Introduction
Mass media has become one of the most important agents in the formation and spreading of attitudes (Horcajo et al., Reference Horcajo, Briñol, Díaz, Becerra, Sabucedo and Morales2015). Thus, the influence that these media, truthful or not, have in societies is significant (Igartua, Reference Igartua, Sabucedo and Morales2015).
False information is not something new for humanity, however, nowadays it has an outstanding importance. During the second decade of the 21st century, mass media has suffered from a growth in Fake News (FN) and disinformation. The raise of new technologies linked to news has enable the growth of disinformation. These new technologies erased the barriers to the creation and diffusion of news and decreased the reputational and economical cost of sharing false information (Allcott & Gentzkow, Reference Allcott and Gentzkow2017). These phenomena have been stated to have outrageous consequences for the population such as the division of society (Olan et al., Reference Olan, Jayawickrama, Arakpogun, Suklan and Liu2024), the growth of cynicism toward politicians (Lee & Jones-Jang, Reference Lee and Jones-Jang2022), or the loss of trust in the media and institutions (Mwangi, Reference Mwangi2023).
One of the main problems linked to FN is the co-occurrence with polarization (see Au et al., Reference Au, Ho and Chiu2022; Deinla et al., Reference Deinla, Mendoza, Ballar and Yap2022; or Ribeiro et al., Reference Ribeiro, Calais, Almeida and Meira2017, for example). Knowing the important consequences of both phenomena, authors are increasingly paying attention to this relationship. Furthermore, since the 2016 United States’ elections and driven by the Pandemic, social, political, military, and scientific institutions are also beginning to pay attention to these two phenomena together.
Thus, the main objective of this research is to determine whether relationships between FN and polarization have been found in the literature, particularly regarding attitudinal polarization. The more specific objectives would be to examine the types of relationships identified between these terms, how they have been operationalized, and what the most recent findings on the topic are.
Fake News
Defining Fake News
We find multiple definitions of the term FN within the literature. Gelfert (Reference Gelfert2018, p. 108) will say that “Fake News is the deliberate presentation of usually false or misleading claims in the form of news.” According to Wang (Reference Wang2020), FN is news that not only contains false and/or misleading information but also has no intention to show the falsity of such information. Baptista and Gradim (Reference Baptista and Gradim2022) state that FN is a type of online disinformation whose content is based on false statements that, being able to be related to real events, are intended to manipulate readers using a structure and aesthetics that mimics that of a real news and whose purpose is to obtain some gain (economic or ideological). Tandoc et al. (Reference Tandoc, Lim and Ling2018) give a definition of FN in which its characteristics are evidenced. According to these authors, FN must have a low factuality (through false, misleading, or fabricated content and connections between false facts), have a format that mimics that of a real news, and have an intention to influence society.
These authors also propose six types of FN based on terms of factuality: satires (present topical news using exaggerations and humor, e.g., The Daily Show), parodies (use humor to present news, but introduce false content or create completely fabricated stories, e.g., El Mundo Today), fabrication (articles that mimic news articles, have no real basis, attempt to create legitimacy, and are intended to manipulate, e.g., false news verified by fact-checkers), photographic manipulation (manipulation of videos or photographs in order to create false narratives, e.g., deepfakes), propaganda (news-format stories created by political entities in order to influence citizens to benefit a group, character, or political entity, e.g., political propaganda) and advertising (use of the news format to promote or advertise services or products, e.g., narrative advertising) (Tandoc et al., Reference Tandoc, Lim and Ling2018). On the other hand, Al-Zaman (Reference Al-Zaman2021) points out six types of FN based on topics and contents: health (news about medicine, viruses, infections, lifestyles, etc.), religion (news with religious or political religious content, about spirituality, religious practices, etc.), politics (news about political institutions, political figures, or political issues), crime (news about violence and criminality), entertainment (news linked to celebrities and popular culture), and miscellaneous (encompassing the rest of news not linked to any of the previous categories).
From these data, we can define a fake news as a piece of deliberately false information, which imitates a real news, and which has an intention of manipulation, influence, or profit. In this case, it is important to emphasize that the pivotal points of this concept are the deliberate action of falsifying information and the intention of manipulation, influence, or profit.
Consequences of Fake News
FN presents severe psychological, political, and social implications. FN can impact consumers’ anxiety and depression levels increasing them and making readers emotionally and physically exhausted (Lin et al., Reference Lin, Hu, Alias and Wong2020; Olagoke et al., Reference Olagoke, Olagoke and Hughes2020; Rocha et al., Reference Rocha, De Moura, Desidério, De Oliveira, Lourenço and de Figueiredo Nicolete2021).
In the political and societal levels, exposure to FN has shown to increase motivated political reasoning (Deinla et al., Reference Deinla, Mendoza, Ballar and Yap2022) and confusion (Rapp & Salovich, Reference Rapp and Salovich2018), decrease trust in trustworthy and mainstream media (Guess et al., Reference Guess, Nyhan and Reifler2020; Lee et al., Reference Lee, de Zúñiga and Munger2023; Ognyanova et al., Reference Ognyanova, Lazer, Robertson and Wilson2020) and threat democracy and political elections by influencing citizens opinions through unreliable information (Lee, Reference Lee2019), and increasing cynicism toward politicians (Balmas, Reference Balmas2014). FN has also been shown to be related to confirmation bias. This bias has been found to enhance belief in FN (French et al., Reference French, Storey and Wallace2025) and is also a relevant factor in reinforcing attitudes after exposure to FN (Zhou & Shen, Reference Zhou and Shen2022), leading to more radical positions.
