Trust and distrust in science: Embedding the interplay among scientists,mass media and public in Italy during the SARS-Cov-2 outbreak

Introduction

The spread of the SARS-CoV-2 virus has put scientists under the scrutiny of the general public and raised questions about trust in science and in science-based policies. ¹ Questioning trust in science is certainly nothing new, however in the last 10 years or so, with the widespread presence of scientists in the media, it has also begun to be increasingly questioned by an audience of non-experts. This phenomenon has had opposite consequences: in some cases, an increase in trust in science can be observed, in others, a decline of the same. Regarding the latter, the symbolic culmination of this process can be pinpointed to 2016, when the Oxford English Dictionary named ‘post-truth’ word of the year, an expression which served to describe a world in which facts ‘are less influential in shaping public opinion than appeals to emotion and personal belief’ due, according to some interpretations of the phenomenon, to widespread ‘science denial’ (McIntyre, 2018). The success of concepts like ‘fake news’, ‘pseudoscience’, ‘disinformation’, and so on expresses a strong concern about both the declining levels of trust in science, and the political and social consequences of its decline. From this viewpoint, the pandemic further strengthened the viral imagery of the circulation of unverified news, considered by some to be as dangerous as the virus itself (Simon and Camargo, 2021). In fact, several studies (Cinelli et al., 2020; Zarocostas, 2020) and the World Health Organization (WHO, 2020) itself have even made mention of an ‘infodemic’ (Rothkopf, 2003), a term that refers to both the circulation of news in excessive quantities as well as the absence of its careful screening prior to publication. Alternatively, or in addition to these concerns, it is believed that it was the display of conflicts in the mass media between scientists during the pandemic that undermined levels of trust in them. The underlying idea is that these conflicts generated disillusionment among people who had a naive view of scientific knowledge that is based on the idea that it proceeds in an objective and a-conflictual manner.

However, this view addresses the issue of trust in science and scientists in an excessively binary categorization that is in terms of a simple presence or absence of trust. Its implicit assumption is that a series of processes that occurred during the pandemic caused a decrease in levels of trust in science. Instead, the problem of trust is more complex and does not easily lend itself to simplistic interpretations. Although the aforementioned concerns are certainly well-founded, the uncritical adoption of terms such as ‘infodemic’, ‘fake news’, and ‘pseudoscience’ is highly problematic, if not misleading (Serafini and Zagni, 2023). The use of these terms implicitly reinforces binary categorizations, as if the entire problem of trust in science boiled down to the seemingly simple distinction between the true (science) and the false (fake news and the like). Moreover, it also appears overly facile to assume that a decline in trust in science has occurred simply because the normal disputes between scientists have been rendered visible in the media. Complex phenomena always require multifactorial explanations. We therefore intend to explore the hypothesis that trust in science and scientists is an articulated phenomenon, which does not necessarily boil down to the mere presence or absence of trust, but instead can best be explained by resorting to multiple dimensions. With this in mind, we will explore the hypothesis that several dimensions may emerge specifically in the context of the mediatization of scientific knowledge that took place during the pandemic. It is possible, in fact, that the public’s view of science, when interrogated in relation to the constant presence of scientists in the media and how they articulated their disputes, may give rise to a complex and multifaceted continuum of attitudes.

In order to explore this hypothesis, we will analyze the concept of infodemic in the next section and propose widening the perspective beyond the limits of the hegemonic paradigm of ‘viral’ information circulation; in addition, we will show how the issue of trust in science has been approached, in the context of the COVID pandemic, in overly binary terms, with many studies merely noting an alleged rise or fall in trust, often relating it to the infodemic drift and/or the ‘performance’ of scientists in the media. In the following two sections, we will reconstruct two conflicts between scientists that affected the Italian public debate during the first (February to June 2020) and second waves (September to December 2020) of the virus’ spread; our aim is not to show the relevance of the mere existence of conflicts between scientists in the media, but of the specific ways in which these conflicts were articulated and the implicit philosophy of science underlying the scientists’ stances. In the last section, we will then explore the question of trust in science in more detail to show how simplistic it is to approach this issue in a dichotomous manner, as if it were something one either possesses or does not possess. Through a survey conducted on nearly 4000 readers of two Italian online newspapers, we show how people have different attitudes toward science. It is worth clarifying that we do not claim the presence of a clear causal effect between the way conflicts are expressed in the media and the levels of trust in science. However, analysis of how the conflict between scientists was staged in the media was important for at least two reasons: first, to show how – in contexts of high uncertainty and a stringent need to take appropriate action – even scientists have expressed a dichotomous view, categorically differentiating between ‘science’ and ‘pseudoscience’, ‘truth’ and ‘fake news’. Second, conflict analysis was preparatory to formulating the web-survey questions. Indeed, respondents were (also) asked about how they perceived the disputes among scientists in the media during the pandemic. It is in connection with this element that their general attitudes toward science were assessed, precisely in order to explore the hypothesis that the reception of such disputes – even disputes in which a binary and dichotomous view of science is expressed – does not simply lead the public to gain or lose trust, but constitutes one of several aspects from which multiple and more complex attitudes toward science and scientists are formed.

