Mimicry boosts social bias: unrealistic optimism in a health prevention case

ABSTRACT Unrealistic optimism bias appears when a person perceives oneself – in comparison to peers – as less at risk from threats. This bias has been widely reported and the consequences are clear: it puts one’s health in danger. The existing body of literature proposes egocentrism as a mechanism leading to a reduction in this bias. The present paper tests a novel mechanism orienting a person toward others – thus linked with egocentrism – i.e., mimicry. Results showed directly opposing effects: mimicry induced a stronger tendency to perceive oneself as less threatened. This result is not only surprising but especially alarming since mimicry may be used in patient-doctor dialogue which may backfire, leading to resistance to medical recommendations provided by the doctor.

Patient adherence to medication is an obvious example of a situation in which scientists and medical doctors provide grounded help, but patients do not follow their recommendations, leading to the possibility of harm to their health, at the very least.A clear example of this problem is the COVID-19 pandemic: it is not only a medical problem, but, as is becoming increasingly clear, it is first and foremost a social problem.The best example in this context is vaccination against COVID-19 and vaccine hesitancy.Other examples include not following medical recommendations (maintaining social distance, hand sanitizing, avoiding large gatherings), and even engaging in spreading false and unchecked information.Much research concentrates on answering why this paradox is so powerful.One possible explanation is the unrealistic optimism bias -a feeling of being less threatened than others.From this reasoning, research on reducing this bias may build a groundwork for COVID-19 eradication.The present paper aimed to join this effort.

Unrealistic optimism bias and its (dis)functionality
When people estimate the probability of different life events, they 'believe that negative events are less likely to happen to them than to others, and they believe that positive CONTACT Paweł Muniak pmuniak@swps.edu.plevents are more likely to happen to them than to others' (Weinstein, 1980, p. 807).On the level of the individual, such an expectation may be fully grounded in facts since s/he might live a very safe, secure, and healthy life.This notion, however, becomes unrealistic when the vast majority of people present such an illusion.Decades of research have proven Weinstein's claim to be true, leading to the theory of unrealistic optimism (Taylor & Brown, 1988).This is a very useful mechanism for -mainly -coping with difficult and threatening situations.On the one hand, it has been shown, for example, that people estimate that risks pose a greater threat to others than to themselves, for example, having a car accident (McKenna, 1993;Rutter et al., 1998), divorce (Lin & Raghubir, 2005), developing heart disease (Prats et al., 2012) or developing addiction in general (Nezlek & Zebrowski, 2001), and specifically alcohol addiction (Kim & Niederdeppe, 2016).In the same vein, participants estimated risks of unintended pregnancy, engagement in alcoholrelated sexual activity, and sexually transmitted infections as higher for others than for themselves (Lopez & Leffingwell, 2020).On the other hand, chances for positive future outcomes are mainly attributed to oneself rather than to others.It was shown that, for example, students overestimate their chances for a higher salary after graduation (Shepperd et al., 1996), and before graduation on passing exams (Lewine & Sommers, 2016) in comparison to others.A more recent study revealed that hospitality employees at the entry-level position are optimistically biased about their promotion expectations (Ngan & Tze-Ngai Vong, 2019).This begs the question: why do we hold this distorted perception of reality and self-perception?In short: to cope with difficult and threatening situations.The research on this issue has shown that unrealistic optimism leads to a greater sense of personal control, which is especially beneficial in uncontrollable, and sudden situations (Taylor, 1983(Taylor, , 1989)).Other research also postulates that when such situations become reality, unrealistic optimism leads to reductions in fear, anxiety, and stress, which is hugely beneficial for one's well-being (Hoorens, 1995;Klein & Weinstein, 1997).