Polarization
Defining Polarization
Polarization is a polysemic concept that has been approached throughout history from multiple disciplines.
From sociology, polarization, commonly named as social polarization, is understood as a phenomenon that arises in conditions of conflict where positions are reduced to opposing extremes (two confronted groups) and where intermediate points (the other possible groups) disappear and are delegitimized (Piñuela & Yela, Reference Piñuela, Yela and Yela2016). This phenomenon is independent of the theme, and we can observe it in diverse fields such as sports (Lindner & Hawkins, Reference Lindner and Hawkins2023) or social media (Marozzo & Bessi, Reference Marozzo and Bessi2018).
Within the fields of political science and philosophy, Almagro (Reference Almagro2023) identifies five distinct types of polarization, each defined by the specific variable that is impacted by polarization (“what is being polarized?”): (1) ideological polarization, the distance between certain ideological positions within a spectrum; (2) partisan polarization, the level of uniformity of ideas and ideology that is associated with a certain political group; (3) platform polarization, how far apart the proposals of different political parties are with respect to certain issues; (4) affective polarization, the tendency to express and have negative feelings toward the out-group and positive feelings toward the group to which one belongs; (5) belief polarization or mixed evidence disagreement, greater acceptance of what is in line with our attitudes, which leads to a reinforcement of our attitudes and to interpretative biases (Lord et al., Reference Lord, Ross and Lepper1979).
From psychology, polarization has been studied from the attitudes’ prism. Moscovici (Reference Moscovici, Granberg and Sarup1992) defined attitude polarization as “a change in attitudes, affected by the group and with effects on the group” (p. 115). Myers and Lamm (Reference Myers and Lamm1976) provided a more detailed emphasis on the importance of reinforcement, pointing out that attitude polarization is the process by which attitudes are reinforced toward the pole to which they were previously directed.
Bringing together the different perspectives presented, polarization can be understood as a phenomenon that is both dynamic and static simultaneously. Sociology and political science have focused more on the static aspect of the phenomenon, addressing the outcome—the formation of rigid, divided groups or poles. On the other hand, psychology has focused on the dynamic aspect of the phenomenon. It emphasizes the process through which these poles and rigid groups are formed. In this way, the perspectives presented are not contradictory, but rather complementary.
Consequences of Polarization
Polarization has consequences at both personal and social levels.
At a psychological level, polarization can lead to the extremization of attitudes. Knowing that attitudes are important influencers of other cognitive and behavioral processes, strengthened attitudes can have important consequences in how we process information, in how we evaluate people, and in how we behave (Howe & Krosnick, Reference Howe and Krosnick2017). Furthermore, polarization has been related to other biases such as motivated reasoning (Stanley et al., Reference Stanley, Henne, Yang and De Brigard2020) or selective exposure behaviors (Stroud, Reference Stroud2010), often pointed as sources of such phenomenon. For instance, Knobloch-Westerwick (Reference Knobloch-Westerwick2012), using unobtrusive measures of selective exposure, showed that exposure to ideologically aligned news content led to increased attitude accessibility. Furthermore, Gvirsman (Reference Gvirsman2014) has shown that exposure to partisan media predicted attitude extremity and attitude cohesiveness, and that this polarization due to selective exposure can be an outcome of familiarity and a desire to avoid dissonance.
At a psychosocial level, polarization can lead to hostility toward the outgroups (Amira et al., Reference Amira, Wright and Goya-Tocchetto2021). Polarization is presented as the outcome of social comparison processes (between the members of the in-group), persuasion processes (with the increase of novel arguments in favor of the predominant position in the group), or self-categorization processes (including a conformity process with the prototypical in-group norms) (García-Ael & Gaviria, Reference García-Ael, Gaviria, Molero, Lois, García-Ael and Gómez2017). It is also linked to the decrease in political participation of citizens (as cited in Casal Bértoa & Rama, Reference Casal Bértoa and Rama2021), the increase in social instability (Bértoa & Enyedi, Reference Bértoa and Enyedi2021; Sartori, Reference Sartori1976), the deterioration of democracy (Caamaño & Bértoa, Reference Caamaño and Bértoa2019), and the greater difficulty in generating stable legislative coalitions to promote the necessary public policies (Binder, Reference Binder2008). Nonetheless, polarization can also serve as a basis for social mobilization (Smith et al., Reference Smith, Thomas, Bliuc and McGarty2024)
The Current Study: Fake News and Polarization
The great complexity of both FN and polarization, the co-occurrence of multiple definitions and perspectives under the same terminology, and the important consequences for society, prove the importance of its study.