Infodemic, media, and trust in science

The circulation of the SARS-CoV-2 virus had a major impact on reconfiguring several mechanisms related to the communication of scientific knowledge; first of all, it reconfigured agenda setting mechanisms, rendering everything that came from the scientific community highly ‘newsworthy’ (Papapicco, 2020): from the statements released by scientists to the media to reports of studies published in draft archives (pre-print) as well as in international scientific journals (Gazendam et al., 2020).

In this context, the term ‘infodemic’ has been used to refer to both the circulation of news in excessive quantity as well as the absence of its careful screening prior to publication. According to this view, an ‘information virus’ then spread, by way of communication channels, which contributed to heightening the amount of alarmism in the population.

All this happened in the context of the acceleration of communication flows typical of social media and the digital public sphere (Schäfer, 2016), within which the simplification of complex messages is accentuated, due in part to the algorithms that favor radicalization and the polarization of opinion (Serafini, 2023). The spread of fake news about COVID-19, therefore, represents an epiphenomenon of the more general and structural connection between fake news, understood as ‘digital entities’ and the dissemination of information in online environments (Gray et al., 2020).

It was the World Health Organization in particular that used the term ‘infodemic’ to refer to the media’s culpability in spreading fake news. In fact, a 2020 report stated, ‘The 2019-nCoV outbreak and response has been accompanied by a massive ‘infodemic,’ an over-abundance of information – some accurate and some not – that makes it hard for people to find trustworthy sources and reliable guidance when they need it’ (WHO, 2020). A few days later, WHO Director-General Tedros Adhanom Ghebreyesus (2020) emphasized the link between the infodemic and the spread of fake news stating the following: ‘We are not just fighting an epidemic; we’re fighting an infodemic. Fake news spreads faster and more easily than this virus, and is just as dangerous’. It has been noted that the journalistic use of the term ‘infodemic’ has strongly adhered to this definition given by the WHO, equating the information virus with the spread of fake news by the media in particular (Coleman, 2020; Hollowood and Mostrous, 2020; Simon and Camargo, 2021). Even within the academic sphere, numerous studies have equated the infodemic concept with the circulation of fake news, either on specific platforms such as Twitter (Pulido et al., 2020), or more generally with reference to the spread of different types of misinformation during the pandemic, from rumors to conspiracy theories (El Kihal et al., 2019; Rathore and Farooq, 2020).

However, the manner in which the ‘infodemic’ notion we have just described has oft been used in both journalistic and academic fields seems to attribute full responsibility to the media for the alleged distortions in scientific communication that took place during the pandemic. The implicit assumption in its utilization is that a monolithic scientific knowledge exists, which proceeds in a non-conflictual and objective way, anchored to absolute certainties, but which is the object of an inaccurate and distorted media representation which threatens our very trust in scientists themselves.

On a more general level, the widespread use of terms like ‘infodemic’ expresses the increasingly strong hegemony of a specific paradigm that understands the relationship between the dissemination of (mis)information and changes in people’s attitudes and behaviors. This approach, which some authors have proposed calling the ‘epidemic paradigm’ (Tosoni, 2022), proposes a poor view of interpretive processes and cultural variables since it is based on the behaviorist assumptions typical of the ‘strong effects’ theory of media (Anderson, 2021). In addition, as has been recently highlighted, the notion of an ‘infodemic’ has been used, both in the media and academia, as a functional metaphor to describe heterogeneous phenomena, which have ended up unduly simplifying a complex situation without clear empirical evidence (Simon and Camargo, 2021). Some studies have shown that the large amount of news about a topic does not necessarily translate for people into a difficulty to distinguish the reliability of sources. Recent research even found that in several countries, people affirmed they felt perfectly capable of seeking reliable information about the pandemic and maintained that the media in fact helped them understand the crisis as it progressed (Nielsen et al., 2020). Other studies have also shown that both the consumption and trust in news from the mainstream media increased during the pandemic (Fletcher et al., 2020; Nielsen et al., 2021); while there was no greater concern for the spread of disinformation compared to the pre-pandemic period (Newman et al., 2021).

