Mapping the misinformation landscape: an interdisciplinary review of false and misleading information on social media during crises

Introduction

Broadscale societal crises, or so-called mega-crises, are often followed by an increased spread of false and misleading information, which can have serious consequences for individuals and society at large (Loomba, 2021; Strömbäck et al., 2022). Social media platforms have further exacerbated this challenge by enabling widespread circulation of both accurate and inaccurate information (Brennen et al., 2020; Papakyriakopoulos et al., 2020). The COVID-19 pandemic, for instance, saw a large spread of misleading and false information across social media platforms concerning the deadliness of the virus as well as its causes and origins (Balakrishnan et al., 2022; Önnerfors, 2021). The COVID-19 pandemic is, however, not the only example. Other crises frequently attract attention due to the circulation of false and misleading information concerning them, such as the terrorist attacks and wars (Chen et al., 2022; Hameleers et al., 2024). It is therefore not surprising that research interest in the phenomenon, both in general and in relation to crises more specifically, has accelerated over the past decade (Broda and Strömbäck, 2024). Studies on false and misleading information can be found across fields, including social sciences, psychology, medicine, and computer science (see, for example, Agley and Xiao, 2021; Lewandowsky et al., 2012; Van der Linden, 2020; Rubin et al., 2015).

As noted by prior scholars, the interdisciplinary nature of research has both enhanced our understanding of the phenomenon as well as contributed to a fragmented field. Scholars rarely reference work outside their own discipline. This applies both to studies on false and misleading information in general, where different disciplines have developed largely independently (for a discussion about this, see Broda and Strömbäck, 2024), and to research specifically focused on societal crises. Journalism and communication scholars have examined the role of false news in shaping political agendas (e.g., Van der Linden, 2020), while psychological researchers have focused on cognitive processing and memory in relation to misinformation (e.g., Lewandowsky et al., 2012) and computer scientists have worked on algorithmic solutions to detect and mitigate false and misleading information (e.g., Chen et al., 2015). Moreover, studies on false and misleading information rarely incorporate theories and concepts from crisis research (such as crisis communication or crisis management). Similarly, crisis communication research tends to engage only to a limited extent with theories and concepts from other disciplines (Johansson and Vigsø, 2020; Palenchar, 2009).

Beyond the lack of theoretical integration across fields, conceptual ambiguity further complicates the picture. As highlighted by other scholars, different terms are sometimes used to refer to the same phenomenon, or the same term is used to refer to different phenomena (for discussion about this, see Broda and Strömbäck, 2024; Kapantai et al., 2021 Wardle, 2018);. For instance, disinformation, misinformation, conspiracy theories, rumours, and fake news are commonly used, but terms like incorrect information, misleading information, and alternative facts also appear in the literature (Broda and Strömbäck, 2024).

The disciplinary fragmentation and conceptual ambiguity hinder the development of a comprehensive understanding of the phenomenon. As argued by Broda and Strömbäck (2024), a more integrated and interdisciplinary approach to the study of false and misleading information is needed, starting with literature reviews. This study therefore aims to provide an overview of the current state of knowledge on false and misleading information during crises across various research fields. Specifically, a scoping review was applied to identify key concepts, methods, relevant factors, and knowledge gaps (Munn et al., 2018). The main focus, however, centred on mapping (1) the factors contributing to the spread of false and misleading information on social media during crises, and (2) the potential consequences of such information at both individual and societal level. We utilize the umbrella term false and misleading information, referring to research on related terms frequently applied in research, such as disinformation, misinformation, false/misleading information, conspiracy theories, rumours, alternative facts, fake news, among others. In total, 179 research studies published between 2014 and 2024, were analysed using quantitative content analysis. By compiling studies from multiple research fields, the overview aims to provide a more holistic view of current state of research, paving the way for interdisciplinary research approaches.

Defining false and misleading information

As noted above, several terms are applied in research to refer to false or misleading information. However, misinformation and disinformation are foundational concepts to describe false and misleading information, occurring across a wide range of different disciplines. While misinformation refers to false information that may be shared without harmful intent, disinformation is spread with the intent to mislead or cause harm (Kapantai et al., 2021). The main difference thus lies in the intent behind it. This also means that information may start as disinformation but spread as misinformation.

