How Did We Forge a Bearable Reality in COVID-19? Biases in Social Comparative Judgments in a Time of Pervasive Threat

Abstract

In this chapter we present a very special form of social illusions: unrealistic pessimism (the expectation that negative events will happen to us in comparison to others who will be less affected by negative events like earthquakes or radioactive pollution). On that basis, we were interested in which illusions, which are associated with appearing in the moment of crisis, will appear during the eruption of the COVID-19 pandemic: that is, unrealistic optimism or pessimism (I perceive myself as less or more threatened than others) and the better-than-average effect.

It turned out that unrealistic optimism was present from the very beginning and did not disappear even when our respondents were delivered clear data (e.g., number of casualties, the rapid growth of cases) showing that all societies around the world were in great danger.

In this chapter we also present global studies showing how widespread this phenomenon is.

Interestingly, in our studies the better-than-average effect was also present in many countries: respondents perceived themselves as more involved in actions preventing COVID-19 infection whilst others were less involved.

2.1 Dynamics of Unrealistic Optimism in a Time of Pervasive Threat

In the previous chapter, we presented research on comparative biases: better than average and unrealistic optimism. These biases have proven to be widespread and strong, and the studies in which they were revealed have been replicated many times. It is worth noting, however, that these studies were conducted in situations (unrealistic optimism, UO) and general characteristics (better-than-average effect, BTAE) that can be described in the simplest terms as typical, or common: getting cancer, having a car accident, getting a lower-paying job, or experiencing unemployment. However, in human life, at least in some lives, there are situations that prove surprising and threatening to a very large portion of the population. Such events can include both natural disasters, such as floods, earthquakes, fires covering huge areas, volcanic eruptions, hurricanes, or locust plagues, and man-made disasters, such as wars, nuclear power plant disasters, or acts of terror such as the two planes hitting the World Trade Center towers. The question arises whether in just such conditions, where the threat is new and at the same time pervasive, people also feel privileged when they compare their threat with that of others who find themselves in analogous conditions. In such emergencies, do we also introduce a censor falsifying our world to make it more bearable, or even optimistic?

There is very little research on this issue. This is because most such events occur completely unexpectedly, and as a result, psychologists are unable to prepare their studies in advance. After all, psychologists didn’t know that terrorists would attack the World Trade Center or that Russia would attack Ukraine, and news of impending tragedies such as a tsunami, the bursting of a dam on a major river, or a hurricane usually comes only hours ahead of the actual occurrence of the tragedy. There is, however, anecdotal data in situations where the research apparatus has been prepared and studies conducted during such acute events.

So when, on the night of April 25–26, 1986, the reactor core overheated at the Chernobyl nuclear power plant (in what was then the Soviet Union, today Ukraine), followed by a hydrogen explosion and the spread of radioactive substances over a growing area, psychologists managed to react in time. This is because the cloud containing radioactive substances moved relatively slowly to the north and west of Europe. This made it possible to study the phenomenon of unrealistic optimism in Poland (Ukraine’s western neighbor, which was the first to be affected by the contamination; our place of birth and residence), even before the radioactive cloud reached that country.

One of the authors of this book – Dariusz Doliński – together with colleagues (Doliński et al., 1987) observed that the effect of unrealistic optimism previously noted by psychologists in cases where the threats about which the respondents were asked concerned individual people (a heart attack affects one person, as does cancer, a traffic accident affects at most a few people) rather than large populations (it is not the case that a mass of people or a whole society suddenly has a heart attack, gets cancer, or has an accident). Moreover, the Polish researchers pointed out that, to date, the threat had been of a prospective, rather than real nature: hitherto, a hypothetical onset of cancer or the occurrence of a hypothetical traffic accident had been analyzed. After all, cancer victims or accidents had not been studied. Finally, to date, known rather than unknown phenomena had been analyzed, e.g., becoming an alcoholic or a victim of a burglary, losing one’s job, or dying in a train crash is predictable; no one predicts a nuclear power plant explosion or a terrorist attack on the WTC until they happen. For this reason, it is difficult to study and measure something that is not there and cannot even be predicted. This time, the researchers wondered what would happen with unrealistic optimism, whether it would also appear in conditions where the threat is widespread, real, current, and entirely novel. All of these conditions were met by the radioactive contamination that was looming over Poland (the country where we live and experienced this tragedy): it was sudden, it had never occurred before – making it unpredictable – and it affected the entire population. For this reason, we also take a look at this unique study because COVID-19 and our research that we present in this book deal with a very similar situation: mass, unpredictable, threatening events. The possibility of falling ill as a result of elevated radioactivity, which was expected to increase dramatically in the coming days, is, from this perspective, very similar to the possibility of contracting COVID-19, where the probability escalates as the pandemic develops.

The radioactive cloud reached Opole, a 100,000-strong university town in southwestern Poland, on April 30, 1986. Studies on unrealistic optimism were conducted a few days later – on May 5 – at a time when contamination was still ongoing due to the breadth of the cloud. Residents of this city, university students, were asked to compare the likelihood of them and the average student of their sex experiencing various potential negative events of a predictable, individual nature: getting cancer, experiencing a heart attack, being the victim of a burglary in their own home, or being the victim of a traffic accident. Here we noted, repeatedly found in psychological research and discussed in detail by us in the previous chapter, the effect of unrealistic optimism. Our participants thought that the bad things mentioned would happen to others rather than to them.

