Better once it's over, worse now: prospective moral behaviors after the coronavirus epidemic and cyberchondria

The COVID–19 crisis might generate effects in what regards people's propensity to engage in moral and immoral actions, both during and after the pandemic. We used an experimental manipulation to explore the impact of the current lockdown on anticipated behaviors and its association with cyberchondria and moral identity. A convenience sample of 865 participants completed an online survey in which they were asked to rate the probability of engaging in certain morally positive and negative behaviors over the next week. The experimental group was required to imagine that the pandemic would be over by then, and all restrictions would be lifted, while the control groups did not receive this instruction. Participants' moral identity and cyberchondria were also measured. Results showed that participants in the experimental condition expressed higher intentions to perform positive behaviors and lower intentions to engage in negative actions. Furthermore, cyberchondria was positively associated both to intended negative behaviors in the control group, i.e., during the lockdown and to positive behaviors in the experimental group, i.e., after the lockdown. We discuss these results in relation to prospective moral licensing and cleansing, and to the role of health anxiety in shaping moral decisions.


3
 We found an increase in intended positive behaviors as an effect of the end of the coronavirus lockdown, in comparison to the current scenario, and a decrease in negative actions.
The COVID-19 pandemic brought significant changes to the political, economic, social, and emotional contexts that we were used to. While we are still adjusting to them, we cannot stop wondering how our world would look like when everything would go back to the normality that we previously knew. Moreover, the need to predict our post-pandemic behavior is essential in fast-changing times, such as these. Stressful events, such as a pandemic, increase violent behaviors (Parkinson, 2019;Zahran et al., 2009). Lockdown measures due to the COVID-19 health crisis increased a series of negative behaviors such as domestic violence or racism. A recent report by UN Women (2020) revealed that in countries across the globe, violence against women and girls has intensified with at least 25% since the middle of March, the average period for most lockdown measures. Discriminatory pronouncements have been documented in Romania, Bulgaria, or Slovakia, targeting Roma populations (Matcahe & Bhabha, 2020), while anti-Asian hate seems to be spreading both online and offline (Devakumar et al., 2020;Macguire, 2020).
Nevertheless, what happens after the COVID-19 crisis is over? This paper addresses this question in terms of the dynamics of moral versus unethical behavior, examined through the lenses of moral licensing and cleansing phenomena. Furthermore, as the current aggressive pandemic intensifies people's health anxiety, and as the internet became an essential source of health-related information, especially during the current lockdown, we also analyse the effect of cyberchondria on prospective moral behaviors.

4
A sense of trust in one's moral conduct activates Moral licensing (ML) mechanisms: one's past good, moral actions might motivate current immoral behavior. More specifically, a person who initially behaves morally can later display morally ambiguous behaviors (Blanken, van de Ven, & Zeelenberg, 2015): a prior good deed licenses a subsequent unethical one (e.g., Merritt et al., 2010). For example, if someone donates to charity, they might "forget" to call back a friend in need, or a daily low-carbs diet might license for a "carbs full" weekend. Conway and Peetz (2012) suggested that recalling moral behaviors, and not moral traits, may lead to morally licensed actions. Their theory was that recalling behaviors might be related to the progress towards the goal of being moral, and later ethical behaviors would become less important since the target has already been achieved. Oppositely, they suggest that recalling moral traits activates the general concept of morality, leading to moral consistency.
Additionally, the prospective moral licensing effect entitles people to behave unethically due to the mere anticipation of future moral deeds. More specifically, Cascio and Plant (2015) showed that merely planning to do good later allows individuals to gain moral credits for their current actions, therefore licensing their present immoral behavior. The opposite mechanisms to ML is moral cleansing, describing the tendency to perform morally positive behaviors in order to re-establish one's moral balance following unethical actions (West & Zhong, 2015).
The current lockdown aiming to slow the spread of the coronavirus imposes significant personal sacrifices for most people, in an entirely different life scenario from the one before this health crisis. One's adherence to the strict social distancing and isolation rules imposed by authorities, together with the idea that these personal sacrifices are acts of altruism aimed at protecting others, might also contribute to the accumulation of "moral credits" (Miller & Effron, 2010). These, in turn, could further generate lenience towards one's moral misconduct, by morally licensing immoral actions. Furthermore, immoral conduct during the highly specific current lockdown circumstances might induce the need to clean one's "moral slate" once the pandemic is over. We aim to examine, at least partially, these dynamics by comparing the intentions to engage in moral and immoral behaviors in two prospective scenarios: the current, i.e., pandemic and under lockdown, and the post-pandemic one.
The present study also examines a personal factor of the tendency to make morallyappropriate decisions, namely moral identity. The positive relationship between moral identity (MI) and moral behavior was found across various studies (e.g., Hardy & Carlo, 2011;Hardy, Bean, & Olsen, 2014;Jennings et al., 2015;Stets & Carter, 2011). Hertz and Krettenauer's meta-analysis (2016) suggested that MI strengthens the people's willingness to engage in prosocial actions, and also refrains them from antisocial behavior.

