"The loss of a shared reality is one of the greatest threats to democratic societies"

Misinformation is increasingly shaping how societies understand reality. But why do current interventions fail to reach those who matter most? Ralph Hertwig explains why the problem runs deeper.
 

How effective are interventions against misinformation?

In recent years, numerous tools have been developed to help people distinguish between trustworthy and untrustworthy information—for instance, accuracy nudges and training in lateral reading. We've been involved in this work, developing and researching interventions that contribute to a kind of anti-misinformation toolkit.

But these interventions are based on the assumption that people are motivated to avoid misinformation. In other words, that they want to get to the truth but may fail to do so due to inattention or a lack of necessary skills. In our own studies, however, it has become increasingly clear that this assumption does not hold for crucial groups.

Which groups are we talking about?

Our research has shown that people with strong political convictions either don't use the tools or don't find them helpful. That doesn't mean that the tools are ineffective. They work well for many people. But they fail to reach precisely those groups that matter most. And that fact was the proverbial elephant in the room.

It became clear to me that the problem runs much deeper. It's less a problem of individual pieces of misinformation or deficits in media literacy, and more a question of differing epistemic standards—that is, which sources and institutions people trust and what criteria they use to decide what is true. Ultimately, it comes down to the question of whether we still have a shared reality.
 

What do you mean by a shared reality?

Of course, the nature of reality is a philosophical question—one we could debate at length. But when I talk about a shared reality, I mean something more pragmatic: that we can at least agree on basic facts. To take an example from the COVID-19 pandemic: Were intensive care units overwhelmed or not? We should be able to agree on the basic facts of the matter. Interpretations can then diverge. Some people might attribute an unprecedented strain on the health services to flawed political decision-making, others to the dynamics of the pandemic. Such differing interpretations are part of normal democratic discourse, but they need to have a shared factual basis.

So what I mean by shared reality is that we don't have to agree on causes or interpretations. But we should be able to agree on whether something actually took place or not.

IMPACT OF SOCIAL MEDIA ON THE BRAIN AND MENTAL HEALTH
In a joint project of the Center for Adaptive Rationality and the Center for Environmental Neuroscience, researchers are investigating how social media use affects the brain and mental health of identical twins. Using questionnaires, the researchers assess how frequently participants use social media and whether there are signs of problematic use. Coupled with magnetic resonance imaging (MRI) brain scans and data on mental wellbeing, this comprehensive approach allows them to investigate how social media use relates to brain structure and function, emotions, and cognitive performance. Studying identical twins offers a unique perspective: Because they are genetically identical, any differences observed between them can be attributed to environmental influences like social media use. This approach provides deeper insights into how strongly such influences affect the brain and mental health.

You say our shared reality is crumbling. What do you mean by that?

We, as a society, are finding it increasingly difficult to agree on a shared description of reality. As recently as 20 years ago, most people obtained their information from the same sources—from what are known as the established or mainstream media. This shared experience created a common ground where, come Monday morning, everyone could chat about the latest stories from the evening news or the weekend’s political talk shows.

Unlike in autocratic states, this form of shared reality no longer exists in democratic societies. Instead, we’re seeing increasing fragmentation. People move in different information spaces, often in digital bubbles that reflect very different versions of reality. Many people’s perceptions of events differ completely from what is conveyed in the established media.

To stick with the same example: During the COVID-19 pandemic, some people were convinced that claims about hospitals being overwhelmed were fabricated. We’re seeing similar divides on issues such as Russia's war on Ukraine, climate change, bank bailouts, and the efficacy of vaccines. When there is no longer consensus on central aspects of reality, we find ourselves facing a fragmented reality.

Why is that so problematic for democratic societies?

In my view, the loss of a shared reality is one of the main dangers facing democratic societies (see info box below). If we can no longer agree on a common factual basis, we’ll become increasingly unable to organize collective solutions to problems. And yet we’re faced with enormous challenges: the climate crisis, digitalization, demographic change, geopolitical conflicts, and resource scarcity. Many people are calling it a "polycrisis."

All these problems require political decisions and societal acceptance of measures that may entail losses of freedoms, resources, or privileges. But if we can't even agree on what the problem is, there’s little hope that potential solutions will be accepted. This undermines the political system's ability to take action and, in the long term, its legitimacy.

Against this backdrop, it's no coincidence that international experts now rank misinformation and disinformation as one of the greatest global risks—in some cases, even ahead of climate change or military conflicts. The underlying idea is this: If we can no longer agree on a shared reality, we won't be in a position to collectively solve all the other urgent problems.

What role does digital media play in this development?

Digital media plays a central role, though it's not the only cause. Misinformation and disinformation have characteristics that give them a structural advantage in digital environments: They are often negative, emotional, and surprising. And those are precisely the features that grab human attention and elicit strong cognitive reactions.

On top of that, algorithms prioritize content that keeps users hooked. Even if they aren't explicitly programmed to spread misinformation, the result is a dynamic interaction between human attention and technical selection that amplifies misinformation and disinformation.