Unrealistic optimism bias in the global/public health domain
As mentioned above, people underestimate their chances of heart failure when evaluating risks for themselves (Prats et al., 2012).Research clearly shows that there is a high degree of unrealistic optimism bias in health evaluation and prevention, as well as medical assessments run by patients.It has been shown that patients opt for unnecessary, harmful, and counterproductive medical interventions by, on the one hand, overestimating chances for success in treatment, and, on the other hand, underestimating risks (Hanoch et al., 2018).Women who were unrealistically optimistic in estimating their risks for breast cancer in fact increased these risks by refraining -in comparison to unbiased women -from mammographic screening (McCaul et al., 1996).The same picture appears in research on lung cancer as a result of smoking.Dillard and colleagues (Dillard et al., 2006) showed that unrealistically optimistic smokers were less eager to quit smoking and underrated their probability of getting lung cancer, whilst this probability, in fact, increased.
In conclusion, the unrealistic optimism bias is a useful tool for maintaining psychological well-being in threatening situations like risk of a disease whilst the same mechanism may be harmful while looking at this issue from a medical perspective.Having said that, it is clear that developing research aimed at reducing or even eradicating this bias may be beneficial not only for the medical context, but also on a global level, such as countries and governments.The current paper presents research run during the COVID-19 pandemic aimed at reducing/eradicating this dangerous bias.

Unrealistic optimism bias in the COVID-19 era
It is absolutely natural that, at the present moment, some readers might feel more secure and safe against COVID-19 infection than others.For example, some of us might live in a place with almost no face-to-face social interactions, be fully vaccinated and recently boosted by the most recent combination of vaccines, be equipped with antiviral medicines preventing serious illness stemming from coronavirus, and finally wear face masks as well as sanitize our own hands.As mentioned above, such a notion becomes a distorted perception of reality when many people present the same biased selfaffirming views, thus proving the existence of the unrealistic optimism bias.
The new, but growing body of literature on the presence of this bias has proved its robustness during the COVID-19 pandemic.Starting from times when no vaccine was available or any cure/medicine was visible on the horizon in many countries, people estimated the risk for COVID-19 infection as higher for others than for themselves, replicating Weinstein's original concept.Its existence was reported in România (Druică et al., 2020), Italy, Switzerland, France, and the United Kingdom (Asimakopoulou et al., 2020;McColl et al., 2022), Belgium and Netherlands (Hoorens et al., 2022), the United States (Salvador Casara et al., 2022), Kazakhstan and Iran (Kulesza et al., 2021), and Poland (Dolinski et al., 2020).
The robustness of unrealistic optimism in COVID-19 risk perceptions was also proven by the fact that it persisted over time, proving its irrationality: availability of treatment changed, as well as the number of cases and deaths, lockdowns were placed and lifted, vaccines were not available, and later on massive vaccination drives took place, but biased perceptions of one's superiority remained regardless of the seriousness of the actual situation.In one study, unrealistic optimism was present in the country of research before the existence of the first case of COVID-19, immediately after the first cases were reported and the lockdown was enforced, and a few days later when the epidemic spread at a high rate (Dolinski et al., 2020).In a study by Kulesza et al. (2021), the bias was reported over the span of a month.Finally, one study reported the presence of this bias for a 12-month period (Izydorczak et al., 2022).