In the last decade, the term fake news has also gained prominence, especially in political and media studies (Aïmeur et al., 2023). As noted by Broda and Strömbäck (2024), “fake news both refers to a weaponized label used to attack and undermine the legitimacy of the news media” as well as false or misleading content “presented as legitimate news” (p. 142). Regarding the latter, researchers have developed typologies to categorize fake news. Tandoc et al. (2018) categorize fake news based on its degree of factual content and the creator's intent to mislead. News satire might blend fact and fiction but is understood by its audience as a form of entertainment, whereas fabricated news is entirely fictional and spread with the aim to deceive (Allcott and Gentzkow, 2017). Due to its emphasis in intentional spread to deceive or harm, scholars often use the term interchangeably with disinformation. Further related to the notion of fake news are terms such as alternative facts and post-truth, referring to alternative constructions of realities, evidence and expertise that goes against established evidence and mainstream discourse (Marwick and Partin, 2024).

Relating to crises more specifically, the term infodemic gained renewed attention during the COVID-19 pandemic. It was coined during the SARS outbreak by David Rothkopf in 2003 and refers to the overwhelming amount of information—both accurate and inaccurate—flooding the public during a crisis (Buchanan, 2020). This overflow of information makes it challenging for individuals to identify trustworthy sources and information, potentially leading to confusion about what is true or false. The notion infodemic is related to the term information pollution, referring to “situations where the information supply is contaminated with low-value (e.g., irrelevant, redundant, inaccurate) material” (Malin and Lubienski, 2022, p. 3).

Finally, the term conspiracy theories also gained increased traction in research following the COVID-19 pandemic. This is mainly due to the upsurge in the amount of conspiracy theories that spread online concerning the deadliness and the origins of the virus (Önnerfors, 2021). Conspiracy theories are defined as attempts to explain significant events as the result of secretive, malevolent actions by powerful groups (Douglas, 2021). While conspiracies can be true, conspiracy theories on the other hand are often not grounded in evidence or goes against established evidence. They also tend to carry mis- and disinformation and are, as a consequence, used interchangeably with these concepts. Finally, rumours have also received significant attention in research departing from crises, defined as unverified information that spreads in situations of ambiguity or perceived threat (Bordia and DiFonzo, 2007).

To sum up, terms like misinformation, disinformation, fake news, alternative facts, post-truth, conspiracy theories, infodemic, and rumours are all applied in research addressing false and misleading information. Although they are sometimes used to highlight different facets of the phenomenon, they are also, as noticed by prior scholars, often used interchangeably (for a discussion, see, for example, Broda and Strömbäck, 2024; Kapantai et al., 2021 Wardle, 2018);. To get a holistic picture of the current state of knowledge, this research overview applies the umbrella term false and misleading information to refer to all the above terms.

False and misleading information: a scattered research area

Although scholars point to the increased spread of false and misleading information during societal crises, few have conducted comprehensive reviews across fields on its spread and effects during such events. While there are examples of literature reviews focusing on topics such as misinformation or fake news (Tandoc, 2019; Broda and Strömbäck, 2024), disinformation (Benkler et al., 2018; Marwick and Lewis, 2017), and conspiracy theories (Douglas et al., 2019), few of them focus specifically on crisis, with most being focused on the COVID-19 pandemic (see, for instance, Pian et al., 2021; Rocha et al., 2023). However, crises are characterized by uncertainty and high emotional engagement (Canet-Juric et al., 2020; Sellnow and Seeger, 2013), which is likely to increase the spread of false and misleading information online. For instance, studies show a large spread of false and misleading information concerning government policies as well as the causes and origins of the threat during crises (King and Wang, 2023; Lu, 2020; Önnerfors, 2021). Yet, studies on false and misleading information rarely incorporate theories and concepts from crisis research. Similarly, crisis communication research rarely engages with theories and concepts from other disciplines (Johansson and Vigsø, 2020; Palenchar, 2009). The interdisciplinary nature of the phenomenon underlines the need to integrate concepts from various fields, however.

In addition to the above, studies addressing false and misleading information within different disciplines have developed largely independently from each other (for a discussion about this, see Broda and Strömbäck, 2024). Journalism and communication scholars have examined the role of false news in shaping political agendas and public opinion (e.g., Van der Linden, 2020). These studies highlight the ways in which false and misleading information can increase polarization, erode trust in traditional news media, and influence public opinion. Meanwhile, psychological researchers have focused on cognitive processing and memory in relation to misinformation (e.g., Lewandowsky et al., 2012). At the same time, computer scientists have developed machine learning models, automated fact-checking systems, and network-based approaches to identify false and misleading information (e.g., Y. Chen et al., 2015). Yet, the specific conditions associated with crises require an interdisciplinary approach to the study of false and misleading information, combining knowledge from different fields.

