Characterization of polarized scientific digital messages: a scoping review

1 Context

Social psychology defines polarization as a deliberate process within a group where members tend to adopt more extreme opinions after discussion [Moscovici & Zavalloni, 1969]. This process fosters a collective narrative, creating a fertile environment for developing and reinforcing identity and cohesion in homophilic groups [Turner et al., 1989; Hogg et al., 1990; McPherson et al., 2001; Postmes et al., 2005; Bliuc et al., 2021]. Within this context, polarization can be categorized into three distinct forms [Arguedas et al., 2022]: (a) ideological polarization, which pertains to the extent of disagreement among individuals regarding political issues; (b) affective polarization, which relates to the emotional responses individuals have toward opposing groups with differing views; and (c) audience polarization, which describes the degree to which audiences of news channels exhibit political partisanship or diversity. As a result, polarized messages often amplify these forms of polarization by reinforcing extreme viewpoints, evoking strong emotional responses, and deepening divides between different audience segments. Consequently, polarized messages not only reflect but also perpetuate polarization processes, reinforcing group identities and solidifying divisive narratives within society. It is noteworthy that these forms of polarization can individually and collectively impact various social aspects, underscoring the importance of understanding these phenomena in scientific research.

In the communication of science, polarization arises from debates over the duality of results and methods, as well as from the discrediting or manipulation of findings by both specialized and non-specialized communities [O’Connor & Weatherall, 2018; Abramowitz & Saunders, 2008; Ploug & Holm, 2015]. As a result, the polarization of scientific discourse can lead to disputes affecting science’s objectivity, progress, social reputation, and impact. It biases debates and public decisions and affects scientists’ motivation, shifting from scientific interest to personal visibility [Sharpe, 2002; van Kolfschooten, 2002; Ziman, 2002]. For instance, the debate over consensus-based anthropogenic global warming versus alternative oil industry’s views has delayed global efforts to reduce carbon emissions significantly [Franta, 2021]. Similarly, discussions on the efficacy of vaccines intensified during the COVID-19 pandemic led to preventable deaths, particularly in countries with political leaders who deny scientific evidence [Morris, 2024].

The polarization process is markedly intensified in digital relationships on social media. Within this context, while compelling empirical evidence for the existence of platform- or algorithmically-induced echo chambers is lacking [Bruns, 2019; Arguedas et al., 2022], people actively choose to engage with content that aligns with their pre-existing beliefs, avoiding contradictory information [Marozzo & Bessi, 2018; Kim & Kim, 2019; Chinn et al., 2020; Choi et al., 2020; Kubin & von Sikorski, 2021]. Consequently, personal characteristics appear to play a more significant influence in the polarization process than content suggested by digital platforms, because users are also exposed to opposing viewpoints in some media [Arguedas et al., 2022]. On the other hand, the role of polarized messages within this scenario remains underexplored, particularly from a scientific perspective. While numerous studies have examined the traits of misinformation and disinformation in digital media, there is a significant gap in the literature regarding the characteristics of scientific polarized messages [Bryanov & Vziatysheva, 2021; Molina et al., 2021; Hamby et al., 2024].

Based on the aforementioned research literature, this scoping review aimed to identify studies that characterized polarized scientific digital messages from the perspective of authors of science communication. To achieve this aim, the PCC (Population/Concept/Context) framework was used to define the research questions. Specifically, this framework is structured around three pillars that guide the theoretical background: population (the group or entity being studied), concept (the main idea or issue being explored), and context (the setting or environment). In this study, ‘population’ refers to science communication messages, ‘concept’ denotes polarization, and ‘context’ encompasses the internet [Peters et al., 2024]. As a result, the review addresses the following questions:

• What are the characteristics of polarized scientific digital messages found in papers, science communication articles, or specialized discourses?

• What definitions and synonyms exist for polarized scientific digital messages?

• How are studies on polarized scientific digital messages characterized by their countries of origin, fields of knowledge, and themes?

• Can polarized scientific digital messages be linked to specific authors, receivers, textual marks, or interest issues?

• Are there distinct patterns in polarized scientific digital messages found in scientific papers, science communication articles, and specialized discourses?

Addressing the main question (Q.1) of this study, the synthesis of content and discourse characteristics of polarized messages in science communication can facilitate their identification across diverse digital contexts. This is relevant to a range of current situations, from the increased production of digital materials by professionals without formal media training to prestige disputes or discourses with hidden agendas posted as pieces of genuine scientific education on social media. Moreover, these results can guide future research in this area, highlighting the knowledge gaps and challenges that emerged from the disproportionate discursive battle observed in fields of public interest within democratic societies. Based on the systematization of findings from this scoping review, we also aim to propose a codification system for identifying and characterizing polarized digital messages in the context of science communication.

2 Materials and methods

This scoping review was developed according to the recommendations of The Joanna Briggs Institute Reviewer’s Manual — Methodology for Scoping Reviews [Peters et al., 2024]. These results are reported following the Preferred Reporting Items for Systematic reviews and Meta-Analysis extension for Scoping Reviews (PRISMA-ScR) checklist [Tricco et al., 2018].

The protocol of this study was registered in Open Science Framework platform in October 20, 2022, which can be accessed in https://osf.io/t8a6f/.

2.1 Eligibility criteria

Initially, according to the registered protocol for this review, the inclusion criteria were limited to studies on discourse and content analysis that focused on the specific syntactic and lexical characteristics of polarized messages within the context of science communication in cyberspace. However, upon identifying a scarcity of literature, the inclusion criteria were revised and broadened to incorporate other studies with distinct designs that presented polarized messages, as recognized by their own authors. In this context, crossover and randomized information intervention studies were included in the analysis. This investigation considered references published in any year and language available on October 29, 2022 (date of collection). Scientific papers lacking sufficient detail for data extraction (see details below) or failing to address the characteristics of the messages were excluded from the study.