Nowadays there is still a lack of systematization in the study of these two phenomena together. Researchers may find some difficulties such as the confusing and nonconsensual definition of FN and polarization or the novelty in the study of both. The aim of this article is to review the general landscape of FN literature and analyze in detail the polarization phenomenon and the relationship between FN, disinformation, and attitudinal polarization in literature. To do so, we will conduct a scoping review to identify the main up-to-date themes in the FN literature, assess the usage of the term “polarization” in the FN literature and analyze the relationship between attitudinal polarization and FN. Other reviews have been conducted in the recent years. For instance, Altoe et al. (Reference Altoe, Moreira, Pinto and Jorge2024) reviewed how FN, spread of disinformation, and belief change are interconnected. While they address similar concepts, they only include papers published after 2020, leaving out some important studies. Their main focus is on belief change and, while they consider polarization, they do not address it as a possible outcome of disinformation but rather as a driver of belief change. Furthermore, we will also focus on analyzing the operationalization of FN and polarization and we will address the evolution of this field over the years. To our knowledge, this is the first scoping review that specifically focus on FN and attitudinal polarization, simultaneously exploring the evolution of the FN literature, the operationalization of these two concepts, and the interconnection between them. Thus, we expect that this scoping review will bring some clarity to this new field, will highlight gaps in the research landscape, and will evidence limitations.
The main questions to be answered through this review are the following:
Q1. What are the main categories and topics explored nowadays in the scientific FN and disinformation literature? How are these main categories related to each other?
Q2. How is the term “polarization” used and defined in the scientific FN and disinformation literature, and with which themes and topics it is usually explored?
Q3. What is the relationship between FN and/or disinformation and attitudinal polarization?
Q4. How are FN and disinformation usually operationalized in the scientific FN literature?
Method
The scoping review was conducted on three different dates, on January 15, February 5, and June 13, 2024. The first two dates correspond to the initial searches conducted by the authors. A subsequent search incorporating the term “misinformation” was deemed necessary. Initially excluded, misinformation does not convey the same negative intent as FN and disinformation (Aïmeur et al., Reference Aïmeur, Amri and Brassard2023). Nonetheless, due to its prevalence in scholarly discourse and its interchangeable usage with disinformation and FN, it was included to explore its connection to polarization. We followed the PRISMA-ScR 2018 (PRISMA extension for Scoping Reviews) criteria (Tricco et al., Reference Tricco, Lillie, Zarin, O’Brien, Colquhoun, Levac, Moher, Peters, Horsley, Weeks, Hempel, Akl, Chang, McGowan, Stewart, Hartling, Aldcroft, Wilson, Garritty and Straus2018).
Scoping Review
Study Identification
The search strategy was developed using SCOPUS and Web of Science databases. To conduct the search, we looked for the terms “Fake News” OR “misinformation” OR “disinformation” AND “polarisation” OR “polarization” OR “attitude polarization” in title, abstract, and keywords. In the search, we included only final publications. All documents had to be published in 2016 or after. We decided to use this year as the cut-off point since various authors have identified 2016 as a key year for the significant expansion of modern FN and the explosion of the post-truth phenomenon (Allcott & Gentzkow, Reference Allcott and Gentzkow2017; Modreanu, Reference Modreanu2017). We limited search results to English or Spanish language and to scientific articles, book chapters, and conference papers. We restricted the access to documents; they had to be open access.
To organize files and conduct screening and review analysis, we used HubMeta platform.
Inclusion Criteria
The studies included in the final in-depth review had to follow the next criteria:
-
1. The document presents empirical research. Theory or review articles were excluded of the in-depth review.
-
2. Document is written in English or Spanish and is full open access.
-
3. The research explores the connection between FN or disinformation and polarization.
-
4. The research is not a simulation, so explores actual empirical data.
-
5. The research must report sample sizes, methodology and analysis used, and concrete results about polarization and FN/disinformation.
Study Selection
In total, 899 studies were identified in the initial search through SCOPUS and Web of Science. Of this initial batch, 476 studies were removed due to duplication, making 423 studies suitable for Title and Abstract review. The two authors (RL and MG) independently evaluated the 423 studies. The Kappa ratio of agreement was 0.88. After this initial Title and Abstract review, 339 papers did not meet the inclusion criteria and were discarded. A total of 84 documents were full text reviewed. After this full text review, other 64 papers were excluded because they did not explore explicitly the relationship between FN/disinformation and polarization, because they were computer simulations and not actual empirical data or because connections between the main variables were not clearly stated. A total of 20 studies met all the criteria for inclusion (see Figure 1 for the schematic PRISMA diagram).
PRISMA-ScR flow diagram.
Theme Analysis
The theme analysis was conducted using Atlas.ti 24. Once the total number of papers to be included after deduplication had been obtained (n = 423), the topics of these documents were analyzed. To do this, by reading and analyzing the information contained in the abstracts of these papers, a series of codes were generated and applied to each study. We followed the bottom-up method for thematic analysis (Terry et al., Reference Terry, Hayfield, Clarke, Braun, Willig and Rogers2017). The codes applied summed up the main theme of the analyzed paper. The process resulted in 49 codes (topics) grouped in 11 categories. These categories are summed up and defined in Table 1.
Categories (group of codes) definitions and topics (codes) used in the qualitative theme analysis
Results
Fake News Literature: Main Categories and Topics
The first noticeable general finding is the growing popularity that the FN and disinformation themes have in scientific literature. In this review, we analyzed papers from 2016 onward. Since this year, the number of publications has grown significantly. We can observe a positive tendency in this field, finding the highest number of publications in 2023 with 121 publications (Figure 2).