Anxiety and Moral Decision Making
Although people may use either healthy or maladaptive mechanisms in order to reduce their anxiety, previous results highlight an association between these emotions and immoral behaviors. Kouchaki and Desai (2015) found that anxiety can enhance unethical, self-interested acts, by serving as a coping mechanism in threatening situations. Bazerman and Tensbrusnel (2011) also suggested that anxiogenic situations marked by isolation, uncertainty, and time pressure (such as the lockdown imposed by the COVID-19 pandemic), can challenge our moral values. Materialistic desires (Chang & Arkin, 2002), consumption of scarce common resources (Kasser & Sheldon, 2000), or infidelity (Mark et al., 2011) were found to be common responses to anxious circumstances. Unethical behaviors may, therefore, serve as pathways to dealing with perceived threatening situations (Zhou et al., 2009).
Though searching for symptoms and medical information using online platforms is a frequent, current behavior (Murphy, 2019), the repeated and excessive online search for healthrelated information (i.e., cyberchondria) is considered to be a dysfunctional behavioral pattern 6 increasing health anxiety (Startevic et al., 2019). The current pandemic is a fertile ground for such maladaptive conducts, and it may cause strong emotional responses, increasing chances for cyberchondria occurrence. Concerns about protecting oneself or loved ones, as well as worrying that regular medical care may be disrupted due to facility reductions, are common reactions in a health crisis such as the present one. Consequently, in the current circumstances (i.e., under the coronavirus lockdown) cyberchondria might be a facet of health anxiety, particularly influential on the dynamics of prospective moral behaviors that we aim to examine.
The present study examined the associations of cyberchondria and moral identity to the intention to engage in specific moral and immoral behaviors, as well as the effect of experimental manipulation of the scenario in which these behaviors would be performed, i.e., either during the current pandemic or after its end.

Participants
We designed and ran a web-based, cross-sectional survey at the beginning of April 2020, several weeks after most European countries, including Romania, imposed lockdown measures (i.e., social isolation) following a substantial increase in COVID-19 cases. The survey was distributed by introductory psychology students enrolled in "Alexandru I. Cuza" University of Iasi, Romania, for partial fulfilment of course requirements to a convenience sample of 977 participants, out of which 112 were dropped from the study due to missing data.
The final sample, therefore, includes 865 participants. Participants were presented with an informed consent describing the purpose of the study, the expected duration of the one's participation (about 15 minutes), and ensuring the confidentiality and anonymity of their answers. A demographic survey accounted for gender, age, and educational level.

Procedure
At the end of the survey, participants were randomly assigned to either the control or the experimental group (see Table 1) before being asked to rate their chances to perform a set of morally positive and negative prospective behaviors. In the control group, these behaviors were presented as potential actions participants may engage in over the following week, thus during the current coronavirus lockdown. In the experimental group, they were required to rate 8 the probability of the same behaviors over the same period, but in the scenario that the "pandemic ends tomorrow, and all restrictions are lifted". The morally positive and negative actions were adapted from previous research related to variations in moral behavior (Blanken et al., 2015;Meijers et al., 2014;Sachdeva et al., 2009;Cain et al., 2005). The set of positive behaviors (Cronbach's alpha = .85) included four actions referring to two hypothetical donating behaviors and fundraising actions (i.e., "I would be actively involved in fundraising campaigns for vulnerable people"; "I would donate money to charity"), and two environmental concern and pro-environmental intentions (i.e., "I would be actively involved in green, proenvironmental activities, such as cleaning the woods"; "I would be actively involved in the global fight against pollution"). The three negative behaviors (Cronbach's alpha = .65) referred to the willingness to pay for luxurious over purposeful goods ("I would buy expensive things, such as clothes or jewelry, even if I didn't really need them"; "I would not refuse myself unnecessary, but indulging pleasures, such as buying a very expensive phone or a very expensive car") and to giving worse advice to increase own gains ("I would take more care of my personal needs, even if it meant disfavoring others").