2.2 Search strategy

On October 4, 2022, a search strategy combining representative keywords of the population (science communication messages) and concept (polarization) was employed for retrieving references on Medline. The first ten articles related to the themes were screened to obtain a list of additional keywords. The relevance of keywords for search strategy was discussed between two investigators (AMJ and TC) and included in the search strategy after consensus. These procedures were repeated on the Web of Science. After that, a final search strategy was determined, as follows: ((“polarized” OR “polarization” OR “partisan” OR “polarised” OR “polarisation” OR “opposing” OR “one-sided” OR “tendentious” OR “authoritarianism” OR “politization” OR “biased judgment” OR “conservative ideology” OR “expert consensus” OR “ideological bias” OR “ideological belief” OR “misleading argumentation” OR “motivated reasoning” OR “perceived consensus” OR “political ideology” OR “political orientation” OR “scientific ideology” OR “partisanship” OR “reactance” OR “scientism”) AND (“scientific news” OR “science news” OR “scientific information” OR “science information” OR “science content” OR “scientific content” OR “science discourse” OR “scientific discourse” OR “scientific communication” OR “science communication” OR “scientific message” OR “science message” OR “science publication” OR “scientific publication” OR “scientific publications” OR “scientific dissemination” OR “science dissemination” OR “scientific divulgation” OR “scientific press” OR “science press” OR “science journalism” OR “scientific journalism” OR “science communicators” OR “science thinking” OR “scientific belief” OR “scientific consensus”)).

On October 29, 2022, references were retrieved by searches conducted on the following databases: Medline, Web of Science, Cochrane, EMBASE, Portal BVS, Clinical Trials, Scopus, ERIC, Human Resources Abstracts and PsycINFO.

2.3 Study selection

The references retrieved were uploaded into Endnote Web^® Reference Manager (Clarivate, London, U.K.) to remove duplicates. Titles and abstracts were screened by two independent investigators (AMJ and ML) to exclude studies not related to polarized scientific digital messages. Papers judged by at least one of the investigators as related to the topics of interest were then fully screened by one investigator (AMJ) to ensure adherence to the selection criteria. Additionally, the reference citations of the selected papers were manually checked by the same investigator to identify any references not retrieved by the databases. Those papers with titles that seemed significant to this review were annotated and retrieved. Finally, the two investigators (AMJ and ML) performed an independent full reading of selected publications. After that, both investigators selected the papers for the scoping review consensually, considering the eligibility criteria aforementioned.

2.4 Data charting

The investigators (AMJ and ML) independently charted data from the included papers using Microsoft Excel. This process involved synthesizing the main characteristics of the studies, such as the authors’ names, type of study (e.g., content analysis, discourse analysis, cross-over, or randomized information intervention), year of publication, country of origin, field of knowledge, subjects of the messages, aims of the study, sample size, characteristics of the messages, and main findings. Additionally, the investigators identified definitions, synonyms, and alternative terms for polarized scientific messages used in these studies. The data charted similarly by both investigators were considered as valid results; otherwise, discrepancies were solved by re-analysis and consensus to validate the results.

This scoping review did not include a critical appraisal of the studies, as its purpose was to summarize the existing evidence on the topic to inform future research, policy, and practice.

The Litmaps (litmaps.com, Wellington, New Zealand) was used to determine the relationships among the references of the included papers. This application was employed to produce a visual map of these relationships and to elucidate the influence of cited references on the development of this issue through citation metrics.

2.5 Codification system

Based on the independent analysis of results, two independent investigators (AMJ and TC) listed possible codes to enable the identification and characterization of polarized digital messages in science communication . After that, redundant codes were removed by consensus. The significant codes were aggregated into categories of similarity.

3 Results

3.1 Selection of sources of evidence

In Figure 1, it is shown that 3,320 references were initially retrieved from the databases. After removing 888 duplicates, the remaining 2,432 titles and abstracts were evaluated for their relevance to polarized scientific digital messages, leading to the exclusion of 2,177 references. The subsequent assessment of titles and abstracts of the remaining 255 references were made against additional eligibility criteria, resulting in the exclusion of 134 studies for lacking digital messages, 30 for not being authored by experts or science communicators, and 51 for methodological designs not focused on message analysis. Then, 40 papers were selected for full reading, with addition of 4 papers found through cross-checking their reference lists. After full reading, 34 out of the 44 papers were excluded for not focusing on polarized scientific digital messages [Kata et al., 2012; Hall Jamieson & Hardy, 2014; Scheufele, 2014; Myers et al., 2015; Pearce et al., 2015; Dalrymple et al., 2016; Dixon, 2016; Winter & Krämer, 2016; Hardy & Hall Jamieson, 2017; Zielińska, 2017; Bolsen & Shapiro, 2018; Harvey et al., 2018; Jiang et al., 2018; Lee et al., 2018; Merkley & Stecula, 2018; Hardy et al., 2019; Iyengar & Massey 2019; Lenzi, 2019; Nagler et al., 2019; Pearce et al., 2019; Yuan et al., 2019; De Cruz, 2020; Landrum & Slater, 2020; Pontalti Monari et al., 2020; Baldi & Gala, 2021; Haupt et al., 2021; Hernandez et al., 2021; Rode et al., 2021; Schneider, 2021; Zhang et al., 2021; Boyd, 2022; Jia, 2022; Wirz et al., 2022; Nagler et al., 2023].

3.2 Characteristics of sources of evidence

The 10 papers included in this review [Kortenkamp & Basten, 2015; Winter et al., 2015; Roper et al., 2016; Dunwoody & Kohl, 2017; Anderson & Becker, 2018; Becker & Anderson, 2019; Lyons et al., 2019; Sol Hart et al., 2020; Wang & Huang, 2021; Iles et al., 2022] are summarized in Table 1.

3.3 What are the characteristics of polarized scientific digital messages in papers, science communication articles, or specialized discourses?

Although the syntax of polarized messages was not detailed in the studies, the lexical characteristics were well-described. These messages typically presented a single viewpoint. Polarizing techniques included associating viewpoints with minority organizations and emphasizing flaws [Kortenkamp & Basten, 2015]. Other forms involved assertive statements and strong intensifiers while avoiding lexical hedges [Winter et al., 2015]. The discourse often featured power vs. resistance, corruption vs. purity, and hysteria vs. reason discourses [Roper et al., 2016]. Strategies like supporting an argument with the consensus of expert opinion [Dunwoody & Kohl, 2017] or employing sarcasm [Anderson & Becker, 2018; Becker & Anderson, 2019] also contributed to polarization.

