False and misleading information has become one of the defining risks of the digital age, and a new interdisciplinary review argues that research on the problem during societal crises is rich but fragmented, heavily skewed toward COVID-19, and too focused on individuals to answer the biggest remaining questions. The review, published in the journal Frontiers in Communication and authored by Sofia Johansson, Peter Rodin, Amanda Srugies and Bengt Johansson, draws on 179 peer-reviewed studies published between 2014 and 2024. It maps the current state of knowledge on why false claims spread on social media during crises and what effects exposure to them has on people and society. Using the umbrella term “false and misleading information,” the authors cover concepts that are often used interchangeably in the literature, including misinformation, disinformation, fake news, conspiracy theories, rumours, alternative facts and post-truth claims. The central distinction highlighted is between misinformation, which may be shared without harmful intent, and disinformation, which is deliberately created and spread to mislead. The study was funded by the Swedish Contingencies Agency and set out to answer two questions: what factors contribute to the spread of false and misleading information on social media during crises, and what effects are associated with exposure to such information. The answers, the reviewers found, are disproportionately drawn from the COVID-19 pandemic, from cross-sectional survey designs, and from individual-level analyses, leaving major gaps around causality, platform design, societal structures, emotions, and emerging technologies such as AI and deepfakes.
The methodology behind the review was designed to capture a broad interdisciplinary picture rather than to compile every existing study. The researchers used a scoping review approach, which is suited to identifying patterns and gaps across a large and heterogeneous body of work. They developed a search string covering three thematic blocks: terms for false and misleading information, terms for crises and crisis types, and terms for social media platforms. An additional exclusion block filtered out studies focused on management, automatic detection and other topics outside the review’s scope. Searches were run in Scopus, one of the largest multidisciplinary bibliographic databases, and a citation-based sampling strategy was used. Studies were divided into two time periods, 2014–2019 and 2020–April 2024, to prevent older publications from having an unfair advantage in citation rankings. Within each period, studies were ranked by citations within 27 research fields, and the top 15 percent of the most cited publications in each field were selected for analysis. After screening, the final sample included 179 unique texts and 275 coded units, because some articles were classified into more than one research field. The results show that the literature spans 21 fields, but three dominate: social sciences account for 26.5 percent of the publications, data science for 18.2 percent, and medicine for 17.1 percent. Geographically, about one-third of researchers were based at institutions in the United States, followed by the United Kingdom and China. In terms of crisis types, health crises dominate overwhelmingly, accounting for 86 percent of studies, with the COVID-19 pandemic alone representing the vast majority of those. Natural disasters, terrorism, political crises, war and organizational crises with societal implications are all far less studied. The methodological picture is also strikingly narrow: surveys and quantitative content analysis together account for nearly half of all empirical studies, while experiments and panel surveys, the designs best suited to establishing cause and effect, are rare.
Turning to the factors that contribute to the spread of false and misleading information, the review finds that individual-level explanations are by far the most common, representing 69 percent of all coded factors. Among those, perceptions are the most studied category, making up 25.2 percent of the total. These studies examine how political attitudes, trust in science, trust in institutions, and ideological beliefs shape people’s willingness to share false content. Some research also discusses the possibility of a backfire effect, in which fact-checking strengthens belief in misinformation among people whose attitudes already align with it. Cognitive factors, including information overload and reflective thinking, account for 11.8 percent of the coded factors, while emotions such as anxiety and fear contribute another 10.1 percent. Behaviours, including media consumption habits and patterns of social media use, account for 8.4 percent, and personality traits such as altruism represent 5.9 percent. Sociodemographic factors are a relatively small share, at 4.2 percent. Content-related factors account for 19.3 percent of the total, with research focusing on framing, tone, the presence of guidance or efficacy information, and source characteristics. Platform-specific factors are less common, at 7.6 percent, and mostly concern platform affordances such as the ease of sharing across platforms and positive feedback mechanisms like likes. Societal-level structural factors, such as the political or economic context, are almost absent, representing only 3.4 percent of the factors identified in the literature. The authors stress that because most of this work relies on cross-sectional survey data, it can describe associations but cannot establish whether these factors actually cause people to share false and misleading information.