Unrealistic optimism reduction/eradication
From the perspective of threats stemming from an unrealistically optimistic perception of reality, it is obvious that researching mechanisms leading to reduction or even eradication may be highly beneficial from the perspective of one's health as well as global health and policies.Probably the most powerful mechanism responsible for reaching this goal is through the reduction of egocentrism.Since this mechanism is under the empirical scrutiny of the present paper, we will discuss this issue in more depth.Weinstein (1983) asked participants to list factors that might influence the probability of these negative events happening to them.One group was presented with a list generated by others, making participants less egocentric.In line with expectations, this manipulation led to a reduction in unrealistic optimism.In a more recent replication study that tested whether reducing egocentrism leads to a reduction in unrealistic optimism, short autobiographies of people affected by a health problem were given to the participants (Kim & Niederdeppe, 2016).Results were consistent with those previously described: in this group this social bias was reduced.
In a more complex design, and importantly one run during the COVID-19 pandemic, which became the basis for the experiment described in this paper, exposure to others as a decreased egocentrism mechanism for reducing unrealistic optimism was tested (Dolinski et al., 2022).In three experiments, participants were presented with media information depicting behaviors of others following (or not) medical recommendations aimed at curbing COVID-19 infections: mask wearing, hand sanitizing, maintaining social distance.
In the first experiment, participants were given a text ostensibly coming from a newspaper depicting (or not) the abovementioned behaviors.A reduction in the optimism bias was reported in the condition in which participants read about others following medical recommendations.In the second study, such a manipulation took place via video and, contrary to the previous results, unrealistic optimism was eradicated when people observed others not following medical recommendations.The third experiment addressed these differences by showing that reading articles possibly requires more effortful information processing than watching a video.
In conclusion, on the grounds of unrealistic optimism/pessimism, there are at least several gaps which our paper tries to fill.First and foremost, in the COVID-19 era we lack clear mechanisms to reduce unrealistic optimism bias which may backfire by fueling vaccination hesitancy.Secondly, on the one hand, in the domain of public/global health, unrealistic optimism has been proven to be extremely spread out, leading to harmful health consequences for the person holding this bias, but on the other hand, the body of literature presents almost no empirical recommendations on how to eradicate this bias.Taken together, the present paper fills an important gap in the lack of the empirical evidence for unrealistic optimism eradication in the specific context of the ongoing COVID-19 pandemic, as well as in the more global domain of public health.
The aforementioned review calls for a summary before the goal of the paper is described.Unrealistic bias may lead to repercussions for the bias holder's health.Researching mechanisms responsible for eradication/reduction seems therefore to be crucial.One of the best grounded and researched mechanisms is reducing egocentrism: exposure to others and their behaviors in relation to health-related risks.The aim of the present paper was to test a new mechanism in unrealistic optimism reduction, which has previously been shown to induce an orientation toward others (vs.oneself), that is, mimicry.
Mimicry is a mechanism widely present in many behaviors of animals.Birds imitate each other forming a 'V' shape to save energy during long distance flights (Voelkl et al., 2015).Fish also form schools, for example, to survive the attack of a predator (Parrish et al., 2002;Partridge et al., 1980;Pitcher, 1979).Mimicry is one of our -humans' -first forms of social behaviors, appearing hours after being born (Meltzoff & Moore, 1983).Below, social goals for mimicry are described.

Mimicry as a mechanism responsible for an "others orientation"
From the perspective of the present paper, it is important to stress that mimicry is tightly associated with an orientation toward others rather than oneself.
Across two decades of research it was shown that, for example, priming participants' attention toward others (showing words like 'we', 'together' vs. 'I', 'alone' leads to a greater tendency to perform mimicry (Van Baaren, Horgan, et al., 2004).Similarly, it was also shown that mimicry is elicited by prior exposure to a prosocial perspective (e.g., Leighton et al., 2010).People holding a high score on self-monitoring are more focused on their social surroundings as well as the behaviors of others (Snyder, 1974) and tend to engage more in mimicry.Discovering similarities to others (for example, in attitudes/ opinions; Castelli et al., 2009;Van Swol & Drury-Grogan, 2017) creates a greater tendency to mimic.
On the other hand, a reverse causal relationship was discovered: mimicry increases the orientation toward others, and prosocial perspective/tendencies.It was, for example, shown that mimicking increased participants' tendencies to provide help not only to the mimicker, but also to unrelated others (e.g., Van Baaren, Holland, et al., 2004).On a more general level, it was proven that mimicry shifted participants' views toward other-oriented political views (Stel & Harinck, 2011).In the same vein, after being mimicked people tend to perceive social relations as more just (Stel et al., 2013), and victims of violent crime as less to blame for it (Stel et al., 2012).
On a more general level, mimicry research has proposed the social glue hypothesis, which states that mimicry is responsible for starting and maintaining social relationships (Dijksterhuis, 2005;Lakin et al., 2003).

Goal of the paper
As outlined above it is clear that the call for research on reducing unrealistic optimism is necessary.The existing body of literature clearly shows risks stemming from this social bias and already proposed mechanisms highlight decreased egocentrism as a key factor in reducing or even eradicating the unrealistic optimism bias.The present paper joins this line of research by testing mimicry as a mechanism that orients mimicked participants toward others/social relations.From this perspective, we expect that mimicked participants will present a reduced or no unrealistic optimism bias.Since the presence of unrealistic optimism has been widely reported during the ongoing COVID-19 pandemic, we tested if exposure to mimicry would reduce the tendency to perceive oneself as less threatened by the coronavirus infection.
part in the experiment.The sample size was determined by feasibility criteria.We strove to recruit as many participants as possible, given the limited time and size of our sampling pool.Participants were randomly assigned to the two experimental conditions and did not receive any payment or course credit.No data were excluded from the analysis.

