Recent research has transformed the public understanding of misinformation from a problem of “fake news” into a systemic feature of digital life with measurable consequences for democracy, public health, and social cohesion. In one of the most cited studies on the topic, Vosoughi, Roy, and Aral (2018) found that false news on Twitter reached more people and spread faster than the truth, particularly when the content was novel and emotionally charged. More recently, Allen, Watts, and Rand (2024) quantified the impact of vaccine-skeptical and misinformation content on Facebook, showing that although such content represented a small fraction of overall engagement, its presence was substantial enough to influence attitudes and behaviors on a population scale. Ecker and colleagues (2024) argued in Nature that misinformation poses a bigger threat to democracy than many appreciate, not only because it deceives individuals but because it erodes trust in institutions, expertise, and even the possibility of shared facts. Surveys of experts in the field, such as Altay and colleagues (2023), reveal broad agreement that misinformation is not simply false content but a complex phenomenon involving deceptive framing, emotional manipulation, and partisan identity. The stakes are high: from vaccine hesitancy and climate doomism to election integrity and public health compliance, the information environment has become a battleground where psychological vulnerabilities are systematically exploited. Van der Linden and Kyrychenko (2024) have argued that a broader view of misinformation—one that includes not just verifiably false claims but also misleading narratives, hyperpartisan outrage, and manipulative tactics—reveals more opportunities for intervention, but also demands more sophisticated scientific and policy responses.
Understanding why misinformation spreads requires looking beneath the content to the psychological mechanisms that make it compelling. One of the oldest and most robust findings is the continued influence effect: Lewandowsky and colleagues (2012) demonstrated that people continue to rely on misinformation even after it has been corrected, especially when the correction fails to offer a coherent alternative explanation. Emotion is a powerful amplifier. Martel, Pennycook, and Rand (2020) found that reliance on emotion promotes belief in fake news, while Carrasco-Farré (2022) showed that deceptive content demands less cognitive effort and appeals more directly to emotions than reliable sources. Outrage, in particular, appears to be a key driver. McLoughlin and colleagues (2024) found that misinformation exploits outrage to spread online, and Watson and colleagues (2024) showed that negative news articles are shared more frequently to social media. The language of conspiracy is also distinctive: Fong and colleagues (2021) found that conspiracy theorists use more uncertainty-related, anxiety-related, and collective-identity language, creating a sense of insider knowledge and persecution. Marks and colleagues (2025) added a crucial social dimension by showing that ingroup solidarity drives social media engagement after political crises, meaning that sharing misinformation can be less about truth-seeking and more about signaling loyalty to one’s group. Fallacious reasoning also plays a role: Blassnig and colleagues (2019) documented how right-wing populist rhetoric relies on informal fallacies, while the classic works of Hamlin (1970) and Walton (1987) remind us that logical fallacies have long been tools of persuasion. Kauk and colleagues (2025) conducted a large-scale analysis of online social data and found that misinformation and reliable information follow different emotional and content dynamics over time, with misinformation maintaining a more negative and arousing profile. Stuart (2025) has highlighted the particular danger of climate doomism, a form of misinformation that induces helplessness and disengagement, showing that misinformation is not always about false facts but also about destructive emotional narratives.
The most promising response to misinformation has come from psychological inoculation, or “prebunking,” an approach rooted in McGuire’s (1964) classic inoculation theory. The idea is simple but powerful: just as vaccines expose people to a weakened virus to build immunity, exposing people to weakened examples of manipulation techniques can build cognitive resistance. Compton and colleagues (2021) reviewed the extensive evidence for inoculation theory in the post-truth era, and Roozenbeek and van der Linden (2024) have developed and tested scalable interventions, including online games that teach people to recognize manipulation tactics such as emotional language, incoherence, and false dichotomies. In a landmark study, Roozenbeek and colleagues (2022) showed in Science Advances that psychological inoculation improves resilience to misinformation on social media across countries and cultures. Meta-analyses have confirmed the effectiveness of these interventions: Lu and colleagues (2023) found that inoculation improves credibility assessment and sharing discernment, while Simchon and colleagues (2025) used signal detection theory to show that inoculation increases sensitivity to misinformation without simply making people more skeptical of everything. Huang and colleagues (2024) found that media literacy interventions also improve resilience to misinformation, though the effects vary by design and context. Crucially, the benefits of inoculation are not necessarily permanent. Maertens and colleagues (2025) demonstrated in Nature Communications that psychological booster shots—brief reminders of the manipulation techniques—can maintain long-term resistance, much like booster vaccines. Pennycook and colleagues (2024) found that inoculation and accuracy prompting, a technique that nudges people to think about whether information is true, work best in combination, though not always alone. Large-scale implementations have already begun: Jigsaw (2023) launched prebunking campaigns in Germany, and the BBC (2025) has explored the idea of vaccinating audiences against mis- and disinformation. The American Psychological Association has also weighed in: van der Linden and colleagues (2025) published a consensus statement on using psychological science to understand and fight health misinformation, emphasizing that interventions must be evidence-based and tailored to the social contexts in which misinformation operates.