Number of papers about FN and disinformation published by year.
Analysis of Categories
Using the previously specified coding scheme (Table 1), we examined the frequency of each category (group of codes). Of the total codes applied across all categories, the four most frequent categories accounted for more than 60% of all codes. These four dominant categories were those related to platforms and technologies, the study of actions carried out by individuals, the study of FN itself, or the study of applications related to FN and disinformation (Figure 3).
Categories’ (groups of codes) frequency.
Using a Sankey diagram (Figure 4), we explore the connections between these categories. The most interconnected categories were “Actions,” “Platforms-Tech,” “Psycho-social” (specifically the topic “echo chambers”), and “Polarization.” These categories had relationships with almost any other.
Sankey diagram representing co-occurrence of groups of codes identified in the review.
Analysis of Specific Topics
Regarding the popularity of specific topics (individual codes), since each study can cover multiple topics, we focused on how often each code appeared, rather than the number of papers within each topic. Excluding research designs, we found that strategies directed toward the fight against FN is the most popular topic (with a frequency of 10.63%). Then, there was the study of spread or diffusion of FN (with a frequency of 9.55%). On third position were papers related to the study of FN and disinformation on social media (with a frequency of 8.65%). Then there was the study of FN characteristics (with a frequency of 8.29%) and the study of believing or trusting FN (with a frequency of 5.95%).
For research designs, we reviewed 63 papers that used machine learning, modeling, simulation, or systematic review methods, and classified book chapters too. The use of a code within the research designs category prevented the use of others, which enabled us to explore the number of studies in each category instead of just code prevalence. Within these categories, 44.44% studies were systematic reviews (28 papers), 22.22% were papers that used modeling techniques (14 papers), 19.05% were simulations (12 papers), 11.11% used machine learning (7 papers), and 3.17% were book chapters (2 documents). Of the 28 systematic reviews analyzed, we found that the most popular topics were FN characteristics (4), FN detection (3), fighting against FN (3), social media (3), and political polarization (3). In fact, no reviews were found in our research exploring the relationship between attitudinal polarization and FN.
Polarization: Analysis as a Category and as a Topic
We wanted to further explore the category polarization within the FN literature.
The category “polarization” appeared in almost 6% of the analyzed papers. It was highly related to the category “Platforms-Tech,” followed by “Intrapersonal characteristics,” and “Applications.” This result mimics the distribution found when analyzing categories generally.
As a category, “polarization” was specially related to echo chambers, networks of users, and social media (Figure 5). There is also a strong relationship between polarization and FN consequences and polarization and fighting FN strategies.
Sankey diagram representing co-occurrence between polarization as a category and topics identified in the review.
Regarding specific topics around polarization, we found five different typologies. Polarization understood as differentiated clusters was the most popular one (5.18%) followed by attitudinal polarization (1.62%), political polarization (1.13%), affective polarization (0.65%), and opinion polarization (0.16%) (Figure 6).
Sankey diagram representing the co-occurrence between polarization topics and other topics identified in the review.
The most interconnected typology was the one referred to clusters. This type of polarization co-occurred with other 17 codes. Clusters or social polarization seems to be studied alongside issues such as social media, emotions, the spread of FN, the fight against FN, and the consumption of FN (Figure 6).
In-Depth Analysis of Full Text-Reviewed Papers
Of the 20 papers included in the final batch (Table 2), all of them included quantitative analysis and empirical approaches. Regarding samples, 70% of the analyzed studies used participants, 10% used tweets or social media posts, 5% used users accounts, 5% used observations, 5% used experts’ evaluations, and 5% used news as their samples. Regarding designs, 50% of the studies analyzed used an experimental design, 45% used a correlational design, and 5% used descriptive design. Focusing on geographical distribution, 65% of the studies were conducted in the United States. There was one study conducted in Australia, one in Austria, one in Brazil, one in Germany, one in Italy, and one in Switzerland. We can see that majority of research is concentrated in the United States.
Included studies for the in-depth analysis. Overview, polarization and stimuli description, and main results
Note: S1 = study 1; S2 = study 2; S3 = study 3; 1 = participants; 2 = tweets/social media posts; 3 = users accounts; 4 = observations; 5 = experts’ evaluations; 6 = news; QT = quantitative; QL = qualitative; OB = observational; L = longitudinal design; E = experimental design; C = correlational design; D = descriptive design; NR = not reported; NM = not measured; N = news; H = headlines; D = disinformation; A = accounts; P = posts; OT = other; (O) = organic stimuli; (RG) = research-generated stimuli.
All analyzed papers explored the relationship between polarization and FN or disinformation. While some of them explored this relationship explicitly, others explored this relationship as something collateral, for instance, when exploring corrections to misinformation, but producing backfire effects.
Regarding the variable polarization, we found six different types of polarization in the analyzed papers, 20% of them explored political polarization, 10% explored attitudinal polarization, 10% explored affective polarization, 5% explored ideological polarization, 5% explored network polarization, and 5% explored sociocognitive polarization. One paper specifically explored attitudinal, political, and ideological polarization at the same time. Of the research articles, 40% did not specify the type of polarization they were analyzing. Regarding definitions, 45% of the analyzed studies did not report a specific definition of the polarization they were analyzing. For the papers that reported definitions, there was not a standardized definition of the polarization, finding 11 different definitions in 20 analyzed papers. For measures, 35% of papers did not specifically report the way they measured polarization and 5% did not measured polarization specifically. Within those that did report their measurements, we found 12 different ways of measuring polarization.