However, and of particular importance, the participants were also asked to compare the probabilities of contracting radiation sickness (i.e., a sudden, unknown, current event, rather than one that may occur in the nebulous future) over the next month, year, and 5 years, relative to themselves and the average student at this university of their sex (the question of sex will return in our present research on COVID-19). While the two measures did not differ too strongly when they estimated the probability of getting sick within the next month, for the longer term (1 year and especially 5 years), the two estimates already differed markedly. Crucially, however, most of the participants considered themselves more likely to get sick than the average student. So it was quite different from previous studies of unrealistic optimism, in which people compare a threat to themselves with a threat to others. We called this effect unrealistic pessimism, because, to reiterate, an inverted trend was noted: the respondents considered themselves more threatened! In the next chapter, we will return to this research in a new context by showing unrealistic optimism and pessimism in the context of whether they help maintain health or ruin it.

We note here at once what we have already emphasized several times above in discussing BTAE and UO: this bias, too, can only be spoken of in a group sense, not in an individual sense. This is because, no doubt, some people may have had rational reasons for believing that they were particularly vulnerable to getting sick: they often went outside, subjecting themselves to more radiation emanating from the radioactive cloud, or they could not give up drinking fresh milk (which, as a result of contaminated feed, must have contained radiation), etc. However, if a clear majority considers itself more vulnerable to radiation sickness than the average other, this is – from a logical or mathematical perspective – just as impossible as the unrealistic optimism or better-than-average optimism we have analyzed earlier.

In order to confirm the existence of a new phenomenon or disprove it (it is always possible that a given study was conducted on a unique sample of people, a different continent, conducted in a different culture of a communist country, the analyzed event was unique and does not replicate with other sudden, massive and unpredictable events, etc.), a few years later, Burger and Palmer (1992) decided to use an event that took place in 1989. This was the earthquake in California, which, from the perspective of interest here, was very similar to the Chernobyl nuclear power plant explosion: it was new, unexpected, and therefore surprising; it was unknown if and when it would come; and it affected an entire society on an unprecedented scale. As in our study, 3 days after the tragic earthquake occurred, researchers asked people how likely they were to become victims of a similar natural disaster in the future and how likely it was to happen to another average person. It turned out that in this case, too, the people surveyed revealed unrealistic pessimism, indicating the universality of the claim that in such situations people think of themselves as more vulnerable than the average person. Burger and Palmer (1987), however, went a step further and repeated this research 3 months after the aforementioned tragic event. As it turned out, the pattern of results this time was very different! People now demonstrated a distinct unrealistic optimism: they believed that an earthquake was more likely to happen to someone else than to them (see Table 2.1). Interestingly, an analogous return of (unrealistic) optimism also occurred a few weeks after the explosion of the Chernobyl power plant (Dolinski & Gromski, 1987), when the majority of students demonstrated the belief that others, rather than them, would fall ill in the future from diseases caused by elevated radioactivity.

Table 2.1 Mean likelihood estimates for being seriously hart in a natural disaster, such as an earthquake

This means that the workings of the falsifying internal censor can be multifarious (the issue of stability over time of comparative illusions will return in subsequent chapters, where we will present studies in which we tested changes in cognitive biases in the context of the sudden and unknown threat emanating from the COVID-19 pandemic).

Before proceeding further (approaching the analysis of cognitive biases in the COVID-19 pandemic), let us briefly summarize the three phenomena described: better than average, unrealistic optimism, and unrealistic pessimism. Everything appears to indicate that the normal (typical) state is a positive illusion, which allows people not to experience constant anxiety and not to worry about their own uncertain future. However, if there is a sudden, unexpected, and widespread threat, such an illusion could be dangerous, because it would prevent people from taking actions aimed at avoiding this threat or, if that is not possible, at least at limiting its negative effects (as we write more about in Sect. 3.1 showing that pessimism evoked a healthy, active desire to protect one’s own health). As it turns out, under the influence of experiencing anxiety, people’s positive optimistic illusions disappear. Dewberry and Richardson (1990) studied their students either in a neutral situation or just before a difficult and important exam. Obviously, in the latter situation, the students were experiencing anxiety (which, by the way, the researchers verified with the help of self-report surveys). It turned out that in a state of anxiety, the participants ceased believing that they were less vulnerable than others to various negative events, such as being mugged or having an obesity-induced heart attack. Helweg-Larsen (1999), on the other hand, demonstrated the absence of positive illusions after people experienced the 1994 California earthquake. Apparently, this was a very powerful experience for those who lived through it, because even 5 months after the event just thinking about it caused such pronounced anxiety that the effect of unrealistic optimism was still not present.

Fortunately for us as a society, events such as the nuclear power plant explosion (1986), the California earthquake (1989), and the terrorist attack on the World Trade Center in New York (2001), unexplored at the time in the context of unrealistic optimism, took place decades ago, sparing us from experiencing similar tragedies. The absence of such negative events is – fortunately! – trouble for scientists who, wishing to study the cognitive biases that falsify reality in the case of such tragedies, have no field, no theater to study, and no material to publish. Of course, it is not as if scientists are waiting longingly for such tragedies to occur. However, if they do occur, they become an opportunity to conduct unique research. The emergence of the coronavirus pandemic in 2019 provided just such a situation.