Procedure
Participants connected with the confederate via Google Meet, were welcomed and given all instructions.The procedure of the experiment was conducted by a female psychology student (in her early twenties) who was trained in verbal mimicry by the experimenter.The confederate was blind to the hypothesis but not to the experimental condition.
The apparent goal of the 10-minute interview was to gather opinions on the ongoing COVID-19 pandemic.For example, the confederate asked how the interviewee felt about the pandemic.During the interview, manipulation of verbal mimicry was introduced, where confederates mimicked words spoken/expressed by the participants.Following recommendations stemming from Kulesza et al. (2014) study in the control condition, the confederate responded simply to the participant by stating 'yes' and 'I understand' without mimicking the participant's words.
After the interview, participants responded (on the scale of 1= Absolutely impossible; 11= Quite certain) to two statements assessing the better than average bias: (Q 1 ) What is the probability that you will be infected with the novel coronavirus (SARS-CoV-2/COVID-19)?(Q 2 ) What is the probability that a friend of yours (of your age and gender) will become infected with the novel coronavirus (SARS-CoV-2/COVID-19)?Finally, the respondents indicated their gender, age, and a debriefing took place.

Results
A 2 × 2 mixed-design ANOVA with one between-subject factor -experimental condition (2: mimicry and no mimicry)-and one within-subject factor -unrealistic optimism bias (2: COVID-infection risk assessment for 'Me' and 'My friend') was run.

The interaction effect of the experimental conditions and unrealistic optimism bias
The interaction effect was not significant, F(1, 252) = 0.56; p = .457;η p 2 = >.0 (post-hoc test power: 1 -β = .55).However, according to Tybout et al. (2001) and Hayes (2005), it is appropriate to test simple effects even though the interaction effect fails to achieve significance.Thanks to this it will be possible to test our hypothesis which claims that in the mimicry condition, the unrealistic optimism bias will be reduced (or will be not present at all).Thus, we decided to perform exploratory post-hoc analysis with Bonferroni correction adjusted for comparing a family of all six estimates.This analysis revealed that unrealistic optimism was found in the mimicry condition.Participants perceived themselves (M = 6.51,SD = 2.8) at a lower risk of contracting COVID-19 than their friend of their age and gender (M = 7.2, SD = 2.4; t = −4.32,p bonf <.001; Cohen's d = −0.28,95% CI [−0.45, −0.1]).Unrealistic optimism was also found in the no mimicry condition.Participants perceived themselves (M = 5.3, SD = 2.4) at a lower risk of contracting COVID-19 than a friend of their age and gender (M = 6.15,SD = 2.28; t = −5.61,p bonf <.001; Cohen's d = −0.34,95% CI [0.51, −0.18]).
Statistically significant differences were also observed when estimating the self-risk assessment ('Me') between the mimicry and no mimicry condition.Participants in the mimicry condition perceived themselves at higher risk of contracting COVID-19 than participants in the no mimicry condition (t = 3.9, p bonf <.001; Cohen's d = 0.49, 95% CI [0.15, 0.83]).The same pattern was also observed when estimating the risk assessment for participants' friends of a similar age and gender as their own ('My friend').Participants in the mimicry condition perceived friends of their age and gender at higher risk of contracting COVID-19 than participants in the no mimicry condition (t = 3.37, p bonf =.005; Cohen's d = 0.42, 95% CI [0.09, 0.76]).
Data visualization can be found in Figure 1.For detailed descriptive statistics please see Table 1.Post-hoc comparisons with Bonferroni correction adjusted for comparing a family of six estimates are presented in Table 2.