Against this backdrop, a mapping of existing research is needed to synthesize the fragmented findings form prior research. Mapping factors related to the spread and effects of false and misleading information in crises can help reveal gaps in current knowledge and inform future research on the development of interdisciplinary theoretical frameworks. This overview therefore aims to answer two overarching research questions:

RQ1. Which factors have previous research identified as contributing to the spread of false and misleading information on social media during crises?

RQ2. What are the effects identified in previous research related to exposure to false and misleading information on social media during crises?

Methodology

To answer the RQs, a scoping review was conducted. This strategy was chosen because it aligns with the research aims, but also due to the vast number of studies conducted on the topic. Unlike a systematic review, which aims to compile all research on a given subject, a scoping review seeks to gather representative studies to identify patterns and gaps in the existing research concerning theories, methods, or results. Specifically, it aims to identify key terms and concepts within current research, analyse the methods applied, as well as identify factors and current knowledge gaps (Munn et al., 2018).

Timeframe

The studies included in this review were published between 2014 and April 17, 2024. 2014 was chosen as the starting point due to the increased attention given to false and misleading information following key events like the 2016 U.S. presidential election. Close to this period, between 2010 and 2012, there was a widespread adoption of smartphones in Sweden, leading to increased social media use (Andersson, 2023). 2014 and onwards thus marks a period of increased social media usage and increased attention given to the topic of false and misleading information. The results further supported the rationale for choosing 2014 as the starting year, as the number of relevant studies increased significantly after 2016, peaking during the COVID-19 pandemic years and continuing to rise thereafter.

Search strategy and study selection

The initial step in collecting the empirical material involved designing the search thread. To do so, we utilized field-specific databases. Specifically, PsycInfo, ERIC, Cinahl, PubMed, Proquest Social Sciences, and IEEE were used to identify common terms for false and misleading information, crises, and social media across various research domains. Identified terms and concepts were structured into a search string comprising three thematic blocks, along with a fourth block aimed at excluding irrelevant studies. The search string, including the three thematic blocks plus the exclusion block, is presented and discussed below.

The first block targeted terminology used to describe false or misleading information, and the second block focused on terminology for societal crises and specific crisis types. It should be noted that crises not explicitly labelled as such were excluded except for COVID-19. For instance, other epidemics like Zika were excluded unless explicitly referred to as a pandemic or crisis. However, this decision was made after testing the search string and noticing that inclusion of the term “COVID-19” added substantially to the total amount of studies, while the equivalent term such as “Zika” did not. It is therefore likely that including COVID-19 did not significantly alter the proportion of crises cases within health crises¹. That said, other crises (beyond health crises) may have been excluded if not explicitly stated as one of the crisis types included in the search string, potentially creating an overrepresentation of health crises. Finally, the third block covered social media and associated platforms.

An exclusion block was added to exclude research outside the scope of the study, focusing on, for instance, organizational management of false or misleading information or automatic detection thereof. The exclusion block therefore included terms such as management, automatic detection, and others. It is important to note that it is possible that the exclusion block filtered out relevant studies. For example, studies examining management or leadership strategies as well as the effects of, or spread of, false or misleading information were excluded. The full search string can be found below².

In addition to the predefined search terms and exclusion block, studies were considered eligible if they were written in English and published as peer-reviewed journal articles, book chapters, or conference proceedings were considered eligible for inclusion. A range of publication types were included to ensure coverage of different research fields. However, the language requirement means that relevant studies in other languages were excluded from the material. Finally, studies had to include the predefined search terms in the title, abstract, or keywords and be published between 2014 (2024) and April 17 (2024) to meet the eligibility criteria (as discussed in more detail above under time frame).

Data collection

Eligible studies were identified using Scopus, one of the largest multidisciplinary bibliographic databases. Thus, the field-specific databases (PsycInfo, ERIC, CINAHL, PubMed, ProQuest Social Sciences, and IEEE) were used solely to develop the search strategy and were not used for the final literature search. Scopus was chosen for data collection as it is one of the largest academic databases covering a wide range of research fields. This enabled the identification of studies from diverse disciplines relevant to the research topic.