Therefore, one could expect that the previously presented full dynamics of comparative judgments (from unrealistic pessimism to optimism) would be present with regard to the coronavirus pandemic, as it fulfilled all the criteria of the Chernobyl nuclear explosion situation and the California earthquake: it was sudden, unknown, affected the entire population, and concerned the present (and not the future as is the case with most studies of unrealistic optimism, the probabilities of heart attack, alcoholism, or cancer). We expected that in the first phase, when people were just anticipating that the pandemic would also appear in their country, they would react with pronounced unrealistic pessimism. Over time, this effect would diminish, followed by unrealistic optimism, which would continue throughout. The looming deadly tragedy also opened up further areas that had not yet been explored. For example, is it possible to modify unrealistic optimism? Is it beneficial – does it encourage vaccination? Or, by falsifying reality, does it calm us and reduce stress, but at the same time does it do harm by causing us, like the Titanic, to wash over an iceberg sunk in a blissful fog of intoxicating positive thinking? The answers to questions may be crucial to predicting human behavior during any future pandemics or other sudden, large-scale emergencies.

During the COVID-19 pandemic, we wanted to test a new, hitherto unexplored falsification of reality by scrutinizing the functioning of the censor, who had to confront an influx of unambiguous epidemiological data and explicit media messages. Note that if there is not a single case of a given infectious disease in a given area, it is reasonable to assume that a given social group believes that the risk of contracting the disease is low. For example, it is difficult to expect Poles, Americans, or Australians to fear the localized, yet highly contagious and dangerous, Ebola virus. Using the research experience one of us obtained from studying the radioactive contamination that resulted from the Chernobyl power plant accident, we decided to capture in our research the moment before the actual threat of contamination and the moment of the advent of the threat and, moreover, to look at how the dynamics of comparative judgments change as the threat becomes progressively greater.

In our first study, we captured the initial phase of the threat by conducting measurements of unrealistic expectations (we deliberately avoid unrealistic optimism, as we could not exclude the possibility that unrealistic pessimism would follow) before reports of the diagnosis of the first case of COVID-19 in Poland (i.e., the country where we conducted this research; Dolinski et al. 2020). We make no secret of the fact that we were quite lucky in doing so, since the first phase of our research on the student population took place on March 2 and 3, 2020, while the government notified Polish citizens of the first identified case the following day (March 4). So we captured the moment of analyzing a risk when it did not yet exist in a particular region of the world. Importantly, our research continued. We thus decided to study a different group of people immediately after the aforementioned press release (i.e., on March 5 and 6). In the following days, there were new reports of COVID-19 cases, illustrating the rapid spread of the disease throughout the country. This prompted us to conduct the next phase of the study on March 9 and 10. In doing so, we intended to conduct successive waves of research, but because of the pandemic situation, all Polish universities were closed. Thus, the possibility of engaging students was limited.

It is worth explaining at this point that we decided to use a slightly different procedure for determining the level of unrealistic pessimism/optimism than Weinstein (1980) did in his pioneering research. This is because we elected to study the estimation of the risks that the participants believed they were incurring and the risks that they believed the average person was incurring on two separate scales. By doing so, the dynamics of comparative judgments could be more precisely tracked. If, for example, there was an increase in unrealistic optimism over a certain period of time, then, using the methodology we chose, it was possible to determine whether this change was due to the participant’s belief that the threat to them had decreased or whether it was due to the fact that the threat to others had increased (or the fact that both processes occurred at the same time). The use of only a single measure, as proposed by Weinstein and often used in studies of unrealistic optimism, does not offer this possibility. If a participant in a study says, for example, that they are less likely to get sick than 75% of the people in their group and 2 weeks later estimates that they are less likely to get sick than 90% of such people, we will know that their level of optimism has increased, but we will not know what the nature of the change is. Did our respondent feel more secure on their own, did they judge that others are now more at risk than before, or did both processes occur together? We also made an analogous methodological assumption for all of our other research that we present in this book, enriching the literature on the subject with not only a more accurate measurement but allowing for a more nuanced analysis.

What were the results of our study? Quite unexpectedly for us, there was unrealistic optimism (rather than the unrealistic pessimism we had expected) in the first phase, which was, after all, implemented before the pandemic appeared in Poland. It persisted in the two subsequent measurements as well, that is, when the threat was becoming real and widespread! Since the dynamics of this bias were slightly different for men and women (and again we highlight sex differences, which we will discuss in more detail shortly), in Figs. 2.1 and 2.2, we present the results separately for both sexes.

Fig. 2.1

Women’s estimates of the likelihood of contracting coronavirus and the likelihood that an average student of the same sex will be infected at three different time intervals

Source: Journal of Clinical Medicine, 9, 1464, Figure 1

Copyright: MDPI

Fig. 2.2

Men’s estimates of the likelihood of contracting coronavirus and the likelihood that an average student of the same sex will be infected at three different time intervals

Source: Journal of Clinical Medicine, 9, 1464, Figure 2

Copyright: MDPI

The student population is certainly quite specific (education, access to expert information, residing in a large city, relatively low age of the participants, life status generally associated with not having to earn a living, often living with parents), and the fact that it is most often studied by psychologists is due solely to the easy accessibility of people from this group to researchers. However, a legitimate question arises: would a similar effect occur with another social group whose characteristics are closer to those of the general population?