The review also maps the consequences of exposure to and belief in false and misleading information during crises. Here the dominance of individual-level outcomes is even stronger, with 87.4 percent of all coded effects concerning individuals and only 12.6 percent concerning society at large. No organizational-level effects were identified in the coded material. Behavioural effects are the most frequently studied, accounting for 39 percent of all effects. In the context of COVID-19, this research largely examines whether exposure to false claims reduces adherence to public health restrictions, increases vaccine hesitancy, and undermines preventive behaviours. Perceptual effects are the second largest category, at 27.4 percent, and include phenomena such as politically motivated reasoning, threat perception, and the third-person effect, whereby people believe that others are more susceptible to misinformation than they are themselves. Emotional effects, such as increased anxiety, worry and stress, account for 11.6 percent of the coded effects. Cognitive effects, including the impact of misinformation on knowledge and the role of media literacy, represent 9.5 percent. Societal effects make up the smallest share, at 12.6 percent, and are dominated by economic consequences, with some attention to political consequences. The authors note that the exclusion of studies focused on management may have removed some research that also examined effects, but they still argue that the overwhelming focus on individuals limits understanding of how misinformation damages institutions, public trust, democratic processes and economic systems during crises.
Beyond the dominance of COVID-19 and individual-level analyses, the review identifies several critical gaps that should shape future research. One of the most important is the tendency to treat social media as a single, homogeneous phenomenon. In 64.2 percent of the studies, authors referred to social media in general without specifying which platforms were being analysed. When specific platforms were examined, X/Twitter was by far the most studied, accounting for 46.8 percent of platform-specific studies, followed by Facebook at 17.7 percent and Sina Weibo at 10 percent. Newer platforms such as TikTok and Snapchat are almost entirely absent, and only five studies compared two or more platforms. This matters, the authors argue, because different platforms have different algorithms, affordances, norms and user bases, and the way false information spreads on one platform may not resemble its spread on another. A second major gap is methodological. The reliance on cross-sectional surveys means that little is known about causal relationships. Experiments and panel studies, both of which can trace change over time and control for alternative explanations, are severely underrepresented. The review also calls for greater attention to emotions. Crises are emotionally charged events, and anxiety, anger and fear likely interact with cognitive biases and personality traits to shape sharing and belief. Yet emotions remain a relatively small part of the research landscape. Structural factors are even more neglected. The authors argue that individual-level explanations cannot fully account for why misinformation spreads more rapidly in some societies than others, and they urge researchers to investigate how media systems, political contexts, economic inequality, and platform governance interact with individual behaviour. Finally, the review points to a striking absence of research on technological developments, including AI-generated content and deepfakes, despite the growing availability of these tools and their potential to amplify false narratives during crises.
Taken together, the findings point to the need for a more interdisciplinary research agenda that treats misinformation during crises as a complex problem at the intersection of communication, psychology, technology and crisis management. The authors recommend moving beyond the COVID-19 pandemic to test whether current findings apply to other types of crises, including natural disasters, terrorism, war and political upheaval. They also call for more varied methods, especially panel studies and experiments, to establish causal mechanisms and to examine how crisis-induced emotional states and motivated reasoning interact. Future research should expand beyond individual-level analyses to include the social and structural forces that shape the information environment, and it should engage directly with the algorithms and platform architectures through which false and misleading information flows. The review acknowledges its own limitations. The inclusion of the term COVID-19 in the search string likely increased the representation of health crises, and the exclusion block may have unintentionally removed some relevant studies. The citation-based sampling strategy, which selected the most cited publications from each field, may also have excluded newer or less-cited but otherwise valuable research. Nevertheless, the authors argue that the review offers the most comprehensive interdisciplinary mapping to date and provides a foundation for developing theoretical frameworks that integrate context-specific crisis dynamics with broader social, technological and psychological mechanisms. Without such a shift, they warn, societies will continue to confront the spread and consequences of false and misleading information during crises without a clear evidence base for understanding, predicting, or mitigating it.