The simplifications and binary categorizations implicit in the use of concepts like the ‘infodemic’ can also be seen in the way the issue of trust in science and scientists was addressed during the COVID-19 pandemic. In Italy, several studies have addressed the issue of a possible decline in trust in science and scientists. According to a survey carried out at the end of October 2020 by the Observa – Science in Society Observatory (https://www.observa.it/), between the first wave (March 2020) and the second wave (October 2020), the positive judgment toward scientific experts dropped by 23%, ‘thus squandering the large consensus accumulated in the spring’. In fact, one in four Italians gave them a negative rating.

Before the spread of SARS-Cov-2, as attested by the International Science Survey 2019–2020 conducted by the Pew Research Center (2020) in Washington, the level of trust toward scientists in Italy was relatively high: 76% of respondents declared they had a lot of trust (33%) or a fair amount of trust (43%), while 20% said they had little or no trust in scientists. Data from 3M’s State of Science Index created by the Ipsos (2021) institute between February and March 2021, and thus 2 months before our survey, are more complex to interpret. Despite very high levels of agreement with the direct statement ‘I trust science’ (91% of respondents were completely or fairly in agreement with the statement, whereas 10% were in disagreement), there were other elements that seemed to go in the opposite direction. For example, 16% said they were skeptical of science and 24% were unsure whether scientific research was conducted following ethical principles. The effects of the pandemic on trust, as well as the duration of the changes, are thus neither clear nor unambiguous. Other research has reported an undulating trend, alternately reporting a rise (Bromme et al., 2022) or a decline (Algan et al., 2021; Battiston et al., 2021) depending on the specific time at which the survey was conducted.

Many of these and other studies also traced the alleged decline in trust in scientists to the confusion they created through their stances in the media, and the way these stances influenced public attitudes. Although identifying the factors that contributed to the spread and reinforcement of the dichotomous view of trust is not among our main purposes, we believe that the dynamic between the media and scientists did play a role which, without any claim to exhaustiveness, we have partially documented. From our perspective, scientists also played an active part in this context and therefore, as mentioned earlier, this role cannot simply be attributed to the media exposure of conflicts but to the concrete ways in which conflict is expressed, which must be investigated specifically rather than taken for granted. This dynamic, in our view, is one of the factors that contributes to reinforcing a dichotomous view of science as something to be accepted or rejected.

It is therefore useful to investigate the way scientists represented the disputes between them in the context of their media stances. From this perspective, analysis of what happened during the pandemic might not be limited to verifying a decline in trust in science, as suggested by other studies, but rather to asking (1) how the communicative dynamics staged may have contributed to the formation of different attitudes in the public toward science and scientists which cannot be reduced to the mere presence or absence of trust; (2) whether or not the vision of scientific knowledge underlying scientists’ stances in the media (binary/dichotomous or more multifaceted) finds correspondence in the view of science expressed by laypeople.

How were the conflicts staged? Asymptomatic individuals and the ‘clinically dead virus’

Let us now examine in further detail how these conflicts were articulated in relation to two topics: (1) the Italian debate surrounding the contagiousness of asymptomatic individuals and (2) the presumed loss of strength of the virus over time, both of which we have subjected to scrutiny through research gathered from newspapers and news sites from the first of January to 31 December 2020. To this end, 819 articles were selected which include both interviews released by a list of the 50 best-known experts at the national level as well as articles in which statements provided in other contexts were reported on (mainly television interviews) and published in the major newspapers and online. ²

By coding the interview texts, 14 main themes were identified ³ that were the subject of conflicting interpretations by scientists. Of these, we explored two themes in particular because we found them especially useful to our research objectives: they involved several scientists, generated some public debate (due also to their relevance to the policies to be applied) and – particularly in the second instance – the scientists concerned displayed a high level of conflict. Moreover, the conflict surrounding asymptomatic individuals persisted over time whereas the ‘clinically dead virus’ implied a certain formalization of the different positions since the protagonists of the debate – as we shall see later – also expressed themselves through joint press releases.

