Selected Publications

False news—given its capacity to distort public opinion and erode trust—has prompted extensive research on potential countermeasures. Yet, there has been no systematic, comparative, and computational investigation of false news sharing and how best to curb it. To address this gap, we apply a semi-integrative experimental approach that (a) compares multiple existing false news interventions, (b) examines how individual and news-level factors predict false news sharing and shape intervention efficacy, and (c) uses drift-diffusion modeling to uncover the decision-making processes underlying all these effects. We find warning labels and media literacy tips to substantially improve news-sharing quality, whereas social norm cues exert a comparatively modest effect, and accuracy prompts yield only subtle benefits. Although numerous individual factors (e.g., age, political conservatism, social media use) predicted news-sharing quality, the observed intervention effects remained broadly robust across these factors, proving effective even within at-risk populations. Intervention outcomes were likewise robust to news-level variation, such as the believability, sensationalism, and political congruence of news content. Despite this robustness, we find each intervention to operate via distinct decision-making pathways. Warning labels shift initial sharing intentions toward sharing higher quality news, whereas media literacy tips operate later, enhancing the processing of news content and increasing cautiousness before making sharing decisions. By applying a multicomponent experimental framework, this work clarifies the risk factors and decision-making processes driving false news sharing and pinpoints which interventions work best, how they operate at the process level, and in which contexts they should be most effective.

Nearly five billion people use and receive news through social media and there is widespread concern about the negative consequences of misinformation on social media (e.g., election interference, vaccine hesitancy). Despite a burgeoning body of research on misinformation, it remains largely unclear who is susceptible to misinformation and why. To address this, we conducted a systematic individual participant data meta-analysis covering 256,337 unique choices made by 11,561 US-based participants across 31 experiments. Our meta-analysis reveals the impact of key demographic and psychological factors on online misinformation veracity judgments. We also disentangle the ability to discern between true and false news (discrimination ability) from response bias, that is, the tendency to label news as either true (true-news bias) or false (false-news bias). Across all studies, participants were well above-chance accurate for both true (68.51%) and false (67.24%) news headlines. We find that older age, higher analytical thinking skills, and identifying as a Democrat are associated with higher discrimination ability. Additionally, older age and higher analytical thinking skills are associated with a false-news bias (caution). In contrast, ideological congruency (alignment of participants’ ideology with news), motivated reflection, and self-reported familiarity with news are associated with a true-news bias (naïvety). Displaying sources alongside news headlines is associated with improved discrimination ability, with Republicans benefiting more from source display. Our results provide critical insights that can help inform the design of targeted interventions.

False beliefs pose significant societal threats, including health risks, political polarization and even violence. In two studies ( N = 884) we explored the efficacy of an individual‐based approach to correcting false beliefs. We examined whether the character virtue of intellectual humility (IH)—an appreciation of one's intellectual boundaries—encourages revising one's false beliefs in response to counter‐information. Our research produced encouraging but also mixed findings. Among participants who held false beliefs about the risks of vaccines (Study 1) and the 2020 US Election being rigged (Study 2), those with higher IH explored more information opposing these false beliefs. This exploration of opposing information, in turn, predicted updating away from these inaccurate health and political beliefs. IH did not directly predict updating away from false beliefs, however, suggesting that this effect—if it exists—may not be particularly powerful. Taken together, these results provide moderate support for IH as a character trait that can foster belief revision but, simultaneously, suggest that alternate pathways to combat false beliefs and misinformation may be preferred.

Numerous polls suggest that COVID-19 is a profoundly partisan issue in the United States. Using the geotracking data of 15 million smartphones per day, we found that US counties that voted for Donald Trump (Republican) over Hillary Clinton (Democrat) in the 2016 presidential election exhibited 14% less physical distancing between March and May 2020. Partisanship was more strongly associated with physical distancing than numerous other factors, including counties' COVID-19 cases, population density, median income, and racial and age demographics. Contrary to our predictions, the observed partisan gap strengthened over time and remained when stay-at-home orders were active. Additionally, county-level consumption of conservative media (Fox News) was related to reduced physical distancing. Finally, the observed partisan differences in distancing were associated with subsequently higher COVID-19 infection and fatality growth rates in pro-Trump counties. Taken together, these data suggest that US citizens' responses to COVID-19 are subject to a deep-and consequential-partisan divide.