In synthesis, the following concepts were associated with polarized messages: controversy, opposition, minority, majority, conflict, sidedness, failureness, topic dependency, skepticism, confrontation, partisanship, unilateralism, unbalancement, singleness, assertiveness, intensiveness, empowerment, resistance, corruptness, pureness, hysteria, reason, supportiveness, sarcasm, persuasiveness, refutation, contestation, politization, and uncertainty.

In contrast, non-polarized messages were recognized for their neutral and balanced arguments [Kortenkamp & Basten, 2015; Wang & Huang, 2021], the presentation of multiple viewpoints [Kortenkamp & Basten, 2015], and the use of lexical hedges and terms that indicate uncertainty [Winter et al., 2015]. Most studies emphasized the presence of diverse opinions [Kortenkamp & Basten, 2015; Dunwoody & Kohl, 2017; Anderson & Becker, 2018; Lyons et al., 2019; Wang & Huang, 2021; Iles et al., 2022]. The following concepts were associated with non-polarized messages: inclusion, multiplicity, impartialism, objectiveness, credibility, equality, neutrality, balancement, attenuation, uncertainty, diversity, and conformity.

To generate a codification system for identifying and characterizing polarized digital messages for science communication, two independent investigators (AMJ and TC) proposed 48 codes according to these findings. Then, 28 codes were removed by consensus of redundancy. Based on similarity, the remaining 20 codes were organized into five categories, called sidedness, criticism, emphasis, discordance, and non-polarized. The complete list of codes and their definitions is presented in Table 2.

3.4 What definitions and synonyms exist for polarized scientific digital messages?

Although the papers did not mention specific definitions of polarized scientific messages, they offered concepts in their introductions or conceptual frameworks that shed light on their approach to the topic. Kortenkamp and Basten [2015] characterized polarized journalistic practice as presenting both sides of an issue while disproportionately criticizing one side, often portraying it as a minority view or one with a conflict of interest. They defined balanced journalism as the effort to include multiple opposing viewpoints as accurately and impartially as possible, aiming to demonstrate objectivity and credibility without taking sides. According to them, balanced reporting involves giving equal space, weight, and credibility to all views, even if they represent unequal positions like majority versus minority opinions. Similarly, Winter et al. [2015] described balanced content as presenting conflicting evidence that does not necessarily lead to a clear consensus on the ‘correct’ view. In addition, the authors sub-divided one-sided messages into basic (presented unilaterally in a neutral style), assertive (conveyed unilaterally with assertive statements and strong intensifiers, leaving no room for doubt or ambiguity), and hedged (presented unilaterally, but with lexical hedges or acknowledgments of potential limitations).

Sol Hart et al. [2020] showed that polarized messages are often disseminated by actors from different poles (e.g. political parties), who craft stories focused on the contentious interactions between competing figures, to accentuate conflict and dramatize the narratives. This approach is pointed to be strategically used to engage and maintain the audience’s attention by spotlighting the adversarial natures of discourses. On the other hand, the ‘center’ of a field, regarded as non-polarized, is occupied by discourses that are seen as ‘neutral’ or representative of common sense, primarily due to their lack of controversy or contestation.

Iles et al. [2022] presented the origin types of conflicting information about cancer. In this scenario, conflict in evidence occurs when different sources (e.g. sources A and B) agree that the evidence on a topic is mixed. Conflict between expert sources about the evidence arises when sources A and B present conflicting viewpoints or data regarding a topic. Lastly, conflict within the same expert source occurs when a single source changes its recommendation or stance on the evidence over time.

In addition, two studies describe the use of “polarization” in science communication. One study discussed climate change news [Roper et al., 2016], where polarization is seen as a strategy used by climate skeptics to attack the arguments of proponents of anthropogenic global warming. Another study examined COVID-19 news [Sol Hart et al., 2020], showing how discussions vary with the involvement of political parties. The term “polarized messages” was represented by 63 synonyms or substitute words. The words that composed synonyms referencing polarization that were more frequently found in the papers were “viewpoint”, “opposing”, and “one-sided” (Figure 2).

Some expressions such as “one-sided message” [Winter et al., 2015; Anderson & Becker, 2018; Becker & Anderson, 2019; Wang & Huang, 2021] and “single-view message” [Dunwoody & Kohl, 2017] described content presenting typically only one speaker’s perspective. Other synonyms like “opposing viewpoints” [Kortenkamp & Basten, 2015; Dunwoody & Kohl, 2017] and “conflicting information” [Iles et al., 2022] indicated information polarization in the context of unbalanced arguments.

Non-polarized messages were characterized for 40 synonyms or substitute words. The words that composed synonyms referencing non-polarization more frequently in the papers were two-sided, balance, and viewpoint (Figure 3).

Terms like “two-sided messages” [Winter et al., 2015; Anderson & Becker, 2018; Lyons et al., 2019; Wang & Huang, 2021] indicated a balanced presentation of two viewpoints. Other synonyms such as “balanced opposing scientific viewpoint” [Kortenkamp & Basten, 2015] reflected a neutral presentation of arguments. The term “conflict in evidence” [Iles et al., 2022] described ambiguous evidence without implying disagreement by authors with two or more different positions.

It is important to note that some words like “viewpoint”, “opposing”, and “opposition” were frequently found among terms equivalently used to express polarized and non-polarized messages, depending on the context they were presented.

3.5 How are studies on polarized scientific digital messages characterized by their countries of origin, fields of knowledge, and themes?

Most studies were conducted in the U.S.A. (n = 7) [Kortenkamp & Basten, 2015; Dunwoody & Kohl, 2017; Anderson & Becker, 2018; Becker & Anderson, 2019; Sol Hart et al., 2020; Wang & Huang, 2021; Iles et al., 2022], followed by Germany (n = 1) [Winter et al., 2015], and New Zealand (n = 1) [Roper et al., 2016]. A multicenter study involved collaboration between researchers from the U.S.A. and the U.K. [Lyons et al., 2019]. The most common study design was a randomized information intervention (n = 7) [Kortenkamp & Basten, 2015; Winter et al., 2015; Dunwoody & Kohl, 2017; Anderson & Becker, 2018; Becker & Anderson, 2019; Lyons et al., 2019; Iles et al., 2022], followed by a discourse analysis (n = 1) [Roper et al., 2016], a content analysis (n = 1) [Sol Hart et al., 2020], and a crossover study (n = 1) [Wang & Huang, 2021]. The studies were published between 2015 and 2022, with the years 2015 [Kortenkamp & Basten, 2015; Winter et al., 2015] and 2019 [Becker & Anderson, 2019; Lyons et al., 2019] having the highest number of papers. The longest study spanned from 2009 to 2016 [Roper et al., 2016], while others ranged from 15 to 120 days [Dunwoody & Kohl, 2017; Sol Hart et al., 2020; Wang & Huang, 2021; Iles et al., 2022]. However, some papers did not specify the period of experiments or data collection [Kortenkamp & Basten, 2015; Winter et al., 2015; Wang & Huang, 2021].