Results
The demographical characteristics of each of the two groups generated by the experimental manipulation and their descriptive statistics on MIQ and CSS are presented in Table 1. Table 1  We also investigated the differences that cyberchondria and moral identity generate in participants' intentions to perform each type of behavior, both independently and in interaction with our experimental manipulation. To this aim, we performed two regression analyses, one on each type of action, using as predictors participants' experimental condition, cyberchondria, and moral identity, as well as the interaction between condition and each of these two psychological measures. The analyses also controlled for the effect of gender, age, and education. In the case of positive behaviors, the overall model accounted for 16% of the variance of these actions, F(8, 856)=18.62, p < .001. Moral identity emerged as a significant predictor (β = .28; p < .001), while its interaction with the experimental condition did not (β =.32; p =.24). Oppositely, cyberchondria was not a significant predictor independently (β = .09; p = .06), but the interaction term between cyberchondria and condition was significant in predicting positive behaviors (β = .37; p = .001). The same set of variables was also significant in predicting negative conducts, F(8, 856) = 23.46, p < .001, explaining 18% of their variance.

Please insert
A similar pattern of results emerged, with moral identity being a significant but negative predictor of these behaviors (β = -.28; p < .001), while its interaction with the experimental condition was not (β = -.09; p =.74). Cyberchondria emerged as a significant positive predictor (β = .15; p = .001), but this effect was qualified by a significant interaction with condition (β = -.26; p = .019).
In order to explore these significant interactions between cyberchondria and experimental condition, we used the simple slopes analysis in PROCESS (Hayes, 2013) Overall, this pattern of relationships suggests that high levels of cyberchondria are especially related to higher tendencies to perform negative behaviors in the near future under the current pandemic lockdown circumstances (i.e., in the control condition of our study) and positive behaviors after the pandemic.

