. friends, family) and the general public. We reasoned that heightened closeness would be stronger concerning significant others. For social distancing intentions, we expected the opposite pattern of findings to emerge (i.e. increased adherence immediately but decreased after a delay).

Additionally, we included a measure of attachment style as those with secure attachments manage existential concerns with interpersonal relationships (Taubman-Ben-Ari et al., 2002). Although we wondered if a basic effect of COVID-19 reminders increasing desire for closeness might occur, we also held a rough confirmatory hypothesis (Fife & Rodgers, 2022) that this would be moderated by attachment style. Based on past research, we anticipated that those low in avoidance and anxiety attachment (i.e. securely attached) would most strongly rely on interpersonal relationships to manage COVID-19-induced death thoughts. Finally, we also included a measure of the perceived effectiveness of social distancing (i.e. response efficacy) in mitigating the spread of COVID-19 for exploratory purposes. One possibility is that this might moderate the extent to which people choose to engage in social distancing to protect themselves from COVID-19.

Participants and procedure

We recruited 256 U.S. participants via MTurk. We excluded 45 because they failed to recall what health condition they had read and/or failed an attention check, leaving a final sample of 211 (M_age = 41.1, SD_age = 12.7), with 113 men, 96 women, 1 other, and 1 not answering this question. Participants were randomly assigned to one of four conditions: COVID-19/no delay (n = 56); COVID-19/delay (n = 56); Control/no delay (n = 53); Control/delay (n = 46). Participants were informed the study concerned health and attitudes (for full materials included in this study see: https://osf.io/apmu6/).

Materials

Participants completed a 9-item measure (e.g. ‘it helps to turn to people in times of need’) of attachment (Fraley et al., 2011) on a 7-point scale, whereby 3 items measured anxiety (α = .90), and 6 items measured avoidance (α = .84). Higher scores reflect higher levels of anxiety and avoidance attachment, respectively.

Subsequently, to ensure we could measure proximal and distal responses to COVID-19-based death reminders, participants were randomly assigned to read information about COVID-19 or dental pain. The COVID-19 information included worldwide mortality rates and estimates of case mortality rates. After reading this, participants then answered two questions regarding how likely they thought that they would die or contract COVID-19 (0–100%). This ensured that death thoughts were conscious. Dental pain participants received information about the prevalence of dental pain and answered a parallel question about experiencing dental pain.

After this manipulation, we randomly assigned participants to complete the dependent measures either immediately or after a delay. The delay consisted of the 60-item Positive and Negative Affect Schedule—Expanded Form (PANAS-X, Watson & Clark, 1999),² and a 33-item Social Desirability Scale (Crowne & Marlowe, 1960). The average length of the delay was 232.5 seconds (SD = 140.9), which is longer than a typical delay length for research in this area (Steinman & Updegraff, 2015). Further, longer delays are preferable in TMHM-based research (Steinman & Updegraff, 2015). Those in the delay condition completed these delay materials before the dependent measures. Those in the immediate condition completed these materials after the dependent measures.

Closeness was measured with 8 items (e.g. ‘I would like to spend time with people that I am close to’), 4 regarding close others (i.e. friends, family), and 4 to the general public on a scale from 1 (strongly disagree) to 7 (strongly agree).

Social distancing intentions were measured with 10 items (e.g. ‘avoiding close contact with others’) on a scale from 1 (not at all) to 7 (entirely) (α = .90). These face-valid items were constructed based on published guidance at the time (e.g. as detailed by the CDC, 2020; for similar procedures see Hagger et al., 2020) and similar to other measures of social distancing (e.g. Pedersen & Favero, 2020; Wu et al., 2022). Higher scores reflect higher intentions to engage in social distancing.³ We also measured the perceived response efficacy of social distancing in two items (α = .91; e.g. ‘I believe social distancing can reduce the spread of the coronavirus’). The order for closeness and distancing was randomized.

Finally, participants provided some basic demographic information.