In line with the aim of a scoping review and to map the characteristics of the research field rather than include all available studies (Munn et al., 2018), a citation-based sampling strategy was adopted. Specifically, citation count was used as an indicator of scientific influence within the respective research field. To ensure that older publications did not have an advantage in the citation ranking, two separate literature searches were conducted from the outset, covering the publication periods 2014–2019 and 2020–April 17, 2024, respectively. Together, the two searches identified 1,189 eligible studies across 27 research fields. Within each search, studies were ranked according to citation count within each research field, and the top 15% most cited publications from each field were selected for further analysis. While citation count was used as an indicator of the scientific influence of a publication within its research field, it should not be interpreted as a measure of study quality. Consequently, relevant studies examining factors associated with the spread or effects of false and misleading information with relatively few citations may have been excluded.

Following the initial screening, 30 articles were excluded after closer review because they did not meet the eligibility criteria. Specifically, this included studies that mentioned the predefined search terms in the abstract but did not operationalize the corresponding concepts in the analysis. Consequently, the final sample consists of 179 unique texts and 275 coded units. A flowchart of the data colleciton process can be found in Figure 1. The number of coded units is higher as some articles were categorized into more than one research field by Scopus based on the journals, books, and conferences associated with a particular scientific text.

Coding and analytical procedures

The collected material was coded using a pre-defined codebook developed by the authors, which was pretested and refined through several iterations. The codebook consists of three sections: 1) metadata (authors, affiliations, journal details, keywords, and research fields), 2) study design (study subjects, crisis type, social media platforms, key concepts, and methodology), and 3) study results (factors contributing to the spread and effects of false information). The coding process was divided among the researchers, with regular meetings to ensure consistency and refine the process. The coded material was analysed applying quantitative content analysis, investigating the frequency of themes within each subsection of the codebook.

Results

Before addressing our main research questions, an overview of the fields and countries represented in research on false and misleading information, types of crises and social media typically studied, as well as the methods employed in the research are provided. After that, the factors contributing to the spread and the effects of false and misleading information are addressed.

Research across disciplines and countries

Publications span 21 distinct fields of research. Table 1 highlights the ten largest disciplines (accounting for 90 percent of the publications), with three fields producing a particularly high share of publications: social sciences, computer science, and medicine.

Table 1. Research on false and misleading information categorized by research field.

The research fields that make up less than two percent of the sample have been included under the category “Other”. The table therefore shows which 10 research fields that research on false and misleading information are most frequently published in.

The results further show that one-third of researchers are based at universities and organizations in the United States, followed by the United Kingdom and China. While up to ten authors were coded for each study, Table 2 presents data on the first two authors to avoid overrepresenting individual countries.

Table 2. Research on false and misleading information categorized by the countries where the authors’ universities/institutions are located.

Crisis types

Most studies focus on global health crises, particularly the COVID-19 pandemic (86 percent) (Table 3). The second-largest crisis type is natural disasters, such as earthquakes (e.g., the 2018 Palu earthquake in Indonesia) and hurricanes (e.g., Maria in Puerto Rico and Harvey in the United States, both in 2017). Other types of crises include terrorism (2.2 percent), such as the Boston Marathon bombing in 2013 and the series of terrorist attacks in several UK cities in 2017. Political crises (1.7 percent) and organizational crises with societal implications (1.1 percent) are less commonly studied.

Table 3. Research on false and misleading information distributed categorized by crises types.

Social media platforms

115 out of 179 studies (64.2 percent) address false and misleading information on social media in general, without specifying particular platforms. This approach is especially common in survey-based studies that explore people's perceptions or usage of social media.

The remaining studies focus on specific platforms, with the majority analysing X/Twitter (46.8 percent) or Facebook (17.7 percent), together comprising two-thirds of the platform-specific studies. The third most frequently studied platform is the Chinese social media site Sina Weibo (10 percent) (Table 4). Most research analysing specific social media platforms focuses on one platform at a time. Very few studies (5 in total) compare two or more platforms.

Table 4. Research on false and misleading information distributed categorized by social media platforms.