Asymptomatic individuals – that is, those who contracted the virus without manifesting symptoms of the disease – were the subject of debate during all stages of the infection. Experts were asked to determine whether or not asymptomatic carriers of the virus were able to transmit infection, or whether it was spread only by people who exhibited symptoms of the disease. This was clearly a relevant issue since it concerned various lockdown policies. The conflict between experts had already begun to emerge in the first few months of 2020: a study published in The Lancet (Chan et al., 2020) in January 2020 highlighted the possibility of infection being transmitted even by asymptomatic individuals. This was joined by a study which presented similar conclusions published in the New England Journal of Medicine in February (Rothe et al., 2020). The two studies led scientists such as virologist Roberto Burioni, immunologist Guido Silvestri, pediatrician and infectious disease specialist Susanna Esposito, and virologist Ilaria Capua to talk about virus being transmitted by asymptomatic individuals in terms that were at times probabilistic and at times of absolute certainty, using expressions like ‘without a shadow of a doubt’. ⁴ However, such certainties were not shared by other scientists: after the case of two Chinese tourists testing positive in Rome on 30 January 2020, Giuseppe Ippolito, infectious disease specialist and scientific director of the Spallanzani Institute, reassured the population with the following words: ‘Citizens can rest assured because the real risk of transmission occurs through symptomatic individuals’. ⁵ Similar positions, between February and March 2020, were taken by other scientists such as the hygienist and toxicologist Fabrizio Pregliasco (‘it’s difficult for an asymptomatic subject to significantly infect another person’ ⁶ ) and the infectious disease specialist Paolo Grossi (‘based on current scientific knowledge, their level of contagiousness is minimal’ ⁷ ). The clash began in June 2020, coinciding with debate on the supposed attenuation of the viral load of the virus – which we will discuss in further detail in the next section. At least three different positions emerged over the course of the summer.

The first position that can be summarized by the formula ‘a COVID-positive person is not a sick person’, was that of those who expressed themselves with certainty about the non-contagiousness of asymptomatic individuals. It made reference to the ‘document’ titled ‘SARS-CoV-2 in Italy today and Covid-19’ ⁸ which was signed by 10 scientists from different disciplines who argued that we were facing an exponential increase in ‘weakly-positive’ individuals who were consequently non-contagious. The other cornerstone on this front was the June 2020 study by the Mario Negri institute, the results of which were disseminated in the media by the director of the institute, nephrologist Giuseppe Remuzzi, who also highlighted a low viral load in asymptomatic subjects (in his institute) and maintained its non-contagiousness. ⁹

This study, however, was strongly contested by other scientists: a counter-document signed by seven experts from different disciplines stated, among other things, that it was not proven that ‘the virus isolated in subjects who have been asymptomatic for several days has no infectious capacity in vitro’. ¹⁰ Other experts, from the hygienist and epidemiologist Paolo Bonanni to the pulmonologist Sergio Harari, affirmed that Negri’s study was too limited to be able to generalize the results. ¹¹ This second front, which we could define as ‘uncertain’, maintained that it was not possible to express the contagiousness of asymptomatic individuals with certainty either in one sense or another.

The third position instead was represented by those who continued to maintain a position of certainty on the contagiousness of asymptomatic patients: among these we can include microbiologist, parasitologist, and expert on tropical diseases Andrea Crisanti, who repeatedly referred to evidence from the study conducted to Vo’ Euganeo, later published in Nature (Lavezzo et al., 2020), as proof of the contagiousness of asymptomatic individuals, expressing himself in a caustic manner toward those who sustained opposing theses (‘When they tell me asymptomatic people don’t transmit infection, I throw up my hands’ ¹² ). Even physicist and expert in computational epidemiology Alessandro Vespignani discredited the theses of those who argued for the non-contagiousness of asymptomatic individuals in numerous articles with phrases like ‘the asymptomatic and pre-symptomatic are contagious and must be traced, otherwise we might as well put a blindfold on and pray to God’. ¹³

It should be pointed out that several scientists involved in the dispute conveyed a scientistic view of scientific knowledge, presenting their theses as certain, based on objective facts, and accusing the other side of incompetence. This resulted in caustic stances (such as those of Crisanti and Vespignani quoted earlier), marked not by the articulation of a debate between different views, but rather by the stigmatization (and even mockery) of theses at odds with their own. This occurred despite the fact that the evidence on this specific issue was far from unambiguous, especially during the first phase of the virus’ spread.

One of the most heated conflicts, however, developed around a phrase pronounced by Alberto Zangrillo, on 31 May 2020, during a television broadcast: ‘the virus from a clinical point of view no longer exists’. In support of this thesis, Zangrillo referenced at that time ongoing study by virologist Massimo Clementi. In truth, Zangrillo’s statement did not indicate the virus was gone, but signaled evidence that no one disputed: a large reduction in hospitalizations and intensive care unit (ICU) admissions.