Influence of Fake News on Polarization
We found mixed evidence of the relationship and influence of FN on polarization.
Three papers found no evidence of influence of FN in attitudes, behaviors, or views of individuals exposed to such news.
Six papers pointed to polarization as a prelude to disinformation. According to these authors polarization, polarized societies and polarized groups create the perfect scenario for disinformation. This polarized context favors the spread of mis- and disinformation and affect the way people interact with each other. Highly polarized environments promote the emergence of echo chambers and can lead users to stop interacting with other counterattitudinal users. This behavior could lead to a reinforcement of echo chambers and polarization by eliminating opposing views.
Eleven papers pointed out the influence of FN in polarization. According to these papers, exposition to disinformation or to FN do influence polarization. The spread and consumption of FN affect individuals by making their attitudes, beliefs, opinions, or ideologies more extreme and engaging in more radical behaviors (such as domestic terrorism). FN spread and consumption also decreases the intention to perform and support prosocial actions, favors in-group favoritism, and increase social distancing and social division. A few papers also explored FN corrections and retractions. They found that when individuals expose to counterattitudinal retractions or when individuals are highly committed to a specific issue, corrections and retractions backfire. In these scenarios, not only FN have an influence over individuals, but the strategies to combat them, far from improving the situation, seem to lead to greater levels of polarization.
We found mixed results when analyzing the influence and relationship of FN and attitudinal polarization. While Bail et al. (Reference Bail, Guay, Maloney, Combs, Hillygus, Merhout, Freelon and Volfovsky2020) claim that exposure to FN and propaganda has no effect on individuals’ attitudes, Greitemeyer (Reference Greitemeyer2023) found that at least a group of participants inside her sample, the sceptics one, reported stronger attitudes after reading proattitudinal fictitious news articles. Weismueller et al. (Reference Weismueller, Gruner, Harrigan, Coussement and Wang2024) showed the importance of emotions in the attitudinal polarization process. According to these authors, negative emotions are linked to polarization and FN usually leads to more negative emotions. Therefore, FN through negative emotions can increase polarization.
Discussion
The field of FN has become increasingly prominent in scientific research. Analysis of publications since 2016 indicates an increase in the popularity of FN studies. Various authors (Abu Arqoub et al., Reference Abu Arqoub, Abdulateef Elega, Efe Özad, Dwikat and Adedamola Oloyede2022; Broda & Strömbäck, Reference Broda and Strömbäck2024; Pérez Escolar et al., Reference Pérez Escolar, Lilleker and Tapia Frade2023) have also noted this growing interest among scholars over the second decade of the 21st century. This trend reflects not only the scientific drive to understand these phenomena, but also highlights FN’s current social significance. Related to FN, polarization has also been seen as an important threat to societies. In our study, we addressed four key questions relating to these two topics:
Q1. What are the main categories and topics explored nowadays in the scientific FN and disinformation literature? How are these main categories related to each other?
Despite the growing popularity of FN research, most analyzed papers are focused on four main categories or themes: (1) the relationship between FN, platforms, and technologies; (2) the analysis of how online users and individuals consume, share, and engage with FN; (3) the examination of FN characteristics and their social consequences; and (4) the exploration of detection methods, regulatory measures, and strategies to counter FN. Our results around categories distribution go in line with those presented by Broda and Strömbäck (Reference Broda and Strömbäck2024), who showed that themes such as detection, characteristics, prevalence, or dissemination were the most popular ones. These authors also underline the huge importance of social media in FN literature and research. We also found interconnectivity between action, digital environments, echo chambers, and polarization. For specific topics, we observed a distribution similar to that of the broader categories. Countermeasures, diffusion of FN, social media, FN characteristics, and belief on FN were the five more prominent topics.
Answering Q1, our results show four key categories in FN literature, some interconnectivity between specific issues, and five central topics. These results could indicate that there is a special interest in understanding how users spread, consume, engage, and believe FN in social media environments, how echo chambers play a role in these phenomena, and how this could be connected to (societal) polarization. However, concentrating research around specific themes and scenarios may lead to the construction of a biased and reductionist picture. Regarding reductionism in social sciences, Verschuren (Reference Verschuren2001) states the following: “…a legitimate question is whether a reductionistic approach leaves underexposed aspects of social reality. At least it leaves open the question whether it can grasp the whole of an object, as this is more than the sum of its parts” (p. 390). While we can see some interconnectivity within FN topics, some research gaps arise. From a more psychological point of view, topics such as cognitive bases of FN or neuroscientific approach in the FN research seem somehow underexplored. Other basic processes such as the effects of interpretation, attitudes, or knowledge on disinformation and FN need more research. In fact, other reviews have stated too the importance of exploring individual differences and characteristics within the disinformation phenomena (Broda & Strömbäck, Reference Broda and Strömbäck2024). We propose the use of experimental studies that combine exposure to FN with the analysis of psychological variables. A necessary step in this process is the development of a well-controlled and validated set of stimuli. While some ecological validity may be lost in the process, this could help clarify causal relationships and reveal how different FN configurations may influence outcomes. Furthermore, with regard to ecological validity, new technologies that mimic real digital environments can be integrated into laboratory experiments, helping researchers preserve realism while maintaining experimental control.