Methods

Most of the studies are empirical studies, accounting for 76 percent. Other types of studies, such as reviews, meta-studies, and theoretical studies, each comprise approximately 7 percent. The most applied methods are survey studies and quantitative content analysis, which together encompass nearly half of all studies. Survey research is the predominant method across almost all research fields (see Table 5). Only three studies utilized data from panel surveys. Finally, there are examples of studies employing sentiment analysis, conceptual analysis, document analysis, and case studies.

Table 5. Research on false and misleading information categorized by research methods.

Factors contributing to dissemination and effects of exposure

To analyse the factors contributing to the dissemination of false and misleading information, as well as its effects, the analysis does not rely on the total number of collected articles (N = 179). Instead, it is based on the frequency with which various dissemination factors or effects at the individual, organizational, or societal levels were coded. Since multiple dissemination factors and effects can be coded for each article, the dataset expands to a total of N = 214. Of this total, factors contributing to dissemination constitute 55.3 percent, while 44.6 percent address the effects (see Tables 6, 7).

Table 6. Research on dissemination categorized by research fields.

The research fields that have only been coded for one variable (i.e., that make up less than 0.8 percent of the total number) have been included under the category “Other”. The table therefore includes fields that have been coded across more than one variable.

Table 7. Research on effects categorized by research field.

The research fields that have only been coded for one variable (i.e., that make up less than 0.8 percent of the total number) have been included under the category “Other”. The table therefore includes fields that have been coded across more than one variable.

Most variables related to dissemination are found within the social sciences (35.3 percent), followed by computer science (19.3 percent) and medicine (13.4 percent).

The largest proportion of investigated factors related to effects is found within the social sciences (27.4 percent), followed by medicine (20 percent) and computer science (16.8 percent).

Factors contributing to dissemination

Factors influencing the dissemination of false and misleading information can be categorized into individual-level factors (e.g., attitudes, personality traits, cognition, emotions), content-related factors (e.g., framing and content markers), technologies and platforms (e.g., AI and platform functionalities), and societal structures (e.g., economic or political contexts). Among these, individual-level factors constitute the most examined variables (69 percent), followed by content-related factors (19.3 percent) and platform-specific factors (7.6 percent) (see Table 8). As noted earlier, most studies in the dataset rely on cross-sectional data, which limits the ability to analyse causal mechanisms.

Table 8. Research on dissemination categorized by factors.

Since no factors related to AI applications are present in the material, the category has been excluded from the table.

At the individual level, perceptions are the most frequently studied factor in understanding people's propensity to spread false and misleading information on social media, accounting for 25.2 percent. Several studies have, for example, analysed the influence of political attitudes and trust (e.g., Agley and Xiao, 2021; Das et al., 2021 Freiling et al., 2023;; Recuero et al., 2022 Stecula and Pickup, 2021;; Su, 2021; van Mulukom et al., 2022). These studies also discussed the potential for a backfire effect, a strengthened belief in false or misleading information following fact-checking, among individuals whose attitudes align with the false or misleading information (Sharevski et al. 2022).

Studies focusing on cognition account for 11.8 percent of the research, with particular attention to phenomena such as information overload (Bermes, 2021). Emotions have also been examined to some extent (10.1 percent). For example, one study explored the interaction between emotions (anxiety) and perceptions (ideology) (Freiling et al., 2023).

Relatively few studies have investigated the influence of personality traits (5.9 percent) and behaviour (8.4 percent). In relation to personality traits, altruism has for instance been studied as a driver behind information spreading (e.g., Apuke and Omar, 2021). Regarding behaviour, some studies have analysed the impact of media consumption habits, focusing particularly on the influence of social and political alternative media (e.g., Romer and Jamieson, 2021). Additionally, some studies have explored the interaction between social media use and cognition (e.g., Stecula and Pickup, 2021). Finally, the interaction between social media use and emotions has also been examined (e.g., Su, 2021; e.g. Freiling et al., 2023),.

Moving beyond the individual level, framing and packaging of information (framing and cues) account for 19.3 percent of the total factors investigated. Studies have analysed elements such as tone (sentiment), efficacy information (guidance on how to respond to the crisis), and characteristics of the information source (e.g., King and Wang, 2023; Song et al., 2022; Jung et al., 2020).