The idea that the virus had weakened had already circulated in Italy around 6th May, following an interview given by Clementi to Corriere della Sera, in which he cited the study he was participating in, together with many Italian hospitals, and whose goal was to ‘have constant monitoring of the sequences of the virus’: the expected result of the study was to show that ‘this new coronavirus may slowly become harmless, like its ‘cousins,’ which are responsible for colds’. ¹⁴ Still, many had already claimed, even if only in terms of a possibility, that the virus was losing strength (without commenting on possible causes). In fact, between mid-April and early-May, hospital radiologist Maurizio Borghetti, immunologist Francesco Le Foche, epidemiologist Massimo Ciccozzi and the aforementioned Palù and Bassetti expressed themselves in this sense. During the month of May 2020, interest in the topic increased progressively, also and primarily because on 4th May restrictive measures were eased in Italy (the so-called ‘phase 2’) and, despite the expectation of a resurgence of infections and hospitalizations, these continued to decline. For some, this was evidence that the virus had mutated, that the viral load had decreased, or that the disease had changed, which might then die out naturally. For others, however, the decline in hospitalizations was simply due to a reduction in the number of infected persons.

At least in the initial phase, opened by Zangrillo’s famous phrase, the clash seemed to concern a different approach to the epidemic: the clinicians (who worked, directly or indirectly, in the hospital in close contact with patients) supported the thesis of the ‘clinically dead virus’ while those who relied on an approach more closely linked to epidemiological models of viral spread contested it. Already by 13 May, when the conflict had yet to assume the harsh tones it would take on later, Clementi explained the difference in perspective with respect to his colleagues who arrived at conclusions different from his own: ‘they start with algorithms, with initial assumptions, with amazing big data, with pre-established keys of interpretation. [But then] they find themselves in front of the clinical reality [. . .] irrefutable! – and the whole castle falls down’. ¹⁵ Thus, Zangrillo, again in the aforementioned interview, affirmed ‘there is only one number that is valid’ and ‘it is the evidence: in this country we heard a professor from Boston a month ago, who is an epidemiologist-statistician named Vespignani, who influenced the government’s choices by saying that 151.000 intensive care units had to be built’. In fact, Zangrillo here seemed to confuse Vespignani with mathematician Stefano Merler, who was responsible for the epidemiological model that was to be the basis upon which the government designed and adopted measures for the so-called phase 2. Yet, the reference did not seem accidental, given that, in an interview about a month earlier, Vespignani defended his approach in this way:

I’m not good at disassembling the motor of a virus, finding a vaccine, but I work with data to understand how many infected there will be tomorrow, in a month, in Milan, Bergamo, or New York [. . .] [There is a] difference between a virus mechanic and an epidemiologist.

And so then ‘virologists – as I will explain – might be luminaries, but sometimes they are the least suitable people for understanding how epidemics develop’. ¹⁶

In one of the phases of greatest tension, two groups of scientists took a strong public position on the matter: the first group, which included 10 of the most well-known scientists at a public level and from different disciplinary backgrounds, published on social media the aforementioned document on 23 June entitled ‘SARS-CoV-2 in Italy today and Covid-19’, in which it was argued that

hospitalization for symptoms attributable to viral infection is now a rare phenomenon [. . .]. The virological evidence, in complete parallel, shows a constant increase in cases with a low or very low viral load. Studies are in progress to explain the cause.

This group of ‘optimists’ – accused of superficiality by the opposing side – was contrasted by another group of more cautious scientists – accused in turn of ‘catastrophism’ – who published a document a few days later (on 28 June) that responded to ‘the 10’ in which they declared themselves to be ‘worried that optimistic declarations about the reduced contagiousness of SARS-CoV-2 could cause people to mistakenly believe that the danger had abated. [. . .] The virus currently circulating is active and contagious’ and is ‘capable of causing extremely serious damage similar to what it did at the beginning of the epidemic’. ¹⁷ It is interesting to note how the parties involved, especially the representatives of the most extreme positions in the most heated phases, did not see the conflict in terms of a difference in interpretation of data and the phenomena under discussion, but rather as a problem of a lack of scientificity or competence of the opposing side. ¹⁸ Appeals to limit oneself to the ‘real’ evidence or the ‘real’ data were leveled at both sides and, in some cases, even led to actual personal attacks. ¹⁹