As a team, we asked ourselves these questions at about the same time. We decided to approach the employees of a multinational company, and since this time we adopted an online research formula, our study was not interrupted by a lockdown; the study was safe for both the respondents and ourselves, the researchers (who worked remotely); and we could conduct a longitudinal study (12 waves conducted from early March of 2020, when vaccines were unavailable in Poland, through early March of 2021, during the period when vaccines became available) for an extended period of time regardless of changes in external circumstances. Of particular importance is that we conducted our study on the same individuals, so the possibility of (lack of) variability over time could not be attributed to different participants being measured at different intervals. In this case, too, we had time to start the study before the first case of COVID-19 was identified in Poland (Izydorczak et al., 2022). As can be seen in Fig. 2.3, employees of the corporation, like students, also reacted consistently – albeit with varying degrees of intensity – with unrealistic optimism. Replication in two independent studies of the same result thus demonstrates the astonishing strength of the effect we noted: with an influx of objective – and threatening! – information, the falsifying censor continued its work! What’s more, these were most likely the world’s first studies of unrealistic optimism that examined potential variability over time to demonstrate how stable this effect is.

Fig. 2.3

Line plot of changes in risk estimates over time

Source: PLoS ONE 17(12): e0278045, Figure 4

Copyryght: PLOS ONE

Note: Each dot represents mean risk estimates for “Self” (blue) and “Other” (red) at a given time. Bars represent standard errors of means. Frames above the graph describe the most important events in the timeline of the pandemic

The aforementioned results and changes in the strength of unrealistic optimism (but not its eradication!) prompted us to investigate post facto what could have caused them. We took two factors into consideration. The first was the objective data on the statistics regularly presented in the media: newly reported cases of COVID-19 and deaths attributable to it. Note that the graph of the number of fatal cases due to coronavirus infection – which often increases not linearly, but rather as an exponential/quadratic function – clearly communicates a deadly threat, while a steady decrease offers reasonable reassurance that the threat is diminishing.

The second factor that may have influenced the variable intensity of the falsification of reality was the Polish government’s decisions to introduce or rescind various restrictions. It could have been the case that the tightening of restrictions on social functioning (closing schools, limiting the number of people in stores, limiting the availability of certain medical services) certainly sent a signal – in the case of an increase in restrictions – of a serious and growing threat; if restrictions were being relaxed, then the clear signal being sent was good news that the threat was being reduced.

The results were astonishing, revealing the sheer dramatic irrationality of our judgments about the threat: the censor was closing its eyes! For it turned out that information on the number of identified cases of new illnesses and deaths was in no way correlated with the level of unrealistic optimism. In other words, such a clear indicator as death, an upsurge in queues at funeral homes, or information that someone in our neighborhood died due to COVID-19 does not have an effect on changes in unrealistic optimism. In contrast, the restrictions that were introduced and withdrawn had a very serious effect on the judgments formulated by participants. When restrictions eased, unrealistic optimism increased; when they escalated, unrealistic optimism decreased (but, we should emphasize, never disappeared completely). Characteristically, the introduction and withdrawal of restrictions mainly changed beliefs about the risk to oneself, and it was this element (rather than the risk to others) that influenced changes in the level of unrealistic optimism. Thus, an important message emerges for policy-makers around health issues: during a pandemic (and perhaps in other sudden global emergencies) the more effective approach is to reduce the scale of falsification of reality by showing changes in the social environment. Various restrictions and limitations are a signal to the public that the situation is dangerous. Unjustified optimism (the illusion that nothing bad will happen) should be abandoned, and the recommendations should be complied with.

During the first wave of the COVID-19 pandemic, studies of unrealistic optimism were also conducted in four different European countries (France, Italy, Switzerland, and the United Kingdom) with a very large sample of a total of 12,378 people (McColl et al., 2022). These surveys were conducted three times: February 12–21, March 11–12, and March 31–April 5, 2020. This made it possible not only to examine unrealistic optimism itself in three different time periods but also to trace the dynamics of this bias from its absolute beginnings to a pandemic reaping a hefty death toll. Interestingly, they found that the level of unrealistic optimism consistently increased in all four countries studied. Taking a closer look at this effect, the researchers found that it was due to the fact that participants thought they themselves were becoming less and less likely to contract the disease, while others were consistently just as likely. A little later, in late April to mid-June 2020 (and thus during the first lockdown announced in Belgium and the Netherlands), research on risk perception was conducted by Vera Hoorens and colleagues (2022). This study, too, found clear unrealistic optimism; taking all this data together shows how strong, robust, and global this social bias is.

The obvious question that arises is why the dynamics of comparative judgments in assessing the future risk of disease are different here than those observed in the aforementioned cases of a nuclear power plant explosion or surviving an earthquake. In particular, why, at the very beginning of the study conducted in Poland (when the first case of the disease in this country had yet to even be announced), no unrealistic pessimism effect was found. We think the critical point here is that people had already been bombarded many weeks earlier in the press with various reports about the pandemic, which was steadily and dynamically spreading around the world. It is highly likely that a conviction taking the form of unrealistic pessimism emerged just then, i.e., a few or perhaps several weeks before the first case of the disease in Poland was announced. By the time we began our study, the comparative pessimism of the participants had already given way to comparative optimism. Of course, this is only a hypothesis that can be neither confirmed nor refuted.