Platform-specific features or technologies account for 7.6 percent of the total dissemination factors in the material. These studies primarily focus on platform affordances, such as the ability to share material across platforms and the impact of positive feedback mechanisms, such as likes (e.g., Agarwala et al., 2023; Ginossar et al., 2022). Finally, structural factors at the societal level constitute 3.4 percent of the total dissemination factors and have primarily focused on the political context (N = 2).

Effects of exposure and dissemination

Like factors contributing to dissemination, the effects of false and misleading information can be categorized into impacts on individuals, organizations, and society at large. Effects on individuals constituted the largest portion of the collected material, accounting for 87.4 percent, followed by societal-level effects at 12.6 percent (see Table 9). No organizational effects were identified in the dataset. As with studies on dissemination, most of the research rely on cross-sectional data.

Table 9. Research on effects categorized by factors.

Since no organizational effects are present in the material, the category has been excluded from the table.

At the individual level, behavioural effects were the most studied, accounting for 39 percent of the research. These studies primarily addressed behaviour in the context of COVID-19, examining the relationship between exposure to, or belief in, false and misleading information on an individuals’ likelihood of adhering to restrictions and willingness to vaccinate (e.g., Jennings et al., 2021; Lee et al., 2020; van Mulukom et al., 2022; Romer and Jamieson, 2020; Sallam et al., 2021; Zimmerman et al., 2023). Some studies have also discussed the potential effects of coordinated disinformation campaigns (such as “The Plandemic”) (e.g., Nazar and Pieters, 2021).

Perception-related effects (27.4 percent) were the second most common, focusing on factors such as politically motivated reasoning (i.e., the tendency to believe information that aligns with already established attitudes and beliefs), threat perceptions, and the third-person effect (i.e., the tendency to believe that others are more susceptible to false or misleading information than oneself) (e.g., Das et al., 2021; Liu and Huang, 2020; van Mulukom et al., 2022; Yang and Tian, 2021).

Few studies have explored the effects of false and misleading information on emotions or affective states (11.6 percent), focusing mainly on the influence on anxiety and stress levels (e.g., Khan, 2021; Lee et al., 2020; Rocha et al., 2023). The effects on cognition (such as knowledge) have also been explored to a lesser extent (9.5 percent), but examples include studies examining the role of self-reported media literacy in relation to belief in false and misleading information (e.g., Borah et al., 2022).

Finally, a relatively small portion examines effects on society at large (12.6 percent). However, it is important to emphasize that the exclusion keyword management in the search string may have contributed to the exclusion of studies investigating both the effects and management of false and misleading information. The most studies societal effect is economic effects (6.3 percent) (Bernard et al., 2021).

Conclusions

This interdisciplinary review is based on an analysis of 179 scientific studies on false and misleading information on social media during societal crises from 2014 to 2024. The focus has been on explaining the spread of such information and identifying its effects. Several key conclusions can be drawn from the review.

Starting with broader general patterns, the results first show that COVID-19 has dominated research on false and misleading information over the past decade. The pandemic was not only a global crisis with significant impacts but also the most frequently studied topic in this field. A critical question is how much of the knowledge gained from research during the pandemic can be generalized to other types of crises. Although most studies originate from the social sciences, a substantial portion comes from medical research. This proportion would likely have been smaller if the pandemic had not dominated the sample. It is therefore crucial for future research to extend beyond the pandemic to ensure that findings can be applied to other types of crises.

Second, research tends to view social media as a homogeneous phenomenon. Notably, 64 percent of the studies examined social media without specifying which platforms were being analysed. When individual platforms were the focus, X/Twitter was the most studied. However, social media is not a monolithic entity; referring to “social media” without specifying platforms is as imprecise as referring to “media” in general. Understanding what respondents mean when referring to “social media” can be challenging. Moreover, there is a lack of research on newer platforms like TikTok and Snapchat.

Finally, cross-sectional survey studies are most common for studying false and misleading information on social media. Given the importance of determining the spread and effects of false and misleading information, the lack of studies establishing causality is striking. Experiments and panel studies, both of which can confirm cause-and-effect relationships, are underrepresented. Future research should focus more on identifying causal mechanisms behind the spread and effects of false and misleading information.