Q2. How is the term “polarization” used and defined in the scientific FN and disinformation literature, and with which themes and topics it is usually explored?
Regarding our Q2, three main results can be drawn from the analysis of polarization. First, nowadays polarization is heavily linked to social media and to digital processes and environments. This is no surprise as social media, Internet, and technologies are quite important in communication and information currently (Althaus & Tewksbury, Reference Althaus and Tewksbury2000; Jost et al., Reference Jost, Barberá, Bonneau, Langer, Metzger, Nagler, Sterling and Tucker2018; Szymkowiak et al., Reference Szymkowiak, Melović, Dabić, Jeganathan and Kundi2021).
Second, the landscape of polarization is confusing, with up to six different types of polarization identified. We found that 40% of the studies did not specify the type of polarization they were examining, and 45% did not provide a clear definition of the polarization they analyzed. Among those that did include definitions, there were inconsistencies. For example, attitudinal, political, and ideological polarization are often defined in similar terms, focusing on division, even though they represent distinct phenomena. Attitudinal polarization should be operationalized as the process in which attitudes strengthen. Therefore, it should provide pre–post attitudinal measures. Political and ideological polarization focus on the division and should be operationalized in terms of distance between opposing groups. Additionally, sociocognitive is conceptualized as an aggregate of individuals’ characteristics offering just a descriptive static picture. Finally, network polarization was not properly described in the studies we analyzed. This complex panorama, which on some occasions is accompanied by inconsistent definitions or even a lack of definition, overlooks the fact that each type of polarization involves different mechanisms, overlaps concepts, and creates a confusing panorama for polarization.
Given this complex scenario, and based on the measures that studies have reported and the definitions of the different types of polarization, we propose a classification that identifies two typologies of polarization: dynamic polarization and static polarization. Dynamic polarization encompasses attitudinal polarization. It refers to the dynamic process of intensification of attitudes and could be underneath the division of groups. In terms of operationalization, dynamic polarization requires the use of longitudinal or pre–post measures. In contrast, static polarization includes ideological polarization, political polarization, sociocognitive polarization, affective polarization, and network polarization. Ideological and political polarization focus on the outcomes of division and are more closely related to social polarization, highlighting the divergence of ideological and political positions. Sociocognitive polarization also emphasizes division, but is primarily characterized by intolerance of ambiguity or xenophobia. Affective polarization focuses division of groups based on emotions. It highlights the experience of positive emotions toward one’s own group and negative emotions toward opposing groups. Finally, network polarization centers on the exclusion of discordant individuals to create homogeneous and clearly defined groups. In terms of operationalization, static polarization can be measured using transversal designs and single measures.
This classification allows for a more comprehensive understanding of polarization as a unified phenomenon that can be conceptualized either as a dynamic process, when focusing on the mechanisms driving polarization; or as a static condition, when emphasizing the resulting fragmented social landscape.
Third, following the previously presented classification, literature is concentrating on static polarization, specifically on social polarization. This approach may be problematic, as it tends to address only one side of the equation. Social polarization, understood as the estrangement and confrontation between two opposing groups, should be seen simultaneously as both a potential cause and outcome of other types of polarization. Specifically related to attitudinal polarization, Myers and Lamm (Reference Myers and Lamm1976) noted that attitudinal polarization could result from the desire for positive self-perception and self-representation within a group. Turner et al. (Reference Turner, Wetherell and Hogg1989) further demonstrated that polarization can emerge from group influence, driven by adherence to group norms and the pursuit of acceptance. These authors show the importance of groups as a preliminary step to the generation of attitudinal polarization. Bliuc et al. (Reference Bliuc, Bouguettaya and Felise2021) found this same effect in online environments. In these contexts, deindividuation strengthens group identification, leading to greater value placed on group members’ judgments, which may ultimately polarize attitudes. Thus, individuals may adopt more extreme positions to present themselves as socially desirable within their group (Lee, Reference Lee2007). This means that an extreme social division and the generation of extreme group positions could be an important first step for attitudinal polarization.
Conversely, attitudinal polarization could precede social polarization. As attitudes become more polarized, they radicalize positions, and at the social level, this intensifies social division and therefore social polarization. For instance, when progressives become more progressive and conservatives more conservative, the reinforcement of these attitudes solidifies group identity and increases the division between groups. Considering this, focusing exclusively on social polarization is insufficient, as it offers only a static snapshot that may reflect both the causes and consequences of an ongoing process. We propose a comprehensive analysis that integrates different types of polarization, and we encourage further research to better understand their interconnections and potential feedback loops.
Q3. What is the relationship between FN and/or disinformation and attitudinal polarization?
Approaching Q3, we found mixed results about the relationship between attitudinal polarization and FN. We analyzed the three reports that indicated using attitudinal polarization, and the research conducted by Chalik et al. (Reference Chalik, Over and Dunham2022) that, despite not specifically reporting the usage of attitudinal polarization, their results show a process of attitude change and reinforcement.