One might ask whether the correlations we have found are not confined to Poland or, somewhat more broadly, to a region of the world directly affected by the Iron Curtain, living for decades under the threat of a nuclear war. Later in this volume, we will show that this is not the case by presenting research from the Americas; now we will present our research conducted both in Poland and outside Europe – in Iran and Kazakhstan (Study #1; Kulesza et al., 2021). Let us emphasize how different the compared countries were. Iran is a country under severe embargoes resulting in a weakened healthcare system and response speed and a country governed by laws derived directly from religion (Islam). Poland, an almost entirely Christian, democratic country, was until recently deprived of some of its rights to self-determination on account of being under the influence of the Soviet Union. Kazakhstan lies more or less between the described extremes. Although Islam is present, it is also no stranger to the influence of other religions. The country was also communist until recently, but with weaker democratic inclinations than Poland. The axis of comparison may also be COVID-19 itself, which hit Iran much harder and faster than Poland or Kazakhstan. Unrealistic optimism emerged and persisted in subsequent waves of research (Izydorczak et al., 2022) in all these populations, with stronger optimism in Asian countries (Iran, Kazakhstan) than in the European country (Poland). This once again demonstrates the robustness of this effect.

It should also be emphasized that this bias has also been found in many other studies conducted in countries around the world (beyond the European continent as well) – e.g., in the United States (Salgado & Berntsen, 2021; Sjastad & Van Bavel, 2021), the United Kingdom (Asimakopoulou et al., 2020; Salgado & Berntsen, 2021), Germany (Kupier-Smith et al., 2021), Brazil (Vieites et al., 2021), Poland (Maksim et al., 2022; Kulesza et al., 2020, 2023a), Portugal (Figueiras et al., 2022), Italy (Druică et al., 2020), and Romania (Druică et al., 2020; Maftei & Petroi, 2022).

At this point, it is worth mentioning an interesting effect demonstrated by Kupier-Smith et al. (2020) who conducted a study on March 16, 2020, in three countries (the United Kingdom, Germany, and the United States). They asked participants not only about the likelihood that they and another average person would become ill with COVID-19 in the next 2 weeks, 2 months, 1 year, and lifetime but also about the likelihood that they would transmit the infection to others if they themselves fell ill. Of course, the researchers also asked about the likelihood that the average person infected with the virus would infect other people. These potentially infected people were divided into specific categories: family members, friends, acquaintances, fellow commuters, and strangers with whom one spends leisure time. It turned out that the study participants believed they would be less likely to infect others than that the disease would be spread by the average person. This belief was manifested in relation to all the categories of “other people” mentioned. Interestingly, the strength of this effect was even greater than the strength of unrealistic optimism regarding the possibility of contracting COVID-19. This means that it is necessary for pandemic managers to take into account unrealistic optimism as a mechanism responsible not only for the increase in infections among those with this falsified worldview but, much more importantly, for the increase in transmissions to others (“I don’t get infected, others do”)!

A pandemic is not only a threat to our health but also a threat to other aspects of our living situation. People’s subjective fear of becoming infected has caused them to avoid visiting restaurants, cafes, bars, pubs, or nightclubs. People were also reluctant to organize weekend trips out of town or holiday tourist trips, which obviously translated into a significant reduction in the number of guests at hotels. The official introduction of lockdown further aggravated the situation for the entire HoReCa (hotels, restaurants, catering) industry. Many food service establishments, motels, and hotels faced bankruptcy, or at least (in more optimistic scenarios) a marked reduction in revenue. This inevitably involved the loss of some or all of the earnings and/or employment of personnel working in this industry. Note how evident the threat situation was. It was both an economic and epidemiological threat: this is a dangerous profession; the government is shutting down this business sector, which is a clear indicator of the danger. This raises the question of whether workers in this particular branch of the economy, particularly affected by the COVID-19 pandemic, are displaying unrealistic optimism about their work situation. Logic would dictate that they should not. They have received far too many signals about the looming danger.

To test this, we asked a group of waiters, cooks, and hotel employees questions about the possibility of losing their jobs due to the pandemic, as well as the chances of the same happening to the average person in their business and to their average fellow countryman (Dolinski et al., 2021). They were also asked what the probability was that the company they work for, an average company in the same industry, and an average company (regardless of industry) would go bankrupt due to the ongoing pandemic. As can be seen in Figs. 2.4 and 2.5, the respondents demonstrate strong unrealistic optimism in all these respects. One particularly striking aspect of their responses is that they seem to overlook the distinctive nature of the HoReCa industry. They feel not only privileged in relation to others from their own company but also in relation to the average person in their country. They also think that their company is less likely to go bankrupt than the average company doing the same business (which makes absolutely no sense, all restaurants in the industry were at risk, and it’s impossible that a particular respondent’s specific workplace is somehow specially protected!) but also less than the average company engaged in any kind of business in their country. Thus, it can be said that – completely contrary to the objective circumstances accessible to this group of workers – they did not think that a pandemic would wreak any particular havoc in specifically the HoReCa industry. In a sense, then, one can conclude that the workers we studied manifested dual positive illusions – about both themselves and the industry that employs them.