While psychological interventions target individual minds, the structural environment of social media platforms creates the conditions under which misinformation thrives. The attention economy rewards engagement, and misinformation often generates more engagement than accurate information. McLoughlin and colleagues (2024) showed that outrage is a particularly effective engagement driver, and platforms’ algorithms are optimized to spread content that provokes strong emotional reactions. The financial incentives are substantial. Bellingcat (2024) analyzed Meta’s ad library during the U.S. presidential election and found that both the Trump and Harris campaigns spent heavily on Facebook and Instagram ads, with dramatic differences in messaging and targeting. Dek and colleagues (2025) mapped what they call the “online manipulation economy,” showing that a shadowy ecosystem of advertisers, data brokers, and content farms profits from spreading deceptive content across platforms. Even ordinary advertisers are part of this ecosystem: Bïrch (2025) reports Instagram advertising costs by CPM, and Grigaliūnaitė (2025) provides click-through rate benchmarks across TikTok, Meta, and YouTube, illustrating how performance metrics drive content strategies. Metalla (2025) offers practical guidance on “ad hooks that drive conversions,” revealing that many advertisers deliberately use emotionally charged language to capture attention—a tactic that misinformation purveyors exploit far more aggressively. Allen and colleagues (2024) found that on Facebook, vaccine-skeptical content was often promoted by a small number of highly active pages, suggesting that the manipulation economy is not a diffuse phenomenon but is concentrated among sophisticated actors. The platform’s own ad library, despite its limitations, has become an essential tool for researchers and journalists to monitor political advertising and detect potential manipulation. However, as Bellingcat’s analysis shows, the ad library also reveals gaps in transparency, making it difficult to fully understand how algorithms amplify misinformation. Searles and Feezell (2023) introduced the concept of “scrollability” to describe how the design of digital news feeds—infinite scrolling, autoplay, and continuous content—reduces the cognitive effort required to consume and share information, which disproportionately benefits sensational and misleading content.
Despite the laboratory success of inoculation and media literacy interventions, translating these findings into real-world impact has proven challenging. Roozenbeek and colleagues (2025) conducted two preregistered field studies on social media and found that misinformation interventions had limited effects on actual sharing behavior, highlighting the gap between what works in controlled experiments and what works in the messy, fast-paced environment of social media. This distinction between efficacy and effectiveness is critical. Molero-Calafell and colleagues (2024) explain the difference between intention-to-treat and per-protocol analyses in clinical research, a methodological distinction that applies equally to misinformation studies: an intervention may work perfectly for people who fully engage with it, but if most people skip it, ignore it, or forget it, its population-level impact will be minimal. Roozenbeek, Remshard, and Kyrychenko (2024) explicitly address this issue, arguing that researchers must move “beyond the headlines” to measure whether interventions change behavior in real-world conditions, not just attitudes in surveys. Another challenge is that some people actively reject psychological interventions. Roozenbeek, Young, and Madsen (2025) examine the wilful rejection of psychological and behavioural interventions, noting that people who are most susceptible to misinformation are often the least likely to participate in inoculation programs or to trust the messengers delivering them. This is compounded by the social nature of misinformation. Traberg, Morton, and van der Linden (2024) study “socially endorsed misinformation,” showing that when false beliefs are supported by one’s social network, they become much more resistant to correction. In such contexts, individual-level interventions may be insufficient. Roozenbeek and van der Linden (2024) acknowledge that misinformation is not just a cognitive problem but also a social and emotional one, requiring interventions that address identity, trust, and group dynamics. The persistence of misinformation after correction, the amplifying role of outrage, and the concentration of manipulation in organized networks all suggest that no single intervention will be a silver bullet. Instead, the evidence points to the need for layered defenses: prebunking to build resilience, debunking to correct false beliefs, accuracy prompts to encourage deliberation, and structural changes to reduce the spread of emotionally manipulative content.
Looking forward, the scientific consensus is that misinformation must be addressed at multiple levels simultaneously. The APA consensus statement (van der Linden et al., 2025) calls for integrating psychological insights with platform governance, public health communication, and media literacy education. Van der Linden’s book Foolproof (2023) argues that the goal is not to make people more cynical but to build “cognitive immunity” so that people can recognize manipulation tactics before they take hold. This requires a shift from reactive debunking—which is often too late and too slow—to proactive prebunking that targets the techniques behind misinformation rather than individual false claims. Jigsaw’s prebunking campaign in Germany and the BBC’s exploration of “vaccinating” audiences against disinformation demonstrate that large-scale implementation is feasible. However, as Roozenbeek and colleagues (2025) note, interventions must be designed with real-world constraints in mind: attention is scarce, trust is fragmented, and engagement metrics reward outrage. Media literacy programs, while helpful, are not sufficient if the underlying platform incentives continue to reward misinformation. The research reviewed here also highlights the importance of addressing misinformation as a global phenomenon. Roozenbeek, van der Linden, and Nygren (2020) showed that prebunking interventions can work across cultures, but local adaptation is essential. The spread of misinformation is not inevitable, but countering it requires sustained investment in research, public education, and platform accountability. As Ecker and colleagues (2024) argue, misinformation poses a bigger threat to democracy than many think, but the good news is that the science of misinformation has advanced rapidly. We now have a rich understanding of why falsehoods spread, who is most susceptible, and what interventions work—at least in some contexts. The challenge for the coming years is to translate that knowledge into effective, scalable, and equitable strategies. The stakes could not be higher: the integrity of elections, the success of public health campaigns, and the possibility of collective action on climate change all depend on a shared information environment. The research summarized here provides both a warning and a roadmap. Misinformation exploits our emotions, our social identities, and our cognitive biases, but psychological inoculation, media literacy, and structural reforms can help build resilience. The question is whether societies will invest in these defenses before the next wave of manipulation arrives.