The selected reports showed a relationship between disinformation, fake claims, FN or fictitious news, and attitudinal polarization, showing that exposition to these types of stimuli could lead to a reinforcement of previous attitudes. Greitemeyer (Reference Greitemeyer2023) found that after participants read attitudinal congruent articles related to COVID-19 vaccination, they reported their attitudes becoming more extreme. This was due to biased information assimilation, where congruent information was rated as more favorably. Weismueller et al. (Reference Weismueller, Gruner, Harrigan, Coussement and Wang2024) found that exposure to misinformation was related to higher engagement with the information (compared to accurate information), to increase of negative emotions, and to polarization. These negative emotions were a driver for the polarization not only when using false information, but also when using partisan biased information. Participants exposed to misinformation about COVID-19 experienced negatives emotions (like anger) and this led to their attitudes about COVID-19 becoming more extreme. Chalik et al. (Reference Chalik, Over and Dunham2022) conducted an experiment which tested how children chose between reliable and unreliable sources, and in-group-favoring and out-group-favoring sources. Children who presented stronger initial biases chose the biased in-group-favoring unreliable source over the reliable source. The ones that prioritized the biased source changed their attitudes, becoming even more biased than initially. Linked to these studies, Fasce et al. (Reference Fasce, Adrián-Ventura and Avendaño2020) found that social axioms, such as religiosity and fate control, mediate the relationship between right-wing authoritarianism, and pseudoscientific beliefs. These findings may help explain the intergroup conflict that drives belief polarization. This conflict pushes extreme believers to reject out-group information and rationalize their own beliefs, leading to a backfire effect where attitudes become more polarized. However, these cause-and-effect relationships have only been explored at a theoretical level.
These studies highlight the apparent relationship between biased, false, or partisan information and attitude change and polarization. These results suggest that cognitive dissonance (Festinger, Reference Festinger1962) may play a key role in how FN contributes to attitudinal polarization. When participants encounter consistent information, they accept it as valid, reinforcing their attitudes (Zhou & Shen, Reference Zhou and Shen2022). Conversely, when they encounter discordant information (either from outgroup sources or content that elicits negative emotions), they tend to reject it to reduce cognitive dissonance, reinforcing their previous attitudes (Nyhan et al., Reference Nyhan, Porter, Reifler and Wood2020). In sum, exposure to FN, whether consistent or contradictory, appears to fuel polarization through these complementary mechanisms of acceptance and rejection. Moreover, exposure to FN has an eminently psychosocial character. It generates a specific intergroup dynamic that not only highlights the differences in the attitudes of the involved groups, but also emphasizes how these differences are accentuated, shaping polarization processes. This can be seen in the fact that the relationship between FN and polarization seems to occur in the presence of emotions and reference groups. This relationship is more evident when high emotions are involved or when messages align with prior attitudes or reference groups. In both cases, attitudes are reinforced, leading to polarization.
On the other hand, Bail et al. (Reference Bail, Guay, Maloney, Combs, Hillygus, Merhout, Freelon and Volfovsky2020) found no effect of disinformative political propaganda in the political attitudes of their participants. Finding, in fact, that exposition to misinformative political propaganda had no effect in both political attitudes and behaviors.
Thus, the selected studies reveal contradictory findings. This suggests that factors beyond mere exposure may play a role in attitudinal polarization and contribute to these inconsistencies. These contradictions may be due to the influence of cognitive, sociodemographic, and ideological variables. For example, factors such as the level of dogmatism (Bronstein et al., Reference Bronstein, Pennycook, Bear, Rand and Cannon2019), level of analysis (Pennycook & Rand, Reference Pennycook and Rand2019), age (Brashier & Schacter, Reference Brashier and Schacter2020; Guess et al., Reference Guess, Nagler and Tucker2019), political biases (Van Der Linden et al., Reference Van Der Linden, Panagopoulos and Roozenbeek2020), or prior congruence (Zhou & Shen, Reference Zhou and Shen2022) have been shown to play an important role in beliefs, attitudes, and behaviors regarding FN. We highly encourage authors to include such variables in future research to explore their potential moderating role.
Additionally, the wide range of methodologies and approaches used across studies further complicates the extraction of clear and consistent conclusions. There is a lack of consistency in the use of stimuli. This can pose a problem when drawing conclusions. Variables related to FN, such as emotional valence (Ali et al., Reference Ali, Li, Zain-ul-abdin and Muqtadir2022), source credibility (Kim & Dennis, Reference Kim and Dennis2019), and relevance (Chaudhuri et al., Reference Chaudhuri, Gupta and Popovič2025), have been found to influence the impact of this type of news.
Moreover, most research has been conducted with the US populations, potentially capturing dynamics specific to the political, social, and informational contexts of the United States, rather than the underlying psychosocial mechanisms driving polarization more broadly. To address these issues, we encourage researchers to adopt systematic methodologies when examining the effects of FN and disinformation on attitudinal polarization. We advocate for the use of experimental designs with well-established and validated stimuli that allow for the isolation of specific characteristics and variables. We also encourage researchers to develop studies with different populations to capture other political realities.
Q4. How are FN and disinformation usually operationalized in the scientific FN literature?