Fig. 2.4

Assesment/comparisons for loosing job

Source: Sustainability, 13, 12562, Figure 3

Copyright: MDPI

Note: Bars represent mean values; error bars represent standard error of mean

*p < 0.05; ***p < 0.001

Fig. 2.5

Assesment/comparisons for bankruptcy in the industry

Source: Sustainability, 13, 12562, Figure 4

Copyright: MDPI

Note: Bars represent mean values; error bars represent standard error of mean

***p < 0.001

As we have repeatedly noted, people in studies on unrealistic optimism compare themselves with those who are similar to them (e.g., students compare themselves with the average student of their sex from their university, people surveyed via online panels with the average user of the panel, employees of a particular company with the average employee of that same company, etc.). Of course, only then can one legitimately employ the term “unrealistic optimism.” After all, if, for example, a 20-year-old healthy student believes that they are less likely to be severely affected by COVID-19 than the average resident of their country (and therefore a person much older than them), then, statistically speaking, they are right: they are younger, healthier, and probably from a wealthier home with better access to resources.

For a number of reasons, however, we may also take an interest in situations where people compare themselves not only with the average person from their own group but also with average people from other but similar groups. In one of our studies (Kulesza et al., 2022a), we created just such a situation by extending the circle of comparisons and retaining the previous, classical comparisons to see if the reference point thus measured changes the distribution of results. We conducted this study via the Prolific panel in October 2021, when COVID-19 vaccines were already widely available, with 660 participants from different countries. The participants were compensated financially (GPB 11.50 per hour of activity). As before, the participants were asked a series of questions regarding the probability of various people contracting COVID-19. On a scale of 1 (absolutely impossible) to 11 (quite certain), the respondents determined this probability in relation to themselves, the average Prolific panel user, the average vaccinated Prolific user, and the average unvaccinated Prolific user. Respondents also answered the question of whether they themselves were vaccinated or not.

As for the respondents who were themselves vaccinated, the results were rather predictable. They believed that they were less likely to contract COVID-19 than the average Prolific user, as well as the average vaccinated and unvaccinated user of the platform. At the same time, they believed that the average vaccinated Prolific user was less likely to contract the disease than one who had not been vaccinated. On the other hand, the results regarding participants who had not been vaccinated were remarkably interesting and surprising. They believed that they were less likely to get sick than the average Prolific user and less likely than those users of this panel who had not been vaccinated. Interestingly, however, at the same time, they believed that they were at the same risk of contracting the disease as the average Prolific user who had been vaccinated, while expressing the opinion that a Prolific user who had been vaccinated would become ill with COVID at a lower probability than one who had not been vaccinated. We present the results of this study in Fig. 2.6.

Fig. 2.6

Estimated COVID-19 risk of infection as a function of vaccinated and unvaccinated participants

Source: Journal of Pacific Rim Psychology, 16. https://doi.org/10.1177/18344909221122573. Figure 1

Copyright: SAGE

Note: Bars represent mean values; error bars represent standard error of mean

*** p < 0.001; ** p < 0.01; ns not significant (p > 0.05)

This means, therefore, that those who are unvaccinated grasp the importance of and need for vaccination, but not in relation to themselves. Others who have been vaccinated are therefore, in their – unvaccinated – opinion, in the right, because by doing so they reduce the likelihood of contracting the disease. They themselves, on the other hand, do not need to vaccinate, because even without vaccination they have just as low a risk of getting sick as those who have been vaccinated. This result thus illustrates a triple intellectual Nelson: while being against vaccination, we believe in vaccines – they lower the probability of getting sick. People who do not vaccinate are making a mistake, because they become more likely to get sick. But I am not affected by this correlation: without vaccination I am as immune to getting sick as those who have been vaccinated.

Although we did not investigate this (essentially because we did not expect to see such a shocking result), it is a reasonable supposition that unvaccinated respondents believe that, for some reason (e.g., genetic), they are less likely to get sick than other people. Either way, this entirely unexpected result shows that the decision not to vaccinate oneself is not necessarily the same as a negative assessment of vaccination per se!

Another example of comparing one’s chances not only with those of one’s own group, but also with those of another group, can be estimating the probability of contracting the disease for oneself, for the average person of the same sex, and for the average person of the other sex. We already know from the various studies reported earlier that most men believe they are less likely to contract COVID-19 than the average man and most women believe they are less likely to contract the disease than the average woman. But will men think they are less at risk than the average woman, and will women believe they are less at risk than the average man?

The issue of such comparisons is interesting insofar as the literature in the field of health psychology contains a wealth of data on cross-sex differences in risk misjudgment (see Byrnes et al., 1999; Courtenay, 2000 for review). This includes, for example, issues such as the consequences of smoking (Weiss & Garbanati, 2006), driving after drinking alcohol (Linkenbach & Perkins, 2005) (see also Dillard et al., 2009), and the wisdom of undergoing preventive screenings (Mahalik et al., 2006). Of particular relevance from the perspective of interest to us in this book, sex differences were also noted in studies conducted on people’s responses during the COVID-19 pandemic, revealing that men were less likely than women to engage in various preventive measures and to follow medical recommendations (e.g., Aranguren, 2022; Lin et al., 2021).

We therefore decided to take a closer look at the topic of comparisons with the average person of the other sex. We conducted three separate studies (in the period from September to November 2021) on a population of American users of the Prolific platform (Kulesza et al., 2023b). These studies differed in several details, but their common feature was that respondents were asked to rate on an 11-point scale the likelihood that they themselves would become ill with COVID-19 and that it would happen to the average Prolific profile user, the average female Prolific user, and the average male Prolific user. The results of the three studies are presented in Table 2.2.