Study 2 results

A factor analysis supported a two-factor solution of our closeness measure, but these did not conform to our expectation of close versus general others. One factor resembled a desire for closeness (α = .74), as the items concerned the extent to which people would miss or like to spend time with others. We labelled the other factor contact intentions (α = .79), as these items were phrased as direct statements about whether the participant would consider meeting up with other people or going out to public places. Supporting this, the contact intentions factor was strongly negatively correlated with social distancing intentions, r = −.58, p <.001, while the closeness scale was only weakly positively correlated with contact intentions, r = .15, p =.025. Consequently, we computed a mean score for the closeness factor to utilize in our main analyses,⁴ and discarded the contact intentions factor. Interested readers can find analyses of the contact intentions factor in the Supplementary Analyses.

Desire for closeness

Homogeneity of variance was met, F = 1.1, p = .347. We conducted a 2 (Health threat: COVID-19 vs. Control) x 2 (Delay: Yes vs. No) ANOVA on desire for closeness. The main effects were not significant, p’s >.20. The only significant effect was the predicted interaction, F (1, 207) = 10.39, p = .001, η_p² = .05. Examining the closeness interaction at both levels of the delay (see Table 1 for descriptive statistics) showed the expected effect of COVID-19 decreasing desire for closeness immediately, t (207) = 2.70, p =.007, d = .51, 95%CI [.13, .89]. After a delay, there was a marginal effect of COVID-19 increasing desire for closeness, t (207) = 1.87, p = .063, d = .38, 95%CI [−.01, .77]. Additionally, there was only an effect of delay in the COVID-19 condition, t (207) = 3.24, p = .001, d = .64, 95%CI [.26, 1.01], but not in the control condition, t (207) = 1.38, p = .168, d = .27, 95%CI [−.13, .67].

Table 1. Means and standard deviations (in parentheses) for desire for closeness in Study 2.

	COVID-19	Control
Immediate	4.45a (1.43)	5.16b (1.36)
Delay	5.29c (1.21)	4.48b (1.49)

Note. Higher scores reflect higher desire for closeness. Cells that do not share a subscript differ at p < .07.

Social distancing intentions

We conducted the same analysis for social distancing intentions. Homogeneity of variance was met, F < 1, p > .50. Again, the main effects were not significant, p’s >.70, but the interaction was, F (1, 207) = 8.45, p = .004, η_p² = .04. Examining this interaction (see Table 2 for descriptive statistics) showed that, contrary to expectations, reading about COVID-19 decreased intentions to engage in social distancing immediately, t (207) = 2.34, p = .021, d = .43, 95%CI [.11, .75]. After a delay, there was a marginal effect of COVID-19 increasing intentions to engage in social distancing, t (207) = 1.79, p = .075, d =.37, 95%CI [−.02, .77]. Looked at differently, social distancing intentions differed as a function of delay in the COVID-19 condition, t (207) = 2.27, p =.024, d =.42, 95%CI [.04, .79]. There was also a marginal effect of delay in the control condition in the opposing direction, t (207) = 1.85, p =.066, d = .39, 95%CI [−.01, .78].

Table 2. Means and standard deviations (in parentheses) for social distancing intentions in Study 2.

	COVID-19	Control
Immediate	5.40a (1.20)	5.86b (.90)
Delay	5.84c (.88)	5.48d (1.09)

Note. Higher scores reflect higher intentions to engage in social distancing. Cells that do not share a subscript differ at p < .08.

Although we had originally speculated whether response efficacy of social distancing might moderate our effects, exploratory analyses showed that reading about COVID-19 had a marginal effect of reducing the perceived response efficacy of social distancing, F (1, 207) = 3.04, p = .083, η_p² = .01, which was qualified by a marginal interaction, F (1, 207) = 3.64, p = .058, ηp² = .02. This interaction showed that immediately after reading about COVID-19, perceived response efficacy of social distancing was significantly lower, t (207) = 2.63, p = .009, d = .51, 95%CI [.13, .89], but there was no effect after a delay (p > .50). We therefore speculated that our effects on social distancing intentions might be due to changes in the perceived response efficacy of social distancing. Perceived response efficacy was also significantly correlated with distancing intentions, r = .55, p < .001, adding credence to the idea that changes in perceived response efficacy might account for the decrease in social distancing intention.