Discussion
We investigated whether intentions to engage in moral and immoral actions over the next period would be influenced by manipulation of the scenario in which these actions would be performed, in terms of them being placed either during or after the current pandemic. The morally positive and negative actions were adapted from previous research related to variations in moral behavior (Blanken et al., 2015;Meijers et al., 2014;Sachdeva et al., 2009;Cain, Loewenstein, & Moore, 2005). The set of positive behavior actions referred to two hypothetical donating behaviors and fundraising actions, environmental concern, and pro-environmental intentions, while the negative behaviors described the willingness to pay for luxurious over regular goods and to giving worse advice to increase own gains.
Even though in the current COVID-19 pandemic situations, when the danger of getting infected by a potentially fatal disease, such behaviors might seem irrelevant to people who were anxious about their health and the future economic situation, we chose these specific behaviors for several reasons: 1) they are well documented within the moral licensing literature, and therefore relevant for the focal topic of the present study (Blanken et al., 2015;Meijers et al., 2014;Sachdeva et al., 2009;Cain, Loewenstein, & Moore, 2005); and 2) we were interested in creating a specific hypothetical post-pandemic scenario (experimental condition) when such potential behaviors would frequently occur in normal circumstances (as before the pandemic).
Using these specific actions, and not COVID-19-related behaviors, we aimed to focus participants' attention on their post-pandemic behavior (experimental condition).
We found an increase in intended positive behaviors as an effect of the end of the coronavirus lockdown, in comparison to the current scenario, and a decrease in negative actions. This suggests that people intend to be more moral after the lockdown compared to the present circumstances, and less immoral than now. The difference in immoral intentions might reflect a moral licensing effect during the coronavirus lockdown, generated by the "moral credits" accumulated by obeying restrictions and thus contributing to public health (Conway & Peetz, 2012;Miller & Effron, 2010). Therefore, we might be licensed to compensate for our civic conformism by indulging ourselves on several levels or performing other acts that otherwise we consider to be below our moral standards. Furthermore, in this perspective of the increase in current immorality, the difference in intended moral behaviors might reflect an 12 anticipated, post-pandemic, moral cleansing phenomenon, aimed at re-establishing moral balance through good deeds (West & Zhong, 2015).
Of course, an alternative explanation of our results is that participants' propensity to engage in moral and immoral actions was not affected by the current lockdown, and that they only show a post-pandemic increase in morality compared to the present circumstances. The absence of measurement of these behavioral intentions before the lockdown is an important limitation of our study. Such data would reveal whether the differences we found reflect a return to the previous and habitual moral praxis or a significant moral cleansing phenomenon, following the current circumstantial alienation from one's ethical standards. Further studies are needed to disentangle these different moral dynamics.
Yet, the moral effect of the coronavirus lockdown is also supported by the associations between cyberchondria, prospective moral, and prospective immoral behaviors. They suggest that health-anxious individuals tend to be more immoral during the lockdown, and more ethical following the pandemic. Thus, the current circumstances might indeed make people more prone to immoral actions, especially those high in cyberchondria and presumably other facets of health anxiety. Previous research revealed the emotion regulation role played by unethical or morally ambiguous conduct when coping with anxiety (e.g., Kouchaki & Desai, 2015;Zhou et al., 2009). In line with this body of research, our results indicate that immoral behaviour during the lockdown, as a way of dealing with cyberchondria, may lead to later ethical actions, as a result of moral cleansing mechanisms.
An alternative explanation of this increase in moral intentions following the pandemic can be conceived as reflecting differences in self-presentational motives (Kurzban & Aktipins, 2006). During the lockdown, moral behavior lacks a more substantial, public audience, while post-pandemic conduct would be less private, and the presence of an audience increases 13 moralistic punishment (Kurzban et al., 2006) and prosocial behavior (Haley & Fessler, 2005;Nettle et al., 2012). Consequently, the change from the private audience imposed by the lockdown measures to a public one might (also) contribute to the prospective increase in moral behavior.
Our results also relate to previous findings on antisocial and unethical actions during stressful events such as the current pandemic (Parkinson, 2019;Zahran et al., 2009).
Implications are ramified in multiple areas and domains, both personal and interpersonal. For example, people may emotionally overeat during isolation and quarantine, due to increased anxiety levels, counting on their post-pandemic dietary regimes. Others may target close people (i.e., family members) as victims of their immoral acts as ways of dealing with stress, implying both short and long-term negative psychological (and most probably legal) consequences for all parties involved.
Further studies should examine in more depth the moral effects of this type of circumstances, by differentiating more attentively the psychological influences of their various facets. For instance, another limitation of our research is that our data cannot distinguish between the effects of the coronavirus pandemic in itself and those of the lockdown and other social and economic dynamics put in place by the disease. Moreover, we did not measure participants' intentions to perform the specific positive or negative behaviors before the pandemic, which is an important limitation of the current study, and including such data would have allowed us more comprehensive assertions of its findings.
Furthermore, even though our results suggested significant evidence to explain people's immoral behavior during the pandemic and more moral intended conducts post-pandemic as a result of both the moral licensing and cleansing effects, as well as a coping mechanism in fighting health-related anxiety, it is important to acknowledge other potential explanations for 14 individuals' less moral actions during the COVID-19 crisis, and more positive intended behaviors post-pandemic. For example, people might want to buy luxurious goods now instead of practicable goods because they are unable to spend money anywhere else (most likely in the control condition group), or want to support local businesses in this area of sales (in both the control and the experimental conditions).
In the current study, we used adapted moral and immoral prospective behaviors, adapted from previous literature within the moral licensing literature (e.g., a donation to charity, engagement in pro-environmental actions). However, future studies might benefit from using examples of moral behaviors related to the current situation the current situation (e.g., volunteering in a COVID-19 information center or helpline, helping senior citizens with their daily shopping in order to prevent exposure to infected areas), or immoral pandemic actions (e.g., to refuse isolation in order to show up for a job opportunity, even though infected with the virus, to refuse to help an old, lonely neighbor in need during the pandemic lockdown). It is also important to acknowledge that, given the stressful context of the pandemic, one might find it difficult to imagine the pandemic ending tomorrow, and this important limitation might be addressed in future studies (exploring similar contexts) by asking participants about their perspective on the matter, and to include their answers as a study variable.
To conclude, the present study found an increase in prospective moral behaviors and a reduction in prospective immoral actions in a hypothetical scenario in which they would be performed after the current coronavirus pandemic compared to the current lockdown circumstances. Cyberchondria emerged as significantly amplifying the moral effects of the current crisis, by making people more prone to display negative conducts during the pandemic and positive behaviors once it is over.