Table 2.2 Risk estimations for all groups in the three individual studies

As we can see, both women and men here demonstrate a strong effect of unrealistic optimism (they believe that they are less likely to get sick than the average user of the Prolific platform of their sex), and they also believe that they are less likely to get sick than the average Prolific user of the opposite sex or the average user of the platform regardless of their sex. While the strength of unrealistic optimism was the same for men and women, the participants’ beliefs that they are less likely to get sick than the average person of the opposite sex were not equally strong – women were more strongly convinced of this than men.

Let us briefly summarize two recent studies. We note that previous studies on comparative risk estimates have dealt with the comparisons participants make with an average person from their own group. Our studies have shown the presence of a positive illusion effect also when people compared themselves with an average person from another group – vaccinated versus unvaccinated and vice versa, women versus men and vice versa. These results are also relevant to global health policies. They should take into account the universality and strength of positive illusions.

To sum up this section, empirical studies consistently demonstrate that unrealistic optimism is strong and widespread, not only in neutral situations but also during a pandemic. Moreover, it is resistant to clear signals indicating the gravity of the situation (clear and present danger). It is present in all the countries where it was studied, showing that it is not related to the culture or history of a given country nor the background of the study participants.

2.2 The Better-than-Average Effect in a Time of Pervasive Threat

We know from the previous section that unrealistic optimism is at work in the COVID-19 pandemic: research participants from many countries unanimously declare that it is others who will get sick, not them (which, of course, makes no sense; if so many people say they won’t get sick, then who will?). Meanwhile, as we already know from Sect. 1.3, unrealistic optimism is not the only means of falsifying reality deployed by the internal censors responsible for our self-image. Another method is the belief that we are, for some reason, better than others. So let’s look at whether the censor triggers better-than-average thinking under conditions of a real, massive, and serious threat – during the COVID-19 pandemic.

The first of our studies to address this issue was conducted online on student populations in three countries: Iran, Kazakhstan, and Poland (Kulesza et al., 2022a, b). The surveyed students were asked to indicate on a scale from 1 (do not follow) to 11 (fully follow) to what extent they follow the recommendations just announced in each country for the functioning of daily life during the pandemic. The comparisons involved issues related to the medical recommendations that a given country’s citizens can follow to fight the pandemic: wearing masks, maintaining social distance, washing hands frequently with soap and water and disinfecting them, etc. In order to see whether participants see themselves in a better light than other people similar to themselves, we asked them, using the same scale, to declare the extent to which the average student of their sex from their country follows such recommendations and then – in order to estimate the possible magnitude of differences glorifying oneself – to estimate to what extent the average resident of their country does so. We repeated this survey with the same people approximately 10 days later.

As we can see, we registered a classic better-than-average effect. The participants declared that they behaved more correctly than the average student of their sex, and they saw an even greater difference in this regard between themselves and the average resident of their own country. Interestingly, the passage of a mere 10 days reduced the participants’ declared belief that they were conscientiously following the recommendations, but similarly, it also reduced it to the same extent for the average student and the average resident of their own country (they, too, according to the respondents, began following the recommendations to a lesser extent than 10 days earlier). Thus, the magnitude of the better-than-average effect remained constant.

An obvious weakness of the above study is that the student population we surveyed is quite specific for many reasons (age, education, place of residence, leisure activities, etc.). A major methodological caveat flows from this finding. It would be unwise to construct recommendations for an entire population on the basis of a study of a specific part of it.

We therefore decided in a subsequent study to examine members of a different population that would be closer to typical average people (Kulesza et al., 2022b, “social psychology journal”). For the study, we selected subscribers to a popular online university newsletter, which features texts and audio broadcasts popularizing scientific knowledge. Since the newsletter appeared in Polish, we were limited to Polish residents. Because of the strict lockdown implemented at the time, the survey was conducted online.

Several hundred such subscribers who agreed to take part in our study were asked – in a similar fashion to the previous survey – first to estimate to what extent they themselves comply with medical recommendations and then to say to what extent they think the average university newsletter subscriber does so; finally, we called on them to estimate the extent to which the average Polish citizen follows such recommendations. As expected, the participants thought they were the ones who behaved the most appropriately; the average newsletter subscriber was less disciplined than them; and the average Pole adhered to the recommendations the least. The better-than-average effect noted in our first study has thus been replicated. Although we still do not have representative data for the entire population, it can already be assumed with greater probability that the pattern of results is typical of residents of the country where we conducted the research (i.e., Poland). However, there is still another problem: perhaps it is only the residents of a particular region of the world who are susceptible to this kind of censorious falsification of reality? In discussing unrealistic optimism, we showed that this is not the case. In the case of better than average, however, further investigations were needed to test the universality of the recorded effects.

Fortunately, studies on the occurrence of the better-than-average effect under pandemic conditions have been conducted in many countries around the world. For example, Kim and Han (2022), surveying 210 respondents from the United States and 214 from South Korea, indicated the prevalence of this effect in both countries; they likewise obtained a result demonstrating the universality of the human tendency to falsify reality. In doing so, the researchers showed that the magnitude of the better-than-average effect correlated with negative attitudes toward people who do not follow medical advice during a pandemic. This result can be interpreted in terms of grievances: “my efforts to stop the pandemic are being thwarted by irresponsible people who totally ignore medical recommendations,” which points to a potentially important conclusion for global health practitioners. Think about building narratives aimed at showing the active involvement of a social group in the fight against a pandemic. After all, consider that we have already shown multiple times that the core cognitive biases, the falsification of reality, are comparisons to the group: it is impossible for “everyone to be superior to the average group member.” The key, then, may be a narrative about group behaviors to change individuals’ perceptions of themselves: you are not necessarily less likely to get sick than others (unrealistic optimism), and you are not necessarily more engaged in protecting yourself than others (better than average). We will return to testing such narratives as a way of curbing the activity of our internal censors, falsifying reality, in Sect. 4.3, where we present a study in which we tested exactly such a method.