As noted earlier, scientists have often represented their different opinions on these two specific topics not as an expression of the usual scientific method, but as a conflict between ‘truth’ and ‘fake news’, ‘competent scientists’ and ‘incompetent scientists’. We are not interested, for the purposes of this study, in whether this occurred because of the scientists themselves, the media, or other factors. Instead, the findings were useful, in the first place, to show how scientists themselves have reinforced binary categorizations and, additionally, for the purposes of constructing the subsequent web-survey on which we will focus in the next section. In fact, several questions were constructed in order to investigate how respondents perceived this specific way of articulating conflict by the scientists themselves and how different attitudes (not reducible to a simple increase or decrease in trust) toward science were also formed in relation to its media representation.

Trust in scientists in Italy: A pilot study

As already mentioned, we believe it is more useful to go beyond a binary view of trust and thus, in order to further reveal attitudes toward science and scientists, we conducted a pilot study ²⁰ on a sample ²¹ of online readers of two Italian newspapers, Il Fatto Quotidiano and The Post Internazionale (TPI). ²²

One of the aims was to analyze how the various elements recalled in the theoretical portion and in the first part of our empirical analysis (the media exposure of scientists and the way they staged the conflict in their debates in particular, as well as the scientistic image of scientific knowledge often underlying their stances) were perceived by the readers of these two newspapers. Thus, an attempt was made within the sample to detect both favorable or critical attitudes toward how scientists expressed themselves during the pandemic, as well as the more general conception of scientific knowledge and its role in society. This is also in order to show how the reception of controversies among scientists does not only lead to an increase or decrease in trust, as the studies we have reported in the theoretical part argue, but can be framed in more complex public views of science.

Between the 10th and 17th of May 2021, 3391 and 587 questionnaires were collected, for a total of almost 4000 standardized online interviews. The questionnaire (included here in Appendix 1, Supplemental material) of the web-survey was made up of 13 questions (completion time was 12 minutes on average). The wording of questions considered the objectives of the research. Accordingly, a first block of questions sought to detect respondents’ attitudes toward how scientists expressed themselves in the mass media during the pandemic. Several of these questions, which qualified the type of conflicts that took place between scientists, could not have been formulated without a prior analysis of those conflicts. For example, respondents were asked to express their degree of agreement or disagreement with statements such as “They [scientists] made overly-authoritative statements, without saying ‘maybe’ or ‘could be.’” This statement directly refers to the results of the conflict analysis between scientists in the media, which showed that many of them were expressing their opinions with absolute certainty on topics on which clear evidence and agreement had not yet formed within the scientific community, using expressions such as ‘without a shadow of a doubt’. Other questions were aimed at comparing how respondents conceived of disputes between scientists with how those disputes were portrayed by the scientists themselves. In other words, having observed that scientists often promoted a dichotomous view of science in the conflict analysis, our aim was to understand whether this dichotomous view was also observable among respondents. For example, respondents who felt that scientists’ stances in the media were contradictory were asked whether this was ‘Because science is unable to arrive at a single truth’ or ‘Because it’s normal that, when faced with a new phenomenon, scientists correct each other and get closer to understanding reality’. Similarly, questions were formulated to test the degree of agreement or disagreement with statements such as (1) ‘Scientific truths are NOT contaminated by the beliefs of the moment, because they are objective and impersonal’ and (2) ‘Scientists act rationally, basing on data and evidence’; or in the opposite direction (3) ‘The natural sciences are also influenced by the personal beliefs and opinions of scientists’. These questions sought to detect whether the vision promoted by various scientists in the media, which emerged from the conflict analysis we carried out (i.e. that science proceeds by absolute certainties, by evidence based on data and empirical observation, and that it does not depend on social factors of any kind), matched the respondents’ vision. This was useful to test the hypothesis that a plurality of latent attitudes toward science can emerge in the public sphere – if properly investigated – regardless of the way scientific knowledge is represented in the media, even by scientists themselves. Other studies, in fact, have associated the dichotomous view of science that emerged in media representations of scientific knowledge during the COVID-19 pandemic with a similarly dichotomous view on the part of the public, establishing correlations (or even, in some cases, causal links) that are, in our view, overly simplistic. Preliminary analysis of the conflicts between scientists in the media thus made it possible to formulate questions – aimed at investigating respondents’ general conception of scientific knowledge – based on the vision of science promoted by the scientists themselves, as well as other questions which opposed this vision.