Regarding operationalization of FN, we found that some authors have used completely fabricated FN (McCright et al., Reference McCright, Charters, Dentzman and Dietz2016; Nyhan & Reifler, Reference Nyhan and Reifler2010), while others have used fabricated FN based on fact-checkers articles (Boukes & Hameleers, Reference Boukes and Hameleers2023). Actual political propaganda, “real” FN, claims, and social media accounts or posts have also been used as stimuli when analyzing the effects of disinformation (Arceneaux & Truex, Reference Arceneaux and Truex2023; Bali et al., 2020; Chipidza et al., Reference Chipidza, Krewson, Gatto, Akbaripourdibazar and Gwanzura2022; Dourado & Salgado, Reference Dourado and Salgado2021; Overgaard & Collier, Reference Overgaard and Collier2023). Other types of stimuli, such as videos created by researchers (Chalik et al., Reference Chalik, Over and Dunham2022), fictional political scenarios (Ecker & Ang, Reference Ecker and Ang2019), or fictional studies (Greitemeyer, Reference Greitemeyer2023), were also used.
Different methodologies have also been used to generate them. Of the 20 papers analyzed in depth, 7 used fabricated stimuli, meaning that researchers created the stimuli; for instance, Boukes and Hameleers (Reference Boukes and Hameleers2023) used a fact-check article about decreasing crime rates as the base for developing two FN articles. Salvi et al. (Reference Salvi, Iannello, Cancer, McClay, Rago, Dunsmoor and Antonietti2021) followed a similar procedure, creating 12 headlines from scratch using online articles as a reference. The headlines were generated in such a way as to mimic the appearance of real news stories. One of them was politicized (increasing crime rates are attributed to Latino immigrants) and the other was nonpoliticized (a general increment in crime rate). On the other hand, 13 studies used organic stimuli, meaning that they used real post, news, or accounts for the disinformation. The use of organic stimuli has the advantage of providing more ecological validity. However, problems often arise with the control of contaminating variables. In addition, this type of stimuli sometimes requires retrospective studies that only provide a correlational approximation. On the other hand, the use of researcher-generated stimuli allows for greater control of the stimuli. However, the use of such stimuli requires testing and validation. In addition to the possibility of losing ecological validity through their use.
The variety of stimuli used in disinformation and FN research makes it difficult to compare results and draw general conclusions. Therefore, we encourage authors to develop and follow standardized protocols for the experimental design of their research.
Limitations
While this scoping review provides valuable insights into the relationship between FN and polarization, it is important to acknowledge certain limitations. Our primary focus was on the link between FN, disinformation, and polarization, with a specific emphasis on attitudinal polarization. Consequently, other dimensions of polarization, as well as broader areas of polarization research, remain underexplored. Therefore, although we identified critical issues within the study of polarization, we cannot confirm that these limitations extend uniformly across the entire field.
Additionally, this review included only papers written in English or Spanish leaving a considerable portion of the scientific literature unexamined due to language barriers.
Finally, we acknowledge the potential review bias from using only open-access papers, which may exclude important studies and favors research from well-funded regions that can afford open-access publishing.
The limitations presented in this review are, however, fully intentional. The researchers recognize that it is impossible to cover any given topic comprehensively within the scientific literature. Given the rise of FN and its apparent connection to polarization, it was necessary to focus on examining the relationship between these two variables, at the expense of exploring other potentially relevant relationships or variables. Moreover, constraints related to language and open access further shaped the scope of this review. These constraints—topic scope, language, open access—though intentional, should be considered when interpreting our findings.
Future Directions and Implications
Despite the boundaries, this scoping review highlights several insights, such as the need for methodological standardization, the differential popularity of themes in the research, and the contradictions found in the relationship between FN and attitudinal polarization. We encourage future researchers to address and overcome the identified gaps by conducting experimental research, with validated stimuli, using samples from different cultural backgrounds and focusing on the psychological variables that underpin the FN and polarization relationship.
The results highlighted in this scoping review have also important implications for policymakers and societies. The review has identified key factors in the relationship between FN and polarization, such as emotions and message sources. Recognizing these factors allows institutions to design targeted interventions, such as focusing on discrediting the sources rather than the message content. This approach could prevent backfire effects (also found in this review) when debunking information, while still addressing the issue.
Additionally, the review highlights the importance of prior attitudes and biases in information judgments. Sharing this knowledge with the public raises awareness of the need to develop critical thinking. It also supports the creation of intervention strategies aimed at improving this skill.
Together, these strategies can enhance institutional efforts to combat misinformation and strengthen public resilience to FN.
Conclusion
FN and disinformation are rapidly expanding research fields in psychology and social sciences. This expansion shows the growing importance that the society and the scientific community give to these issues. Polarization has been pictured as a closely related issue. Thus, research exploring the relationship between these themes is increasing. Despite this, definitional and methodological issues are extremely present in these fields. The concentration of research mainly in the United States, the confusion in definitions and measures of polarization, the avoidance of specific polarization types, the use of correlational and retrospective approaches, or the nonsystematization of stimuli employed in the experimental designs pose important challenges in current research. A more systematic approach in order to disentangle basic principles behind the FN and polarization phenomena is highly needed, and we encourage researchers to create and include them in future studies.
Author contribution
Rocío Lana-Blond: conceptualization, data curation, formal analysis, investigation, methodology, and project administration. Miguel García-Saiz: conceptualization, data curation, formal analysis, investigation, and methodology.
Funding statement
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interests
The authors declare none.
Open access