Respondents’ belief that they behave better, more responsibly, than other people during a pandemic was also tested in European studies conducted in Belgium and the Netherlands (Hoorens et al., 2022). In these studies, participants answered a series of questions on issues such as the medical authorities’ recommendation to wash their hands with soap and water more often and more thoroughly than usual and to maintain physical distance from others; they also commented on the frequency of behaviors that are clearly not recommended by medical authorities, such as leaving home for fun, meeting relatives and friends, visiting other people in their homes, and running an errand that is not necessary. The research participants were asked to estimate how often they themselves exhibit such behavior and how often the average person of their age and sex does so. A clear better-than-average effect was noted. Evidently, the Belgians and Dutch they surveyed were convinced that they themselves were more likely to follow medical advice than their fellow countrymen. This effect was particularly pronounced with older (vs. younger) respondents. The researchers also asked the participants about their predictions for the aforementioned behaviors in the near future (how things will be next month). It turned out that while the better-than-average effect also surfaced then, its strength was less.

In doing so, let us note that the predictions of the Belgian and Dutch respondents as the situation in a month’s time coincided exactly with the results of the real-time estimation studies we presented above, conducted in Iran, Kazakhstan, and Poland by the first author of this monograph and his colleagues (Kulesza et al., 2022a, b). Perhaps people realize that, under the influence of an emerging threat, they begin to engage in behaviors recommended by authority figures quite intensively, but they habituate existing threats relatively quickly, and the aforementioned avoidance-oriented activities decrease in intensity. This hypothesis may give rise to further practical recommendations. Perhaps changes in narratives that reduce the sense of being superior are needed to avoid weakening the impact of such messages in the long term.

Very interesting results linked to the better-than-average effect (and providing further evidence that the better-than-average effect is not restricted to just one country or residents of one area of Europe) were obtained by Salgado and Berntsen in a study conducted on the UK population (Salgado & Berntsen, 2021). They asked participants about the likelihood in various situations that they would buy and wear masks, use hand sanitizers, and maintain physical distance from others and the likelihood that such behaviors would be exhibited by close others and acquaintances. They also asked questions about the extent to which it is necessary for themselves and close others and acquaintances to manifest these very behaviors in order to effectively guard against the threat of infection. It turned out that, in the respondents’ estimation, it is equally necessary for everyone to manifest the aforementioned behaviors, but they themselves exhibit them at a slightly higher frequency than close others, and at a significantly higher frequency than acquaintances – see Fig. 2.7.

Fig. 2.7

Attitudes towards protective behaviour against COVID-19

Source: Journal of Applied Research in Memory and Cognition 10, p. 373

Copyright: Elsevier

Note: Left panels illustrate the necessity of using facemasks (A1), hand sanitizer (B1), and keeping a physical distance (C1). Right panels show the likelihood of buying and using facemasks (A2), hand sanitizer (B2), and the likelihood of keeping a physical distance (C2). All panels illustrate the data as a function of the target of the tasks: self, close other, and acquaintance. Bars denote the group means, and error bars represent 95% Confidence Intervals of the mean

The better-than-average effect was also revealed by Kupier-Smith et al. (2020), surveying 828 people from additional European countries (the United Kingdom and Germany) and the United States in March 2020. The participants expressed the belief that they take pandemic-related medical recommendations more seriously and responsibly than the average person. In particular, they believe that they reduce the number of direct interpersonal contacts to a greater extent than other people and are more careful than others about personal hygiene during this exceptional period. And here arises a practical conclusion analogous to the one we presented when discussing the research by Kim and Han (2022). In cases of ubiquitous threats, it seems sensible to make people realize that ... it’s simply impossible! As we have previously noted, we are afflicted by cognitive egocentrism: we are more aware of our own actions than of other people’s behavior. This causes us to falsify our self-image and that of others. Perhaps, then, active presentation of other people’s health-promoting actions in various media could reduce this cognitive bias. We will return to this issue in Chap. 4 with a thorough presentation of a study directly dedicated to this very issue.

The results presented in this chapter clearly indicate that we (i.e., people) believe that we are better than others about following medical recommendations designed to protect us from coronavirus. Thus, we manifest the classic illusion of a better-than-average effect. Not surprisingly, we consequently have a sense of being less likely to contract COVID-19, so we manifest another classic illusion in the form of unrealistic optimism. In other words, it is clear that we defend ourselves against an uncontrollable threat that is challenging to understand by employing a set of cognitive biases: they are double-edge swords cutting in unison, coherently, in multiple directions.

Having said this, the practical recommendations are clear: global health politics should aim at reducing not one but at least two social cognition biases. In the short term, they would take away procedures aimed to reducing fear, anxiety, and depression even but in a long run would prolong life and quality of life (since COVID-19 is extremely dangerous in terms of not only mortality but also in the long-term perspective which we present in the last chapter). Under conditions of real danger, one should be less concerned with well-being and stress reduction and more concerned with protecting human life. There will be time to augment well-being after the grave threat has subsided.