Another important block of questions, closely related to the analysis of media conflicts among scientists carried out in the first part of the research, was aimed at testing opinions on how scientific knowledge can be communicated and how it actually succeeds in being communicated by scientists. Respondents were then asked to express their level of agreement with statements such as ‘Scientific knowledge cannot be communicated to other people because its contents are too complex’, or ‘Scientific knowledge can be communicated to ordinary people, but scientists often don’t know how to communicate effectively’. Finally, some control questions were included to test the consistency and stability of respondents’ opinions.

Almost all questions in the survey are original and were constructed specifically for the purposes of this investigation so that responses would not be collected that were uninformative, stereotypical, predictable, or with biased assumptions, as is the case in many questionnaires (see Gobo and Mauceri, 2014). Only in a few cases were questions taken from previous research in order to have some comparison. In particular, question Q3 was taken from a survey by the Sigma Consulting company, conducted in February 2021; question Q6 and some items in question Q7 were taken from question QC4 and question QC6.8 of the Special Eurobarometer on Science and Technology report (2010) respectively. For the original questions (e.g. Q8A and Q8B), an attempt was made to take some of the concepts or statements of Popper (1972), Kuhn (1962) and Feyerabend reported in their books. ²³

The rationale that guided the formulation of questions was to seek positions, attitudes, and opinions that were articulated and differentiated, in order to restore complexity to a topic (science) that is often simplified, radicalized, polarized and dichotomized. For this reason, questions were constructed that could express all positions along a continuum ranging from realism to radical relativism, inspired by the proposal of Gobo and Marcheselli (2022: 111).

The primary source used by respondents to gather information on the SARS-CoV-2 virus and the COVID-19 disease was TV (48%), followed by newspapers, both in print and online (19%) and the Internet (11% – Google, Wikipedia, YouTube). Those who consulted the websites of institutions (of the Region, Ministry, etc.) were 4%, the family doctor 1%, blogs and forums 0.6%, and family and friends (also through chat) 0.4%. Surprisingly, 9% of respondents stated they did not use any source (among the many listed in the question), and the much-discussed social media (Facebook, Twitter, WhatsApp, Telegram, Signal, Instagram, etc.) were the main source for only 3%. Certainly, the latter data point is partially due to the fact that online newspaper readers were interviewed. However, the contiguity of the mode used (online), which also includes social media, might have led one to imagine a higher percentage.

Discussion and conclusions

After the outbreak of the pandemic, scientists had a larger media presence than they had before. This also meant that conflicts became more visible, which was to be expected given the uncertainty and complexity of the situation. Since 2020, several studies in Italy and Europe have sought to measure changes in the levels of trust in science, which had also been documented by previous comparative research. Although some of these studies did identify a decrease in trust, others instead documented an increase. More generally, levels of trust seem to vary greatly depending on the specific juncture in time in which they are examined. This is clearly a methodological problem. Different results could depend on the questions and their quality, the textual premises, the order, sharpness, and distinctiveness of the items, the articulation of the response alternatives, and so on (see Gobo and Mauceri, 2014). With this in mind, we constructed the web survey with the aim of exploring a hypothesis in partial contrast to those that frame the issue of trust in science in binary terms. In fact, many studies simply record an increase or decrease in trust and also orient survey construction on the basis of these hypotheses. However, in our opinion, it is possible to trace the topic of trust in science along a plurality of dimensions and attitudes that can be formed among laypeople. In order to understand how to construct a survey capable of rendering this variety of attitudes, it was therefore useful to carry out a preliminary study on the way scientists had hitherto represented their disputes in the media. In this way, the plurality of dimensions and attitudes could be referred to the mediatization of scientific knowledge that took place during the pandemic and could also be compared with the attitudes and visions of scientific knowledge conveyed by scientists themselves.

Preliminary analysis revealed that scientists in conflict have – as seen above – at times interpreted opposing positions to their own as unscientific or as not adhering to ‘facts’; or have presented the results of ongoing research and uncertain data as if they were already well established. This has likely contributed to clouding the constructive nature of the scientific debate. Essentially, some scientists have represented the normal discussions that always play a role in scientific progress (especially in situations of marked cognitive uncertainty such as that presented by a new, unknown virus) as a diatribe between science and pseudoscience and have appealed to their colleagues to stick to ‘the facts’ rather than just providing opinions. In fact, as emerged from our analysis, a number of scientists from the natural, physical, mathematical, and even medical sciences made public accusations of incompetence on a disciplinary basis against their colleagues and conveyed a scientistic image of scientific knowledge, which they claimed was based on ‘objective facts’. In other words, scientists expressed a binary and tendentially ‘scientist’ and positivist view of scientific knowledge.