Discussion
Participants in both the mimicry and no mimicry condition presented an unrealistic optimism bias.However, being verbally mimicked increases average risk perceptions of contracting COVID-19.Specifically, verbal mimicry makes people estimate that the risk is higher not only for themselves, but also for others (friends of their age and gender).
The question arises as to why the pattern of results is exactly the opposite than expected.In the field of mimicry research this unexpected pattern of results may have its roots in several sources.Firstly, in the introduction we summarized the body of literature on mimicry showing that mimicry has close links with orientation toward others.On this basis we stipulated that mimicry should lead to the reduction of egocentrism.In the light of the present data, it is possible that mimicry orients our social perception toward others, but at the same time does not lead to reduced egocentrism (which would be a key factor in the reduction of unrealistic optimism).As a result, the predicted relationship could not be supported, and from this perspective the reduction of unrealistic optimism could not take place since reduction of egocentrism did not take place, and this line of reasoning aligns with expectations of the pattern of results.To conclude, future studies should address the question of whether mimicry leads to reduced egocentrism or not.
Secondly, it is also possible that mimicry not only leads -as already mentioned above -to a strengthened self-perceived connection with others (Lakin et al., 2008), but also to a greater tendency to meet others in person.As a result of such meetings, the   expectation of contracting COVID-19 due to higher direct exposure to others is reasonable and grounded since direct interactions are an essential risk factor of COVID-19 infection.
Thirdly, on a more general level, the present experiment supports the under researched area of mimicry and costs stemming from this process.Anecdotal data show the costs stemming from mimicry.For example, mimicry leads to decreased selfesteem (Kot & Kulesza, 2016), recognition of facial expressions of emotions (Kulesza et al., 2015) as well as lie detection (Stel et al., 2009), and compliance with stereotypes (Leander et al., 2011).The present study supports this line of research.

Results in the light of the theoretical frameworks of unrealistic optimism and mimicry
Unrealistic optimism is the tendency for people to 'believe that negative events are less likely to happen to them than to others, and they believe that positive events are more likely to happen to them than to others' (Weinstein, 1980, p. 807), for example, when estimating the risk of a car accident (McKenna, 1993;Rutter et al., 1998), and in the domain of public health of developing heart disease (Prats et al., 2012) and developing addiction (Kim & Niederdeppe, 2016;Nezlek & Zebrowski, 2001).The theory states that the mechanism responsible for holding this bias is to deal with threatening moments in our lives, especially in uncontrollable times.Finally, the mechanism responsible for the reduction of unrealistic optimism bias is egocentrism reduction (Dolinski et al., 2022;Kim & Niederdeppe, 2016;Weinstein, 1983).The present study aimed to test another, new method of reducing egocentrism (via mimicry).Unexpectedly, mimicry not only did not reduce this tendency, but fueled it.
The theoretical consequences are at least twofold.First, within the scope of the unrealistic optimism theory, this paper may be the first to call for changes in the claim that egocentrism leads to the reduction of unrealistic optimism.It is possible that this mechanism does not always lead to this expected outcome.
From the scope of the unrealistic optimism theory it is worth mentioning that, this bias has previously been grounded in a solely interpersonal nature (we are comparing ourselves with others) and was also grounded on the intrapersonal level ('I am concentrating on myself and comparing myself to general others').In other words, in past research participants were confronted with 'general others' and were not confronted with specific people.In our study participants compared themselves with specific others, utilizing a solely interpersonal comparison.It turns out that such a comparison changes the pattern of results, broadening the horizon for future studies on issues like: (a) personal presence of 'compared with others', (b) what they are doing, and how they are behaving.Future studies might not only replicate our findings, but more importantly, on the grounds of unrealistic optimism theory, reshape this original concept.
Second, within the scope of the theory of mimicry, it has been claimed that mimicry leads to great social benefits for the mimicker, such as liking (Chartrand & Bargh, 1999) and prosocial tendencies (e.g., Kulesza et al., 2014;Van Baaren, Holland, et al., 2004), leading to the conclusion that mimicry lays a social glue between the mimicker and the mimicked (Dijksterhuis, 2005;Lakin et al., 2003).The present paper clearly does not support this claim.From a more general perspective, mimicry fuels biases that are harmful for societies.