The papers reported results from three fields of knowledge: Agricultural and Biological Sciences (n = 6), covering environmental risks [Kortenkamp & Basten, 2015], climate change [Roper et al., 2016; Anderson & Becker, 2018; Becker & Anderson, 2019], pharmaceutical water pollution [Dunwoody & Kohl, 2017], and genetically modified foods [Lyons et al., 2019]; Medicine (n = 3), focusing on COVID-19 [Sol Hart et al., 2020], e-cigarettes use [Wang & Huang, 2021], and cancer [Iles et al., 2022]; and Psychology (n = 1), exploring the psychological effects of computer games on children and adolescents [Winter et al., 2015].

Litmaps identified 417 references cited by the 10 papers included in this scoping review (Figure 4).

The earliest of these references was published in 1946 in a psychology journal [Sherif & Cantril, 1946]. The most cited reference was a programming book published in 2013, which has accumulated 29,513 citations [Hayes, 2017]. Notably, only two of the included papers [Dunwoody & Kohl, 2017; Becker & Anderson, 2019] cited another paper also included in this review [Kortenkamp & Basten, 2015; Anderson & Becker, 2018], one of which was produced by the same research group. Additionally, some papers cited studies from their own groups [Winter et al., 2015; Roper et al., 2016; Anderson & Becker, 2018; Becker & Anderson, 2019] or from other groups that had studies included in this review [Kortenkamp & Basten, 2015; Anderson & Becker, 2018; Becker & Anderson, 2019; Sol Hart et al., 2020; Iles et al., 2022].

Study limitations commonly referred by authors were small or demographically unrepresentative samples [Kortenkamp & Basten, 2015; Winter et al., 2015; Dunwoody & Kohl, 2017; Anderson & Becker, 2018; Becker & Anderson, 2019], inappropriate grouping by variables [Wang & Huang, 2021], or overly specific focuses [Kortenkamp & Basten, 2015; Roper et al., 2016; Anderson & Becker, 2018; Becker & Anderson, 2019]. Participants’ low reactance or unfamiliarity with topics might skew responses [Lyons et al., 2019; Iles et al., 2022]. Environmental contexts, like the COVID-19 infodemic, could influence participant sensitivity and responsiveness [Iles et al., 2022]. Limitations also raised from single-topic studies [Winter et al., 2015; Iles et al., 2022], short stimuli [Dunwoody & Kohl, 2017; Becker & Anderson, 2019], and lack of long-term effect analysis [Iles et al., 2022]. Finally, one quantitative study measured polarization extent but not its nature [Sol Hart et al., 2020].

3.6 Can polarized scientific digital messages be linked to specific authors, receivers, textual marks, or interest issues?

The authorship of messages found in these studies was not always clear. One study distinguished authors’ message between journalists and scientists to assess each group’s credibility among laypeople [Kortenkamp & Basten, 2015], while another inferred journalistic authorship from the source of the analyzed messages (broadcast news) [Wang & Huang, 2021]. In addition, other authorships were deduced from representative sources, such as news articles from the New Zealand Climate Science Coalition websites [Roper et al., 2016], videos from The Onion and The Weather Channel [Anderson & Becker, 2018; Becker & Anderson, 2019], and a video from an environmental conference [Lyons et al., 2019]. In contrast, all papers contained information about the media vehicles where messages were made available, including digital channels of traditional news media [Kortenkamp & Basten, 2015; Dunwoody & Kohl, 2017; Anderson & Becker, 2018; Becker & Anderson, 2019; Sol Hart et al., 2020; Iles et al., 2022], blogs and websites [Winter et al., 2015; Roper et al., 2016], and social media [Lyons et al., 2019; Wang & Huang, 2021]. The authors of half of the studies (n = 5) created specific polarized messages exclusively for being applied in experimental research conditions [Kortenkamp & Basten, 2015; Winter et al., 2015; Dunwoody & Kohl, 2017; Wang & Huang, 2021; Iles et al., 2022].

The messages in randomized information intervention studies mainly targeted unspecified adults [Dunwoody & Kohl, 2017; Lyons et al., 2019; Wang & Huang, 2021; Iles et al., 2022], or specific adult groups including undergraduate students [Kortenkamp & Basten, 2015; Anderson & Becker, 2018; Becker & Anderson, 2019] and parents or caregivers of children and adolescents [Winter et al., 2015]. In two other papers [Roper et al., 2016; Sol Hart et al., 2020], the specific receptors of the messages were not identified, but they are presumed to be laypeople or news consumers based on the context of studies, using a website and traditional news media as vehicles of information.

One study demonstrated that scientists with one-sided views were seen as less biased and more credible than those with balanced views. Conversely, journalists with balanced views were deemed less biased. Despite scientists generally being viewed as more credible, those opposing the majority were seen as less trustworthy [Kortenkamp & Basten, 2015]. The impact of message-sidedness on scientists’ credibility varied across studies, ranging from no effect [Wang & Huang, 2021] to a negative influence of conflict in evidence and between sources on public health perceptions [Iles et al., 2022].

The studies often reported that receptors agreed with the majority opinions [Kortenkamp & Basten, 2015] or expert-backed views [Dunwoody & Kohl, 2017]. Also, the engagement with two-sided messages was higher among those with advanced epistemological beliefs or related experiences [Winter et al., 2015; Wang & Huang, 2021]. Such messages fostered a balanced perspective [Winter et al., 2015] and did not negatively affect behavioral intentions [Iles et al., 2022]. However, sarcastic one-sided messages could reduce interest in topics like global warming [Anderson & Becker, 2018], while satirical messages could increase engagement [Becker & Anderson, 2019].