Rerunning the analysis controlling for response efficacy found that the interaction was still significant, F (1, 206) = 4.86, p = .029, ηp² = .02, though the pattern of this interaction was meaningfully different. There was now no effect of COVID-19 on social distancing intentions immediately (p > .40). The effect of COVID-19 on social distancing intentions after a delay was significant, F (1, 206) = 4.20, p =.042, η_p² = .02. There was also a marginal effect of the delay in the COVID-19 condition, F (1, 206) = 2.98, p = .086, η_p² = .01, but no effect in the control condition after controlling for response efficacy (p > .25). Part of the interaction therefore appeared to be driven by the perceived response efficacy of social distancing. Supporting this view, the immediate decrease after reading about COVID-19 was significantly mediated by response efficacy perceptions, b = −.14, 95%CI [−.26, −.04], but did not mediate the effect after a delay, b = .01, 95%CI [−.11, .12].

Attachment style

We examined whether our effects on desire for closeness were qualified by attachment (see Supplementary Analyses for social distancing intention). We utilized Model 3 in PROCESS (Hayes, 2018), inserting health threat as the independent variable (−1 Control, +1 COVID-19), and the delay as a moderating variable (W). Anxiety and avoidance attachment were inserted as an additional moderating variable in separate models (Z).

The three-way interactions were not significant for avoidance, β = −.03, t (203) = .44, p = .659, 95% CI [−.19, .12], or anxiety, β = −.06, t (203) = 1.02, p = .308, 95% CI [−.17, .05]. However, our study was likely underpowered to detect our rough confirmatory hypothesis, and more related to the simple effects in low anxiety and avoidance rather than the interaction term. As we preregistered these analyses a-priori, we continued with our analysis plan. The two-way interactions between Health Threat X Delay at low levels (−1 SD) of attachment style suggested that the interaction was significant in participants low in anxiety, β = .42, p = .002, and marginal for those low in avoidance, β = .25, p = .056. These interactions were not significant at high levels (+1 SD) of anxiety or avoidance, p’s >.10.

Analyses showed that COVID-19 reminders decreased desire for closeness among participants low in anxiety, β = −.51, t (203) = 2.70, p = .008, 95% CI [−.88, −.14], but not avoidance (p >.40) in the immediate condition. After a delay, there were marginal effects of COVID-19 reminders increasing a desire for closeness in participants in low avoidance, β = .36, t (203) = 1.95, p = .053, 95% CI [−.00, .73], and anxiety, β = .33, t (203) = 1.67, p = .096, 95% CI [−.06, .71].

Discussion

Supporting our prediction, the findings demonstrated that reading about COVID-19 decreased desire for closeness immediately but increased it after a delay, albeit the latter was only marginal. Our findings also suggest that this effect was strongest in securely attached individuals (i.e. low anxiety and avoidance), which mirrors past research (e.g. Taubman-Ben-Ari et al., 2002). However, these attachment findings were marginal. More generally, our findings were identified in an ad-hoc measure. Thus, to increase confidence in these effects we sought to address them in a new sample with a stronger measure of closeness.

Although we found a significant interaction for social distancing intentions, contrary to our prediction, social distancing intentions decreased immediately after reading about COVID-19 but increased after a delay. Concerning the former, the effect was driven by a decreased perceived response efficacy of social distancing. Though not consistent with our prediction, such a finding would correspond with other TMHM research that has shown that proximal reactions are dependent on whether behaviors are perceived as effective in protecting one’s health (Cooper et al., 2010), and that death reminders can decrease people’s perception that other COVID-19 health behaviors (i.e. face masks) are effective in managing the virus (Fairlamb et al., 2022). Importantly, our manipulation did not include any statements regarding the perceived efficacy of social distancing. Regarding the latter, the delayed increase in social distancing intentions might suggest that social distancing is perceived as a culturally valued behavior. We investigated this latter claim in Study 3.

Study 3

In Study 3, we dropped the delay manipulation and focused on examining distal responses only. This is because our findings regarding social distancing intentions increasing after a delay was unexpected and required further clarification of the underlying processes. One possibility is that social distancing is perceived as a culturally valued behavior. Reading about COVID-19 may constitute a reminder of this norm, as well as a death prime. As such, we expected our COVID-19 prime would increase the perceived cultural value of social distancing. Additionally, if social distancing has become culturally valuable, we also expected that participants would increase their willingness to punish those who violate social distancing guidelines. We anticipated that this increase in punishment would be greatest among those high in anxiety and avoidance attachment (Mikulincer & Florian, 2000).