Limitations and future directions
The main caveats of this study stem from the methodology.First, the experiment was run via the Internet.Since participants do have access to the Internet it is possible that this group is highly specific.Thus, replications in face-to-face interactions are necessary.
Second, this experiment was run during and was grounded in the COVID-19 pandemic (the topic of the interview).It is unclear if boosting unrealistic optimism would also take place while facing other threats like car accidents (Rutter et al., 1998), divorce (Lin & Raghubir, 2005), or health related issues unrelated to COVID-19, like heart disease (Prats et al., 2012).More research is needed to test the generalizability of the effect since it may be limited to solely this specific threat.
The third caveat stems from the fact that since only verbal mimicry was employed it is unclear whether other forms of mimicry, like nonverbal (Chartrand & Bargh, 1999) and mimicry of facial expressions or emotions (Kulesza et al., 2015), would produce the same pattern of results.
Fourth, since the pattern of the results was contrary to our expectations, the planned methodology and data cannot address the question of 'why' mimicry boosts unrealistic optimism.Mediators and moderators of this relationship might shed light on the actual process behind this surprising phenomenon.
Fifth, it is possible that self-focus may be a key moderator responsible for the abovedescribed surprising effect.It was shown (Guéguen, 2011) that self-consciousnessa mechanism linked to mimicry by heightening this factor -decreases (social) anxiety.As mentioned in the introduction, reducing fear (a concept similar to anxiety) is a key motivator for the usefulness of unrealistic optimism, making fear a common denominator in mimicry phenomena.It is possible that unrealistic optimism is heightened by mimicry since mimicry creates heightened self-consciousness, which is, in turn, clearly linked with egocentrism; a key factor behind unrealistic optimism.Future research might use a precise manipulation of self-awareness (a concept similar to self-consciousness) introduced by Duval and Wicklund (1972) in which sitting in front of a mirror increases self-awareness in participants.If, in this case, unrealistic optimism is heightened, such results would explain the aforementioned surprising pattern.Future research is highly recommended in this line of research.
Sixth, in the domain of unrealistic optimism research, controllability was postulated by its creator as another -besides egocentrism -key mechanism fueling this bias (Weinstein, 1980).Another key motivator -stemming from mimicry being linked with both self-consciousness (Guéguen, 2011) and unrealistic optimism -is controllability in the face of difficult situations.
Finally, from a more general point of view, future research should address the question of why the pattern of the results was exactly opposite to the expected pattern.On the one hand, future research might deliver the answer by, for example, including empathy as a critical factor responsible for this pattern of results.Indeed, mimicry has been shown to be highly linked with the cognitive aspect of mimicry (Chartrand & Bargh, 1999;Experiment #3), and it may be a key factor explaining the pattern of results.On the other hand, as briefly described above, it is possible that mimicry leads to a strengthened self-perceived connection with others (Lakin et al., 2008), and thus to a greater possibility of meeting others in person, resulting in higher direct exposure to others which might result in COVID-19 infection.To test this possibility, future studies should employ a pretest and posttest of risk assessment.

Practical implications
Mimicry is taught as a mechanism that deepens the interaction between a patient/client, and, for example, a therapist.In general, it is postulated that mimicry deepens mutual understanding (Gallese et al., 2004;Meltzoff & Moore, 1983;Stel et al., 2008), as well as elicits trust (Maddux et al., 2008;Swaab et al., 2011) and rapport (Muir et al., 2020).An explanation for why mimicry is responsible for deepening mutual understanding, which is crucial in the therapeutic process, is theory of mind: to understand others' ideas, intentions, and emotions, one has to mimic (Premack & Woodruff, 1978).
The present study shows another aspect of mimicry employed in therapeutic relationships.Mimicry performed by the clinician boosts a possibly dangerous social bias.Pairing this effect with an increased need for (psycho)therapy (e.g., APA, 2021) due to the two years of the coronavirus pandemic, clinicians may -contrary to their best interests -fuel their clients' unrealistic optimism bias.The same effect may be true for the patient-doctor relationship since mimicry is taught as a useful tool for effective medical communication (Wu et al., 2020).
Finally, as research shows, at least 50% of all patients do not follow recommended doses for medicine (Cramer & Rosenheck, 1998).In the light of the present data, it may be due to boosted -by the mimicking doctor -unrealistic optimism.In other words, one may think that s/he is less exposed, and after an appointment at the clinic, this selfperception may increase due to the mimicry performed by the doctor.

Conclusions
The aim of this study was to test the new method oriented toward unrealistic optimism reduction -in the specific context of the COVID-19 pandemic and as a promising avenue for global health policies -which is mimicry.Contrary to what we expected, the result was exactly the opposite: mimicry fueled unrealistic optimism.

Disclosure statement
No potential conflict of interest was reported by the authors.

Figure 1 .
Figure 1.Unrealistic optimism bias in the mimicry and no mimicry condition.Note: Dots represent means.Vertical bars represent standard errors.

Table 1 .
Descriptive statistics of all study variables.

Table 2 .
Post-hoc comparisons with Bonferroni correction adjusted for comparing a family of six estimates.