On the basis of this evidence, a questionnaire was then constructed with the aim of creating a combined analysis of the way these conflicts were interpreted by the sample queried as well as their more general view of scientific knowledge and the role of science in society. The questions were constructed in order to test whether the binary and dichotomous view of science conveyed in public debate by scientists themselves was mirrored in the attitudes of laypeople, or whether these attitudes could highlight a more complex and multifaceted view of science instead.

This study clearly has limitations and further studies will need to be conducted in the future. More specifically, it is a pilot web-survey, which does not aim to be statistically representative. The composition of the sample may therefore have influenced the prevalence of some attitudes over others. For example, the prevalence within the sample of people politically oriented toward the center-left might partly explain the preponderance of respondents who displayed a positivist attitude. Indeed, during the most acute phase of the pandemic, the center-left government expressed the view that policy decisions should be inspired by scientific evidence and a strong belief in the work of scientists. In contrast, the center-right parties advanced many more doubts about scientists, hypothesizing conflicts of interest and criticizing their way of expressing themselves in the media. The non-representativeness of the sample may also place further limitations on the number of attitudes that emerged in the survey: a sample composed differently might ensure that the prevalence of some attitudes over others be reversed or that more or fewer attitudes emerge. The classification of five attitudes for this study in fact does not exhaust all possible attitudes toward science. Furthermore, the research did not aim to directly test whether or not there was a causal link between the media performance of scientists and the confidence levels of respondents. However, the analysis procedure used could contribute to enriching the approaches through which the topic of trust in science and scientists is investigated through subsequent investigations on statistically representative samples. In our view, the results of current surveys are often the product of simplified questions. Consequently, we wanted to observe whether we could discover less simplistic attitudes and opinions that cannot be traced back to a dichotomous view of science itself by presenting a more articulated and complex view (and thus closer to the reality of science). This also implies that future studies, beyond the representativeness of the sample to which they will be addressed or otherwise, may draw upon this way of formulating a questionnaire to verify the presence of attitudes even further divergent from those identified in this article. What matters, in fact, is not so much (and only) the specificity of the attitudes found, but rather the analysis procedure established that is replicable even where there is a desire to investigate the presence of different attitudes.

Furthermore, the fact that respondents did not only express themselves (positively or negatively) on the communication methods of scientists during the pandemic, was important. Restricting it to this element, and linking it with dichotomous questions on trust in science, would have produced binary results, in line with other research published during the pandemic. Conversely, by testing opinions on how science is communicated in connection with the general view of scientific knowledge and the role of science in society, it was possible to differentiate five distinct attitudes, thus highlighting a plurality of nuances. It was true, in fact, that for a proportion of respondents, positive or negative views toward the communicative modes of scientists were reflected in their general conception of scientific knowledge. These were respondents who manifested a positivist or, to the contrary, skeptical and disenchanted attitude toward science. However, in other cases, the connection between these various aspects produced more complex and nuanced positions. Those that manifested a realist attitude, in fact, perceived the conflicts between scientists during the pandemic as proof of the non-communicability of scientific knowledge to ordinary people. Although the realist attitude displays an overall positive attitude toward science and scientists, it differs from the positivist position. In fact, respondents from the latter stated they wanted to see more scientists on television and thus considered scientific knowledge to be fully communicable to ordinary people. Articulation of the link between the communication of science and the general view of scientific knowledge was therefore different between the two groups. The respondents who manifested a constructivist attitude, however, emphasized a view of scientific knowledge that was ‘communicable’, while at the same time expressing a negative opinion of the way scientists actually manages to communicate it.

Generally speaking, that binary and dichotomous vision of science conveyed in the public debate by the widespread use of concepts such as ‘fake news’ and an ‘infodemic’, partly promoted by the way scientists themselves staged their conflicts in the media, did not match the (more articulate) visions of science expressed by the laypeople who responded to our survey.

To conclude, what we have proposed in this article, the joint analysis of media conflicts between scientists in Italy and the attitudes of the Italian population toward science, is intended to help overcome the binary categorizations of a generic trust or distrust in science. Only by overcoming these dichotomous oppositions, in our opinion, is it possible to shed light on the plurality of attitudes toward scientific knowledge, which must be investigated in their complexity and in relation to the specific historical, social, and cultural variables that contribute to generating them.

Supplemental Material