Additionally, although Study 2 suggested that a desire for closeness did not undermine intentions to engage in social distancing (on the contrary, we observed an increase), we sought to examine this further by measuring intentions to engage in social distancing from the general public, as well as close others. Presumably, desire for closeness would most likely heighten intentions to see close others at the expense of social distancing, rather than the general public.

Participants and procedure

We recruited 128 U.S. participants via MTurk using the same cover story in Study 2. We excluded 15 participants because they failed to recall what health condition they had read, and 1 participant who did not complete all the dependent measures. This left a final sample of 112 participants (M_age = 41.5, SD_age = 13.2), with 59 women, 52 men, and 1 identifying as ‘other’. Participants were randomly allocated to the COVID-19 (n = 59) or control (n = 53) condition. Full materials included in this study can be found online (https://osf.io/fc95k/).

Materials

The study was similar to Study 2 in that participants completed the avoidance attachment (α = .91), and anxiety attachment (α = .95) scales before COVID-19 (vs. dental pain) manipulation. The COVID-19 manipulation was updated to reflect accurate pandemic statistics at the time of participation. All participants then completed the delay materials before the dependent measures. We introduced a third measure to the delay, which consisted of rating the emotions of 8 faces, though whether participants completed this delay task depended on the speed in which they completed the first two delay tasks. This was to ensure all participants had a sufficient delay, and to standardize the length of the delay across participants. The average delay time was 236.7 seconds (SD = 128.9).

Mirroring prior research (e.g. Rosenblatt et al., 1989), we assessed social distancing-based worldview defense by asking participants to assign a hypothetical fine ($0–10,000) to people who violated social distancing guidelines.⁵ We also assessed whether participants perceived social distancing as culturally valued (e.g. ‘those who are staying at home as much as possible are making a valuable contribution to society’) with 4 items (α = .88). Higher scores reflect a higher perceived value of social distancing.

We assessed general social distancing intentions with 4 items that were almost identical to Study 2 (α = .92), and 4 items to measure social distancing intentions towards close others (α = .88; e.g. ‘inviting friends or family over to your home’). A factor analysis supported that these measured distinct aspects of social distancing intentions. We reversed the close other social distancing intentions scale so that higher scores on both scales reflect greater intentions to adhere to social distancing guidelines. We also measured perceived response efficacy of social distancing in two items used in Study 2 (α = .93) to use as a covariate in our analyses.

Finally, participants also completed a 5-item measure of desire for closeness similar to the previous studies (α = .89; e.g. ‘I would like to spend time with people that I am close to’). All dependent measures were completed on 7-point Likert scales, and the order of the measures was randomized.

Study 3 results

We first assessed the effect of reading about COVID-19 on our dependent measures, controlling for perceived response efficacy as specified in our pre-registration plan.⁶

There was no effect of COVID-19 reminders on social distancing from close others, F (1, 109) = .16, p =.690, η_p² = .00, or desire for closeness, F (1, 109) = .00, p =.965, ηp² = .00. The effect of COVID-19 reminders increasing social distancing intentions was also not significant, albeit marginal, F (1, 109) = 2.88, p =.093, η_p² = .03. COVID-19 reminders significantly increased the perception of social distancing as culturally valued, F (1, 109) = 5.64, p =.019, η_p² = .05, and willingness to give out larger fines for violating social distancing guidelines, F (1, 109) = .5.52, p =.021, η_p² = .05. Table 3 presents the descriptive statistics.

Table 3. Means and standard errors (in parentheses) of the dependent measures in Study 3.

	COVID-19	Control
Hypothetical fine ($)	1970.65a (298.76)	949.99b (315.22)
General social distancing	6.49a (.09)	6.27a (.10)
Close others social distancing	6.33a (.15)	6.24a (.16)
Culturally valued social distancing	6.23a (.07)	5.98b (.08)
Desire for closeness	5.33a (.17)	5.34a (.18)

Note. Rows that do not share a subscript differ at p < .05.

Attachment style

We examined the moderating role of attachment style on our worldview defense and closeness measures using Model 1 in PROCESS (Hayes, 2018). Analyses of the social distancing intentions measures can be found in the Supplementary Analyses.