4 Discussion

These findings indicate that research on the characteristics of polarized scientific digital messages is relatively recent, limited in scope, and geographically constrained. Regarding this scenario, 4 out of 5 questions outlined in the protocol of this study were addressed fully or partially. The earliest identified study was published in 2015, with most studies concentrated in the United States. These studies predominantly focused on fields related to human health (such as medicine and psychology) and planetary health (including climate change, environmental risks, and GMOs). They mainly examined traditional news media, identifying authors as journalists or scientists, while recipients were often lay adults. Despite the scarcity of syntactic and lexical analysis, the research revealed a plethora of terms describing polarized and non-polarized messages. This variety of terms and some found concepts highlighted various aspects that researchers associated with polarization, such as controversy, topic dependency, skepticism, partisanship, and unilateralism. Characteristics of polarized messages were observed to include presentation from a single viewpoint, discrediting opposing views, emphasizing the minority and flaws of concurrent discourses, and using assertive statements, intensifiers, sarcasm, vague lexicons, and expert opinion support. Conversely, non-polarized messages tend to feature neutral and balanced arguments, use hedges and terms indicating uncertainty, include phrases that highlight the limitations of the statements, and present a diversity of opinions or points of view within the same message. Based on these results, however, the elucidation of distinct patterns of digital messages found in scientific papers, science communication articles, and specialized discourses was not possible because of the lack of comprehensive literature on the subject.

In general, leveraging consensus in science communication seems effective as it resonates with individuals lacking in-depth understanding or experience in the topic. While uncertainty drives scientific advancement, societies often influenced by cultural fears, crave certainty for comfort and reassurance [Wilson, 2000]. This dynamic can lead science and media to disseminate one-sided messages. However, simplifying and polarizing scientific information for the public might not be ideal, as it risks undermining critical thinking. In addition, treating consensus as an unchallengeable fact may foster group solidarity among its supporters while alienating skeptics, thereby intensifying issue polarization. This situation poses a challenge within the research community; the dominance of a hegemonic discourse may intimidate scientific progress by discouraging the questioning of established norms, in contradiction to the principle of falsifiability [Rekker, 2021], which is crucial for challenging and refining established theories.

The issue of polarization related to scientific topics seems underexplored. Studies on polarized scientific messages have typically focused on traditional media, with journalists as the primary sources of information. In this scenario, the production and publication of information are decentralized, allowing all users the potential to become communicators [Frewer et al., 2003]. This digital realm is typified by its open and asynchronous nature, fostering an environment where individuals can express their identities while maintaining privacy. The widespread access to information, combined with the intellectual pursuit of forming and expressing personal opinions, plays a pivotal role in shaping individuals’ health values. This is guided by the principle of ‘self-care’, a fundamental human impulse to participate in one’s own care, driven by the intrinsic need for self-preservation. [Foucault, 2005]. Consequently, there is a growing expectation for active engagement in healthcare, both directly and indirectly, reflecting the profound influence of Internet access on individual health participation and the broader healthcare discourse. This shift could result in a scenario where political ideologies and financial interests increasingly drive the polarization of science communication in digital media. This is evident in the emergence of violence-related concepts within the discourse, such as segregation and confrontation, which are also linked to hate speech. Remarkably, most studies referenced in this review were published subsequent to the intensification of the ‘fake news’ discourse, which gained prominence during the 2016 U.S. presidential campaign led by then-candidate Donald Trump [Bovet & Makse, 2019]. In this sense, it is imperative to conduct further research to investigate how decentralized communication affects the quality and polarization of scientific messages on the Internet.

This review presents some limitations. First, the strict inclusion of peer-reviewed scientific papers, although beneficial for limiting uncertainties about the evidence’s strength and practical significance of studies, may omit important insights from gray literature and preprints. Second, these findings also considered artificially constructed messages for research approaches, which may not accurately represent spontaneous discourse in real-life contexts of communication. Third, the analysis of messages not initially designed for digital dissemination, but later published on online media channels, might impact on the authenticity of the characteristics described in this review. Lastly, the search strategy’s confinement to certain databases may have introduced selection bias, overly emphasizing the connection of the studies to human and planetary health while neglecting other fields like mathematics and social sciences.

In summary, this scoping review elucidated the characteristics and concepts of polarized scientific digital messages, as validated and employed by researchers to craft content for studying the impact of polarization on consumer perceptions. The identified characteristics not only inform the development of future infodemiology studies focused on detecting polarization in scientific discourse on the internet but also underscore the necessity of researching the public impacts of both polarized and balanced scientific communications within specific contexts. Additionally, it is crucial to investigate how political ideologies and financial interests may contribute to the polarization of science communication in digital media.

5 Declarations

5.1 Declaration of conflicting interests

The authors declare that there is no conflict of interests.

5.2 Funding

This work was supported by the São Paulo Research Foundation (Grant # 2022/12885-6).

References

Abramowitz, A. I., & Saunders, K. L. (2008). Is polarization a myth? The Journal of Politics, 70(2), 542–555. https://doi.org/10.1017/s0022381608080493

Anderson, A. A., & Becker, A. B. (2018). Not just funny after all: sarcasm as a catalyst for public engagement with climate change. Science Communication, 40(4), 524–540. https://doi.org/10.1177/1075547018786560

Arguedas, A. R., Robertson, C. T., Fletcher, R., & Nielsen, R. K. (2022). Echo chambers, filter bubbles, and polarisation: a literature review. https://doi.org/10.60625/risj-etxj-7k60

Baldi, V., & Gala, A. (2021). Anticorpos à cultura da desconfiança: o regresso da ciência e da prudência na agenda mediática. Observatorio (OBS*), 15, 115–126. https://doi.org/10.15847/obsOBS0020211954

Becker, A., & Anderson, A. A. (2019). Using humor to engage the public on climate change: the effect of exposure to one-sided vs. two-sided satire on message discounting, elaboration and counterarguing. JCOM, 18(04), A07. https://doi.org/10.22323/2.18040207

Bliuc, A.-M., Bouguettaya, A., & Felise, K. D. (2021). Online intergroup polarization across political fault lines: an integrative review. Frontiers in Psychology, 12, 641215. https://doi.org/10.3389/fpsyg.2021.641215