Addressing the research questions, the results show that prior research has mainly focused on individual-centred explanations and effects. Approximately two-thirds of the studies attribute the spread of misinformation to individual factors. The dominance of the individual perspective is even greater in studies examining effects, with over 80 percent focusing on individuals. These studies often explain the spread of false and misleading information based on people's perceptions. However, crises are affective-laden events where emotions play a significant role in human behaviour. It is likely that emotions such as anger or anxiety interact with cognitive factors or personality traits, influencing the likelihood of people spreading false or misleading information. Likewise, exposure to and belief in false information may trigger strong emotions that, in turn, affect behaviour in various ways. Future research should pay greater attention to the role of emotions in both the spread and effects of misinformation.

Moreover, structural factors influencing the spread and effects of false and misleading information are absent. This limits our understanding of how structural factors interact with individual behaviours. One reason for this may be that studying individuals is methodologically easier than analysing societal-level influences. However, research on false and misleading information must expand beyond individual-level explanations to consider broader societal dynamics. This includes examining the role of media systems, political contexts, and platform algorithms, as well as how these structures interact with individual behaviours and perceptions. Future research should adopt a more holistic approach to fully understand the phenomenon.

Lastly, technological aspects and applications are rarely explored. It is striking how few studies focus on examining AI applications, and research on deepfakes is entirely absent. Given the rapid advancements in these technologies and their potential impact, this gap is particularly noteworthy. One likely reason is that AI-related false and misleading information has only recently become a central societal concern, and academic research takes time to develop. Our selection criteria may have further contributed to this, including studies with the highest number of citations. Nevertheless, future studies on misinformation during societal crises must address the role of AI and emerging technologies in shaping the spread and impact of false and misleading information.

Taken together, what is lacking is an interdisciplinary research agenda that accounts for the context-specific dynamics of crises. The spread of false and misleading information during crises represents a complex problem at the intersection of psychology, technology, and communication. Moving forward, research should adopt interdisciplinary approaches that examine how key drivers identified in this review, such as crisis-induced emotional states and motivated reasoning, interact. At the same time, future studies should move beyond only focusing on individual-level factors and also incorporate structural and technological dimensions, which remain underexplored. Doing so would support the development of frameworks that integrate context-specific dynamics with broader theoretical mechanisms to explain and mitigate false and misleading information during crises.

Of course, this study is not without its limitations. Most notable is the inclusion of the term COVID-19 in the search string, which may have generated an overrepresentation of that case in relation to other crises types. Moreover, the exclusion block might have (unintentionally) led to an exclusion of relevant studies. However, the strategy was deemed necessary given the large number of studies published on the topic. Future reviews should therefore aim to replicate the findings while also further refining the search string and method, for instance by using computational and automatised methods.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Marta Natalia Lukacovic, Angelo State University, United States

Reviewed by: Ahlam Alharbi, Imam Abdulrahman Bin Faisal University, Saudi Arabia

Sergey Samoylenko, George Mason University, United States

Footnotes

¹Adding Zika to the search string added 12 studies, while adding Covid-19 and pandemic added 2102 studies.

²TITLE-ABS-KEY-AUTH(misinformati* OR disinformati* OR rumor* OR rumour* OR "false informati" OR {information pollution} OR "alternative fact" OR {fake news} OR "multiple truth" OR post-truth OR mislead* OR misrepresent* OR conpirac) AND TITLE-ABS-KEY(crisis OR crises OR disaster OR emergency OR emergencies OR catastroph* OR calamity OR calamities OR devastation* OR "volcano eruption" OR tornado OR "Covid-19" OR "pandemic" OR hurricane OR storm OR "terrorist attack" OR shooting* OR earthquake* OR landslide* OR flood) AND TITLE-ABS-KEY("social media" OR "digital media" OR "online media" OR instagram OR twitter OR facebook OR telegram OR whatsapp OR tiktok OR wechat OR douyin OR "facebook messenger" OR kouaishou OR "Sina Weibo" OR qq OR pinterest OR ict OR "computer mediated communication" OR internet OR "electronic communication") AND NOT TITLE-ABS-KEY(classification* OR benchmark* OR classifier* OR "rumor identification" OR "rumour identification" OR "misinformation identification" OR "fake news detection" OR "misinformation detection" OR "rumour detection" OR "rumor detection" OR automati* OR {management} OR computational*) AND (LIMIT-TO (DOCTYPE,"ar") OR LIMIT-TO (DOCTYPE,"cp") OR LIMIT-TO (DOCTYPE,"re") OR LIMIT-TO (DOCTYPE,"ch")) AND (LIMIT-TO (LANGUAGE,"English"))