The Health Threat X Anxiety Attachment interaction was approaching significance for hypothetical fines, β = 216.12, t (107) = 1.74, p = .085, 95% CI [−30.54, 462.78]. Simple slopes analyses suggested that COVID-19 increased fines for those with high (+1 SD) levels of anxiety attachment, β = 891.79, t (107) = 2.92, p = .004, 95% CI [285.83, 1497.76], but not at low levels (−1 SD) of anxiety attachment, p >.50. For avoidance attachment, the interaction was not significant, β = 131.46, t (107) = .75, p = .454, 95% CI [−.215.39, 478.30], though simple slopes analyses suggested the effect of increasing fines was significant in those with high levels of avoidance, β = 708.48, t (107) = 2.21, p = .029, 95% CI [74.08, 1342.90], but not at low levels of avoidance, p >.20.

For desire for closeness, there was a significant two-way interaction between Health Threat X Avoidance Attachment, β = −.27, t (107) = 3.15, p = .002, 95% CI [−.45, −.10]. Simple slope analyses suggested that at low levels of avoidance, COVID-19 increased desire for closeness, β = .39, t (107) = 2.49, p = .014, 95% CI [.08, .69], but decreased desire for closeness at high levels of avoidance, β = −.33, t (107) = 2.07, p = .041, 95% CI [−.65, −.01]. The Health Threat X Anxiety Attachment interaction failed to reach significance (p > .15).

Mediation analyses

We next examined whether the perception of social distancing as culturally valued mediated our effects to social distancing intentions and fines for violating social distancing guidelines. We utilized Model 4 in PROCESS (Hayes, 2018) to test for a basic mediation effect (i.e. COVID-19 → culturally valued behavior → fine/intentions). Bootstrapping with 5000 samples was used, whereby significant mediation occurs if the confidence interval does not include zero. These analyses suggested that the perception of social distancing as a culturally valued behavior mediated the effect of COVID-19 on intentions to engage in social distancing, b = .04, 95%CI [.01, .11], and fines for violating social distancing, b = 99.37, 95%CI [6.26, 226.77].

Study 3 discussion

Study 3 provides additional support for the notion that closeness concerns are heightened during the pandemic. However, such an effect was limited to participants who were low in avoidance attachment (i.e. securely attached). One possibility is that our measure of closeness was more strongly associated with avoidance than the anxiety dimension. For example, the avoidance dimension is primarily concerned with emotional distance, and the anxiety dimension concerns fear of rejection (Shaver & Hazan, 1993). Our measure about being around others might not have been particularly anxiety-provoking in comparison to past research that concerned desire for personal intimacy (Mikulincer & Florian, 2000) or willingness to strike up social interactions (Taubman-Ben-Ari et al., 2002).

Our findings are consistent with Study 2 and supported our prediction that social distancing may have become a culturally valued norm that increases people’s intentions to adhere to guidelines. We found converging support for this view. First, COVID-19 reminders increased assignment of larger fines for those who violated social distancing guidelines. This effect was also strongest among those high in avoidance and anxiety attachment, mirroring past research (Mikulincer & Florian, 2000). Second, COVID-19 increased perceptions of social distancing as a culturally valued behavior. Third, increases in social distancing intentions and fines for violating social distancing guidelines were mediated by the perception of social distancing as a culturally valued behavior.

We did not find that COVID-19 increased intentions to be socially distant from close others. Although this finding might be due to ceiling effects, it is also possible that it reflects the limit to distally-induced adherence to social distancing. Indeed, our general social distancing intentions measure was almost identically worded to instructions disseminated by government agencies (CDC, 2020). This might have made the measure more sensitive to tapping into the culturally valued nature of social distancing.

General discussion

The present research examined how COVID-19-induced death thoughts influence desire for closeness and intentions to socially distance. Our findings demonstrated that COVID-19 is implicitly associated with death and suggested that responses to death thoughts related to the disease depended on whether those thoughts were consciously or non-consciously accessible. Our findings provide a stronger understanding of how people might respond to social distancing-based messages by world leaders or medical experts in a pandemic. Additionally, they provide a novel theoretical contribution to the TMHM, whereby health behaviors themselves can be adopted as distal defenses against mortality threats.