Another important factor is the emergence of networks of like-minded people. Digital media makes it very easy to connect with people who hold similar views. If I spend most of my time in those kinds of environments, it will reinforce my perspective—even if what I think is inconsistent with established facts.

Let's turn to your current research project. What are your plans?

Together with my colleagues Anastasia Kozyreva, Philipp Lorenz-Spreen, Pietro Nickl, Lisa Oswald, and Carolin Ziemer, I am interested in how people end up in these alternative realities. Which individual life experiences, social setups, and specific events lead to them losing their trust in quality media, institutions, or science. We want to understand whether it's a sudden shift or a gradual process, and whether it's limited to specific topics or is more generalized.

We're asking three core questions. First: Are there identifiable tipping points or triggers at which people begin to question the established description of reality? Second: How does this process continue—is it limited to a single topic, or does it extend to ever more areas? And third: What role do trust, belonging, and social validation play?

A central theme is loss of trust. Many people say that they have doubts not only about individual news reports, but also about the media, government institutions, and in some cases science in general. In our view, this loss of trust is key to understanding why traditional interventions fall short.

What methods are you using?

Because we're dealing with such a complex phenomenon, we deliberately use a mix of methods. We began with qualitative, semi-structured interviews lasting about 90 minutes each. These interviews allow us to trace biographical trajectories, identify personal turning points, and assess subjective interpretations in a lot of detail.

"Semi-structured" means that there is an interview guide with key questions and topics to be covered, but that interviewers can respond flexibly based on the interviewees' responses. This avoids a rigid question-and-answer format and instead fosters a free-flowing conversation that allows for the unexpected to emerge. The individual interviews are supplemented by focus groups, in which participants discuss topics together. These group settings are particularly interesting because they clearly reveal social dynamics, mutual validation, and alienation.

Based on these qualitative insights, we can then develop quantitative assessments—large-scale surveys that allow us to assess how widespread certain patterns and attitudes are. And for a smaller group of participants, we use digital tracking to better understand the information spaces that they frequent and how their media usage relates to their beliefs.

How do you succeed in recruiting people for the interviews?

That was one of the biggest challenges. Many of the people we needed to talk to are not scientifically inclined and might not voluntarily come to a research institute. With that in mind, we collaborate with an experienced public opinion research firm that has the necessary contacts.

It was particularly important for the interviewers to maintain a respectful, neutral stance, even when confronted with highly controversial statements. This approach builds trust and is a necessary condition for people being willing to speak openly about their views.

Can you share any findings yet?

We're seeing a pattern of individual turning points— for some, it was the pandemic for others, the war in Ukraine or the Gaza conflict. Almost without exception, interviewees report a discrepancy between what’s reported in the "mainstream media" and what they are seeing in their own information channels.

The process is rarely abrupt it unfolds gradually. Individuals reassess and ultimately withdraw their trust in established media sources, often permanently.

Do you think there's a chance of returning to a shared reality?

We can't simply accept that societies will inevitably splinter into irreconcilable realities. At the same time, we need to acknowledge that the instruments developed so far aren’t enough. It's not just a problem of information; it’s primarily a question of trust—in the media, institutions, and science. And trust isn't something you can prescribe or legislate. We need to understand why it was lost in the first place. Only then can we start thinking about ways to rebuild it.

That calls for an interdisciplinary perspective. Psychology alone cannot explain these developments; it must learn from sociology, political science, and other fields. We can't focus solely on individual behavior; we also need to consider societal processes, macro-level developments, and cultural change.

In the long term, I hope that this project and our work at the Center for Adaptive Rationality will contribute to a better understanding of these dynamics and pave the way for new, more effective approaches. Not as ready-made solutions, but as a sound foundation for social and political decisions.

STRENGTHENING DEMOCRATIC RESILIENCE
Democracies worldwide are under mounting pressure, often through the gradual erosion of democratic norms. A 2024 study by researchers at the University of Potsdam, the MPI for Human Development, and the University of Bristol examined how democratic backsliding unfolds—and how it can be stopped. The focus is on "near-misses" in countries including Colombia, Sri Lanka, and South Korea, where political elites undermined democratic rules—for example, by concentrating power in the executive or putting pressure on the media—but where courts, civil society, and protest movements were able to halt the drift toward authoritarianism. The analysis draws on the "drift-to-danger" model from accident research, which describes how systems gradually move toward tipping points: While isolated breaches of democratic norms may have no immediate consequences, their cumulative effect over time can erode established safeguards. Risk factors include polarization, populism, and misinformation. The study's findings have been incorporated into teaching materials for schools.

Abels, C. M., Huttunen, K. J. A., Hertwig, R., & Lewandowsky, S. (2024). Dodging the autocratic bullet: Enlisting behavioural science to arrest democratic backsliding. Behavioural Public Policy. Advance online publication. https://doi.org/10.1017/bpp.2024.43