Bolsen, T., & Shapiro, M. A. (2018). The US news media, polarization on climate change, and pathways to effective communication. Environmental Communication, 12(2), 149–163. https://doi.org/10.1080/17524032.2017.1397039

Bovet, A., & Makse, H. A. (2019). Influence of fake news in Twitter during the 2016 US presidential election. Nature Communications, 10(1), 7. https://doi.org/10.1038/s41467-018-07761-2

Boyd, K. (2022). Trusting scientific experts in an online world. Synthese, 200, 14. https://doi.org/10.1007/s11229-022-03592-3

Bruns, A. (2019). It’s not the technology, stupid: how the ‘echo chamber’ and ‘filter bubble’ metaphors have failed us. International Association for Media and Communication Research. https://eprints.qut.edu.au/131675/

Bryanov, K., & Vziatysheva, V. (2021). Determinants of individuals’ belief in fake news: a scoping review determinants of belief in fake news. PLoS ONE, 16(6), e0253717. https://doi.org/10.1371/journal.pone.0253717

Chinn, S., Sol Hart, P., & Soroka, S. (2020). Politicization and polarization in climate change news content, 1985–2017. Science Communication, 42(1), 112–129. https://doi.org/10.1177/1075547019900290

Choi, D., Chun, S., Oh, H., Han, J., & Kwon, T. (2020). Rumor propagation is amplified by echo chambers in social media. Scientific Reports, 10, 310. https://doi.org/10.1038/s41598-019-57272-3

Dalrymple, K. E., Young, R., & Tully, M. (2016). “Facts, not fear”: negotiating uncertainty on social media during the 2014 Ebola crisis. Science Communication, 38(4), 442–467. https://doi.org/10.1177/1075547016655546

De Cruz, H. (2020). Believing to belong: addressing the novice-expert problem in polarized scientific communication. Social Epistemology, 34(5), 440–452. https://doi.org/10.1080/02691728.2020.1739778

Dixon, G. (2016). Applying the gateway belief model to genetically modified food perceptions: new insights and additional questions. Journal of Communication, 66(6), 888–908. https://doi.org/10.1111/jcom.12260

Dunwoody, S., & Kohl, P. A. (2017). Using weight-of-experts messaging to communicate accurately about contested science. Science Communication, 39(3), 338–357. https://doi.org/10.1177/1075547017707765

Foucault, M. (2005). The hermeneutics of the subject: lectures at the Collège de France 1981–1982. Picador.

Franta, B. (2021). Early oil industry disinformation on global warming. Environmental Politics, 30(4), 663–668. https://doi.org/10.1080/09644016.2020.1863703

Frewer, L., Hunt, S., Brennan, M., Kuznesof, S., Ness, M., & Ritson, C. (2003). The views of scientific experts on how the public conceptualize uncertainty. Journal of Risk Research, 6(1), 75–85. https://doi.org/10.1080/1366987032000047815

Hall Jamieson, K., & Hardy, B. W. (2014). Leveraging scientific credibility about Arctic sea ice trends in a polarized political environment. Proceedings of the National Academy of Sciences, 111(supplement_4), 13598–13605. https://doi.org/10.1073/pnas.1320868111

Hamby, A., Kim, H., & Spezzano, F. (2024). Sensational stories: the role of narrative characteristics in distinguishing real and fake news and predicting their spread. Journal of Business Research, 170, 114289. https://doi.org/10.1016/j.jbusres.2023.114289

Hardy, B. W., & Hall Jamieson, K. (2017). Overcoming endpoint bias in climate change communication: the case of Arctic sea ice trends. Environmental Communication, 11(2), 205–217. https://doi.org/10.1080/17524032.2016.1241814

Hardy, B. W., Tallapragada, M., Besley, J. C., & Yuan, S. (2019). The effects of the “war on science” frame on scientists’ credibility. Science Communication, 41(1), 90–112. https://doi.org/10.1177/1075547018822081

Harvey, J. A., van den Berg, D., Ellers, J., Kampen, R., Crowther, T. W., Roessingh, P., Verheggen, B., Nuijten, R. J. M., Post, E., Lewandowsky, S., Stirling, I., Balgopal, M., Amstrup, S. C., & Mann, M. E. (2018). Internet blogs, polar bears, and climate-change denial by proxy. BioScience, 68(4), 281–287. https://doi.org/10.1093/biosci/bix133

Haupt, M. R., Li, J., & Mackey, T. K. (2021). Identifying and characterizing scientific authority-related misinformation discourse about hydroxychloroquine on Twitter using unsupervised machine learning. Big Data & Society, 8(1). https://doi.org/10.1177/20539517211013843

Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis: a regression-based approach. Guilford Press.

Hernandez, R. G., Hagen, L., Walker, K., O’Leary, H., & Lengacher, C. (2021). The COVID-19 vaccine social media infodemic: healthcare providers’ missed dose in addressing misinformation and vaccine hesitancy. Human Vaccines & Immunotherapeutics, 17(9), 2962–2964. https://doi.org/10.1080/21645515.2021.1912551

Hogg, M. A., Turner, J. C., & Davidson, B. (1990). Polarized norms and social frames of reference: a test of the self-categorization theory of group polarization. Basic and Applied Social Psychology, 11(1), 77–100. https://doi.org/10.1207/s15324834basp1101_6

Iles, I. A., Gillman, A. S., O’Connor, L. E., Ferrer, R. A., & Klein, W. M. P. (2022). Understanding responses to different types of conflicting information about cancer prevention. Social Science & Medicine, 311, 115292. https://doi.org/10.1016/j.socscimed.2022.115292

Iyengar, S., & Massey, D. S. (2019). Scientific communication in a post-truth society. Proceedings of the National Academy of Sciences, 116(16), 7656–7661. https://doi.org/10.1073/pnas.1805868115

Jia, H. (2022). More engagement but less participation: China’s alternative approach to public communication of science and technology. Public Understanding of Science, 31(3), 331–339. https://doi.org/10.1177/09636625221090729

Jiang, K., Anderton, B. N., Ronald, P. C., & Barnett, G. A. (2018). Semantic network analysis reveals opposing online representations of the search term “GMO”. Global Challenges, 2(1), 1700082. https://doi.org/10.1002/gch2.201700082