First, our findings supported our hypothesis that proximal reactions to death thoughts would decrease desire for closeness. However, we expected that proximal reactions would increase social distancing intentions, but we instead observed a decrease. This difference might reflect that conscious death thoughts arouses concerns of losing loved ones (Firestone, 1993), but this does not necessarily entail being socially distant. Instead, COVID-19 might induce a fatalistic attitude that decreases people’s intentions to engage in safe practices. We believe our findings correspond with research that shows that the stronger COVID-19 is associated with death, the less likely people are to engage in preventative health behaviors (Jimenez et al., 2020). The decrease in preventative behaviors is a worrying concern for public health. We suspect, and indeed was supported by our analysis in Study 2, that such behaviors might be caused by perceptions that preventative measures are perceived as ineffective. This is consistent with research concerning how death thoughts can undermine confidence in the effectiveness of other COVID-19 related health behaviours (Fairlamb et al., 2022). Public health officials should make it a priority to highlight the response efficacy of such strategies when discussing COVID-19. Otherwise, well-intentioned messaging might backfire.

In terms of non-conscious death thoughts, our findings support the prediction that death concerns can instigate desires to feel close to others, particularly in securely attached individuals. However, this desire for closeness did not lead to decreased intentions to adhere to social distancing guidelines. Instead, social distancing intentions increased due to perceptions of social distancing as culturally valued.

Our findings therefore provide the novel illustration that, in some cases, existential concerns might promote contradictory intentions inasmuch that participants increased a desire to be close to others, but at the same time, showed heightened intentions to remain socially distant. This finding reflects how worldviews, and ways in which people manage death anxiety, are multi-faceted and could produce incoherent or conflicting defenses. For example, reminders of death can increase negative feelings towards worldview critics, but at the same time willingness to support their rights to free speech (Fairlamb & Cinnirella, 2021). In this case, one can manage existential concerns via affirming close relationships that provide a sense self-esteem and security, or by living up to cherished cultural values that promote the importance of limiting physical contact with others.

Future research could consider how people might negotiate ideological dilemmas (Billig et al., 1988) when reminded of death. Possibly, in the case of closeness and social distancing, people might be able to resolve such a dilemma through using internet-mediated communication, like video conferencing. More generally, it is possible that death-thought-induced dilemmas could be anxiety-provoking. For example, people who violate social distancing guidelines to satisfy a sense of closeness might feel a sense of guilt or anxiety because they transgressed against worldview norms, or exposed others to risk. Prior research has shown that death reminders heighten feelings of tension and apprehension when having to desecrate cherished cultural symbols (Greenberg et al., 1995). This is an important avenue for future research to explore because our worldviews are multi-faceted, and how people navigate, and cope, with these dilemmas is an untapped avenue of investigation.

The social distancing findings support the idea that proximal health-related responses to death can become distal defenses if they become embedded within cultural worldviews (Courtney et al., 2020). This is an important theoretical contribution to the TMHM, as it demonstrates that the same behavior can act as proximal or distal defense, though engagement with those behaviors may have different underlying motivations. Supporting this, in Study 2 we found that perceived response efficacy drove proximal, but not distal, responses. In Study 3, distal responses were driven by ideological concerns, but not response efficacy. Future research could further support these conclusions by measuring a myriad of proximal and distal concerns both immediately and after a delay.

From an applied perspective, the increased intentions to engage in social distancing as a distal defense against death thoughts is a promising finding in the continued need to manage the spread of the virus. As we initially expected, distal defenses against death have the potential to impede health-facilitating behaviors (Arndt & Goldenberg, 2017). Our findings imply that social distancing has become culturally significant and valuable, at least within American worldviews at the time these studies were conducted. The TMHM suggests that framing health choices as culturally valuable is more likely to lead to successful health strategies long-term. World leaders and medical experts would be well-advised to emphasize the extent to which engaging in preventative behaviors constitutes a valuable contribution to society, especially given how death thoughts are more accessible in the context of a pandemic. Reducing existential fears by tapping into concerns of individual self-worth, rather than health, is a better strategy for promoting a level of social distancing stringent enough to save lives.

Despite the studies being pre-registered, our research has some limitations. First, our self-report measures, particularly social distancing intentions, might be liable to social desirability effects. Although we cannot fully discount this possibility, we conducted exploratory analyses using the social desirability measure from the delay materials in Studies 2 and 3. This did not alter the overall outcome of the findings. The self-report nature of our dependent measures might also account for why some of our findings were only marginal and did not reach conventional thresholds of significance. Despite the marginal nature of some findings, most of our findings in this paper were replicated or clarified in subsequent studies, which should increase the confidence that those findings reflect genuine effects.

However, an additional limitation regarding the use of self-report is that our measurements therefore reflect intention not actual behaviour, and one should be careful in presuming that intention will lead to behaviour (e.g. Armitage & Conner, 2001). It is possible that despite intentions to engage in social distancing, people might still seek out social connections that undermine this intention. Our data might hint at this given our finding that despite increased intentions to engage in social distancing, participants also reported an increased desire to be close to others.

A second limitation concerns the fact that this research was conducted as the United States was approaching its first peak in COVID-19 cases (Institute for Health Metrics and Evaluations, 2020). One possibility is that as the curve flattens, emphasis on social distancing as a culturally valued behavior might wane. However, desire for closeness should remain stable across time, which might lead to a health-defeating impact on social distancing intentions in a less-controlled context. Thus, our findings might be limited in that they are temporally bound, and future research could consider doing follow-ups at different timepoints of the pandemic to see if a different pattern of findings emerge.

Finally, an additional limitation concerns whether the effects are the result of TMHM processes or can be explained by alternative accounts. The present research took the recommended experimental-causal-chain approach wherein we tested separate components of terror management theorizing across different studies (Hayes & Schimel, 2018). Therefore, although our data does suggest that COVID-19 reminders can increase DTA, as well as theoretically consistent distal outcomes, it is limited in its ability to demonstrate that these responses are the result of increased DTA. It is possible that our findings might reflect some other processes beside terror management, such as heightened social isolation (Hart, 2014) or uncertainty (Van den Bos, 2009). However, such issues were beyond the scope of our research aims, which was to assess how behavioral intentions in a pandemic may not be driven purely by rational, deliberated processes.

In sum, our studies shed light on some ways that death concerns influence responses during the COVID-19 pandemic. Additionally, they provide a new perspective on the theoretical application of the TMHM, whereby potentially life-saving individual health behaviors become intertwined with cultural worldviews and better inform the manner in which those behaviors might be communicated. Our findings provide both promise, and concern, for public health officials; but overall, are useful in informing effective communication strategies in tackling the pandemic in the long-term.

Open practices statement

Pilot Study was not pre-registered, but studies 1–3 were pre-registered. The anonymized data has been made available at: https://osf.io/ua3bh/.

Disclosure statement

The authors declared no conflicts of interest with respect to the authorship or the publication of this article.

Funding

This work was funded by the Research Initiative Fund at Royal Holloway, University of London.

Notes

1 In this study, we also measured desire for closeness using the same measure from our pilot study after the DTA measure. There was no effect of the manipulation on COVID-19. Most likely this was because measuring DTA would cause downstream effects on our closeness measure (see Hayes & Schimel, 2018) and this possibility was specified in our pre-registration plan (see https://osf.io/7e5kj/), thus it will not be discussed further.

2 Results concerning affect across all studies can be found in the Supplementary Analyses.

3 Because of possible social desirability and ceiling effects, we measured both personal and perception of other people’s social distancing intention with 5 items each. Asking about other’s intention can reveal masking of one’s own attitudes. As the findings for both scales were similar, and combining them produced a reliable composite, we opted to combine them for parsimony. We provide the findings for each scale separately in the Supplementary Analyses.

4 Inspecting our key variables suggested that the data was not normally distributed. Transforming the data and running a non-parametric equivalent of our main analyses does not change the findings reported.

5 We also measured how much they would reward a person for adhering to such guidelines but found no reliable effects. We suspect this is because one’s worldview promotes the sense that people should engage in social distancing without the need for incentivisation, insofar as social distancing may be perceived as the “right thing” to do.

6 Data was not normally distributed (p’s <.05) and for the analysis regarding fines, homogeneity of variance was not met, F (1, 110) p = .002. We therefore transformed the data and conducted non-parametric equivalent tests. These largely did not alter the findings, except the finding for social distancing as culturally valued was now marginal, F (1, 109) = 3.47, p =.065, η_p² = .03. Because transforming the data did not significantly alter the findings, we report the unadjusted raw scores in text.