Kata, A. (2012). Anti-vaccine activists, Web 2.0, and the postmodern paradigm — an overview of tactics and tropes used online by the anti-vaccination movement. Vaccine, 30(25), 3778–3789. https://doi.org/10.1016/j.vaccine.2011.11.112

Kim, Y., & Kim, Y. (2019). Incivility on Facebook and political polarization: the mediating role of seeking further comments and negative emotion. Computers in Human Behavior, 99, 219–227. https://doi.org/10.1016/j.chb.2019.05.022

Kortenkamp, K. V., & Basten, B. (2015). Environmental science in the media: effects of opposing viewpoints on risk and uncertainty perceptions. Science Communication, 37(3), 287–313. https://doi.org/10.1177/1075547015574016

Kubin, E., & von Sikorski, C. (2021). The role of (social) media in political polarization: a systematic review. Annals of the International Communication Association, 45(3), 188–206. https://doi.org/10.1080/23808985.2021.1976070

Landrum, A. R., & Slater, M. H. (2020). Open questions in scientific consensus messaging research. Environmental Communication, 14(8), 1033–1046. https://doi.org/10.1080/17524032.2020.1776746

Lee, N. M., VanDyke, M. S., & Cummins, R. G. (2018). A missed opportunity?: NOAA’s use of social media to communicate climate science. Environmental Communication, 12(2), 274–283. https://doi.org/10.1080/17524032.2016.1269825

Lenzi, D. (2019). Deliberating about climate change: the case for ‘thinking and nudging’. Moral Philosophy and Politics, 6(2), 313–336. https://doi.org/10.1515/mopp-2018-0034

Lyons, B. A., Hasell, A., Tallapragada, M., & Hall Jamieson, K. (2019). Conversion messages and attitude change: strong arguments, not costly signals. Public Understanding of Science, 28(3), 320–338. https://doi.org/10.1177/0963662518821017

Marozzo, F., & Bessi, A. (2018). Analyzing polarization of social media users and news sites during political campaigns. Social Network Analysis and Mining, 8, 1. https://doi.org/10.1007/s13278-017-0479-5

McPherson, M., Smith-Lovin, L., & Cook, J. M. (2001). Birds of a feather: homophily in social networks. Annual Review of Sociology, 27, 415–444. https://doi.org/10.1146/annurev.soc.27.1.415

Merkley, E., & Stecula, D. A. (2018). Party elites or manufactured doubt? The informational context of climate change polarization. Science Communication, 40(2), 258–274. https://doi.org/10.1177/1075547018760334

Molina, M. D., Sundar, S. S., Le, T., & Lee, D. (2021). “Fake news” is not simply false information: a concept explication and taxonomy of online content. American Behavioral Scientist, 65(2), 180–212. https://doi.org/10.1177/0002764219878224

Morris, R. D. (2024). How denialist amplification spread COVID misinformation and undermined the credibility of public health science. Journal of Public Health Policy, 45(1), 114–125. https://doi.org/10.1057/s41271-023-00451-4

Moscovici, S., & Zavalloni, M. (1969). The group as a polarizer of attitudes. Journal of Personality and Social Psychology, 12(2), 125–135. https://doi.org/10.1037/h0027568

Myers, T. A., Maibach, E., Peters, E., & Leiserowitz, A. (2015). Simple messages help set the record straight about scientific agreement on human-caused climate change: the results of two experiments (K. Eriksson, Ed.). PLoS ONE, 10(3), e0120985. https://doi.org/10.1371/journal.pone.0120985

Nagler, R. H., Vogel, R. I., Rothman, A. J., Yzer, M. C., & Gollust, S. E. (2023). Vulnerability to the effects of conflicting health information: testing the moderating roles of trust in news media and research literacy. Health Education & Behavior, 50(2), 224–233. https://doi.org/10.1177/10901981221110832

Nagler, R. H., Yzer, M. C., & Rothman, A. J. (2019). Effects of media exposure to conflicting information about mammography: results from a population-based survey experiment. Annals of Behavioral Medicine, 53(10), 896–908. https://doi.org/10.1093/abm/kay098

O’Connor, C., & Weatherall, J. O. (2018). Scientific polarization. European Journal for Philosophy of Science, 8(3), 855–875. https://doi.org/10.1007/s13194-018-0213-9

Pearce, W., Brown, B., Nerlich, B., & Koteyko, N. (2015). Communicating climate change: conduits, content, and consensus. WIREs Climate Change, 6(6), 613–626. https://doi.org/10.1002/wcc.366

Pearce, W., Niederer, S., Özkula, S. M., & Sánchez Querubín, N. (2019). The social media life of climate change: platforms, publics, and future imaginaries. WIREs Climate Change, 10(2), e569. https://doi.org/10.1002/wcc.569

Peters, M. D. J., Godfrey, C., McInerney, P., Munn, Z., Tricco, A. C., & Khalil, H. (2024). Scoping reviews. In E. Aromataris, C. Lockwood, K. Porritt, B. Pilla & Z. Jordan (Eds.), JBI manual for evidence synthesis. JBI. https://doi.org/10.46658/JBIMES-24-09

Ploug, T., & Holm, S. (2015). Conflict of interest disclosure and the polarisation of scientific communities. Journal of Medical Ethics, 41(4), 356–358. https://doi.org/10.1136/medethics-2014-102114

Pontalti Monari, A. C., Santos, A., & Sacramento, I. (2020). COVID-19 and (hydroxy)chloroquine: a dispute over scientific truth during Bolsonaro’s weekly Facebook live streams. JCOM, 19(07), A03. https://doi.org/10.22323/2.19070203

Postmes, T., Spears, R., Lee, A. T., & Novak, R. J. (2005). Individuality and social influence in groups: inductive and deductive routes to group identity. Journal of Personality and Social Psychology, 89(5), 747–763. https://doi.org/10.1037/0022-3514.89.5.747

Rekker, R. (2021). The nature and origins of political polarization over science. Public Understanding of Science, 30(4), 352–368. https://doi.org/10.1177/0963662521989193

Rode, J. B., Iqbal, S., Butler, B. J., & Ditto, P. H. (2021). Using a news article to convey climate science consensus information. Science Communication, 43(5), 651–673. https://doi.org/10.1177/10755470211027235

Roper, J., Ganesh, S., & Zorn, T. E. (2016). Doubt, delay, and discourse: skeptics’ strategies to politicize climate change. Science Communication, 38(6), 776–799. https://doi.org/10.1177/1075547016677043

Scheufele, D. A. (2014). Science communication as political communication. Proceedings of the National Academy of Sciences, 111(supplement_4), 13585–13592. https://doi.org/10.1073/pnas.1317516111

Schneider, J. (2021). Big COVID, red state: the value of over-communication in a public health crisis. Frontiers in Communication, 6, 653665. https://doi.org/10.3389/fcomm.2021.653665

Sharpe, V. A. (2002). Science, bioethics, and the public interest: on the need for transparency. The Hastings Center Report, 32(3), 23–26. https://doi.org/10.2307/3528110

Sherif, M., & Cantril, H. (1946). The psychology of ‘attitudes’: part II. Psychological Review, 53(1), 1–24. https://doi.org/10.1037/h0058561

Sol Hart, P., Chinn, S., & Soroka, S. (2020). Politicization and polarization in COVID-19 news coverage. Science Communication, 42(5), 679–697. https://doi.org/10.1177/1075547020950735

Tricco, A. C., Lillie, E., Zarin, W., O’Brien, K. K., Colquhoun, H., Levac, D., Moher, D., Peters, M. D. J., Horsley, T., Weeks, L., Hempel, S., Akl, E. A., Chang, C., McGowan, J., Stewart, L., Hartling, L., Aldcroft, A., Wilson, M. G., Garritty, C., … Straus, S. E. (2018). PRISMA extension for scoping reviews (PRISMA-ScR): checklist and explanation. Annals of Internal Medicine, 169(7), 467–473. https://doi.org/10.7326/m18-0850

Turner, J. C., Wetherell, M. S., & Hogg, M. A. (1989). Referent informational influence and group polarization. British Journal of Social Psychology, 28(2), 135–147. https://doi.org/10.1111/j.2044-8309.1989.tb00855.x

van Kolfschooten, F. (2002). Can you believe what you read? Nature, 416(6879), 360–363. https://doi.org/10.1038/416360a

Wang, W., & Huang, Y. (2021). Countering the “harmless e-cigarette” myth: the interplay of message format, message sidedness, and prior experience with e-cigarette use in misinformation correction. Science Communication, 43(2), 170–198. https://doi.org/10.1177/1075547020974384

Wilson, K. M. (2000). Drought, debate, and uncertainty: measuring reporters’ knowledge and ignorance about climate change. Public Understanding of Science, 9(1), 1–13. https://doi.org/10.1088/0963-6625/9/1/301

Winter, S., & Krämer, N. C. (2016). Who’s right: the author or the audience? Effects of user comments and ratings on the perception of online science articles. Communications, 41(3), 339–360. https://doi.org/10.1515/commun-2016-0008

Winter, S., Krämer, N. C., Rösner, L., & Neubaum, G. (2015). Don’t keep it (too) simple: how textual representations of scientific uncertainty affect laypersons’ attitudes. Journal of Language and Social Psychology, 34(3), 251–272. https://doi.org/10.1177/0261927x14555872

Wirz, C. D., Cate, A., Brauer, M., Brossard, D., DiPrete Brown, L., Chen, K., Ho, P., Luter, D. G., Madden, H., Schoenborn, S., Shaw, B., Sprinkel, C., Stanley, D., & Sumi, G. (2022). Science communication during COVID-19: when theory meets practice and best practices meet reality. JCOM, 21(03), N01. https://doi.org/10.22323/2.21030801

Yuan, S., Besley, J. C., & Ma, W. (2019). Be mean or be nice? Understanding the effects of aggressive and polite communication styles in child vaccination debate. Health Communication, 34(10), 1212–1221. https://doi.org/10.1080/10410236.2018.1471337

Zhang, Y., Wang, L., Zhu, J. J. H., & Wang, X. (2021). Conspiracy vs science: a large-scale analysis of online discussion cascades. World Wide Web, 24(2), 585–606. https://doi.org/10.1007/s11280-021-00862-x

Zielińska, I. (2017). The influence of the social media on science communication: strategies of GMO opponents on Facebook. Adeptus, (10), 1513. https://doi.org/10.11649/a.1513

Ziman, J. (2002). The continuing need for disinterested research. Science and Engineering Ethics, 8(3), 397–399. https://doi.org/10.1007/s11948-002-0060-z

About the authors

Ana Maria Jucá. Department of Pediatric Dentistry, Orthodontics, and Public Health, Bauru School of Dentistry, University of São Paulo, Bauru, Brazil.
Ana Maria Jucá is a clinical dentist with a Master’s degree in Pediatric Dentistry from the Bauru School of Dentistry, University of São Paulo. Her research focuses on health information and health communication, with a particular interest in the polarization of health information on digital media.

E-mail: anamariajucanr@usp.br

Matheus Lotto. Department of Pediatric Dentistry, Orthodontics, and Public Health, Bauru School of Dentistry, University of São Paulo, Bauru, Brazil.
Matheus Lotto is a Ph.D. candidate at the Bauru School of Dentistry, University of São Paulo. His research focuses on information pollution, health information, and health communication, with a specific interest in analyzing health misinformation on digital media.

E-mail: matheus.lotto.souza@usp.br

Agnes Cruvinel. Bauru School of Medicine, University of São Paulo, Bauru, Brazil.
Agnes Cruvinel is an Assistant Professor at the Bauru School of Medicine, University of São Paulo. Her research spans clinical communication, medical humanities, and medical education, with a special focus on teaching communication skills to medical professionals.

E-mail: agnescruvinel@usp.br

Thiago Cruvinel. Department of Pediatric Dentistry, Orthodontics, and Public Health, Bauru School of Dentistry, University of São Paulo, Bauru, Brazil.
Thiago Cruvinel is an Associate Professor at the Bauru School of Dentistry, University of São Paulo. His research areas include information pollution, health information, and health communication. He is particularly interested in the identification, characterization, prebunking, and debunking of health misinformation on digital media.

E-mail: thiagocruvinel@fob.usp.br