Advances in AI require a revision of the psychological and socio-technical dynamics by which individuals are radicalized to embrace violent extremism. This review synthesizes process models of radicalization with research on social and personality risk factors, AI, and psychological mechanisms to propose a four-stage framework mapping the AI architecture of radicalization: (1) Exposure, where recommender systems and virality features create initial attraction to extreme content; (2) Reinforcement, where filter bubbles and group recommendations leverage biases to strengthen extremist beliefs and create echo chambers; (3) Group Integration, where ideologically homogenous clusters, AI bot swarms and companions foster group belonging and readiness for action; cumulatively resulting in (4) Violent Extremist Action. We examine how established social, cognitive, personality, and contextual vulnerability factors heighten psychological risk in the AI-driven radicalization process, as well as the emerging role of generative AI. We conclude by outlining a stage-based framework for governance and future research.

Recently, the world has experienced a wave of violent protest, and in particular Islamist and right-wing extremism have become increasing challenges for many societies. We argue that especially the experience of relative deprivation, that is the perception that oneself or one's group is undeservingly worse off than others, can explain various, contemporary forms of violent extremism, including (a) low-power groups' violent attempts to challenge the unequal status quo, (b) high-power groups' violent defense of their privileged position, and sometimes even (c) people's violent attempt to help out-groups in need. In light of recent research and growing social inequalities, we expect relative deprivation to be a key factor driving violent extremism across cultures and contexts in the 21st century.

Research has documented an overlap between people's aversion toward nonsocial pattern deviancy (e.g., a row of triangles with 1 triangle out of line) and their social prejudice. It is unknown which processes underlie this association, however, and whether this link is causal. We propose that pattern deviancy aversion may contribute to prejudice by heightening people's dislike of statistical minorities. Infrequent people in a population are pattern deviant in that they disrupt the statistical regularities of how people tend to look, think, and act in society, and this deviancy should incite others' prejudice. Nine studies ( N = 1,821) supported this mediation. In Studies 1.1 and 1.2, adults' and young children's nonsocial pattern deviancy aversion related to disliking novel statistical minorities, and this dislike predicted prejudice against Black people. Studies 1.3 and 1.4 observed this mediation when experimentally manipulating pattern deviancy aversion, although pattern deviancy aversion did not directly impact racial prejudice. Study-set 2 replicated the proposed mediation in terms of prejudice against other commonly stigmatized individuals (e.g., someone with a physical disability). Importantly, we also found pattern deviancy aversion to affect such prejudice. Study-set 3 provided additional support for the mediation model. Pattern deviancy aversion predicted prejudice dependent on group-size, for instance, greater racial prejudice in cases where Black people are the statistical minority, but decreased racial prejudice when Black people are the statistical majority. Taken together, these findings indicate that people's aversion toward pattern deviancy motivates prejudice, and that this influence is partially driven by a dislike of statistical minorities. (PsycInfo Database Record (c) 2020 APA, all rights reserved).

The amount of scholarly attention directed at resolving the question why people turn out to cast a vote is vast. In a research field dominated by empirical studies—such as the one on voter turnout—an overview of where we stand and what we know is not superfluous. Therefore, the present paper reviews and assesses the empirical evidence brought forward through a meta-analysis of 83 aggregate-level studies. We thereby concentrate on the effect of socio-economic, political and institutional variables. The results argue for the introduction of a ‘core’ model of voter turnout—including, among other elements, population size and election closeness—that can be used as a starting point for extending our knowledge on why people vote.

Research about voter turnout has expanded rapidly in recent years. This article takes stock of this development by extending the meta-analysis of Geys (2006) in two main ways. First, we add 102 studies published between 2002 and 2015 to the initial sample of 83 studies. Overall, we document only minor changes to the original inferences. Second, since different processes might conceivably play at different levels of government, we exploit the larger sample to separately analyse the determinants of voter turnout in national versus subnational elections. We find that campaign expenditures, election closeness and registration requirements have more explanatory power in national elections, whereas population size and composition, concurrent elections, and the electoral system play a more important role for explaining turnout in subnational elections.

Democratic backsliding in European Union (EU) member states is not only a policy challenge for the EU, but also a potential existential crisis. If the EU does too little to deal with member state regimes that go back on their commitments to democracy and the rule of law, this risks undermining the EU from within. On the other hand, if the EU takes drastic action, this might split the EU. This article explores the nature and dynamics of democratic backsliding in EU member states, and analyses the EU’s capacity, policy tools and political will to address the challenge. Empirically it draws on the cases that have promoted serious criticism from the Commission and the European Parliament: Hungary, Poland, and to a lesser extent, Romania. After reviewing the literature and defining backsliding as a gradual, deliberate, but open-ended process of de-democratization , the article analyzes the dynamics of backsliding and the EU’s difficulties in dealing with this challenge to liberal democracy and the rule of law. The Hungarian and Polish populist right’s “illiberal” projects involve centralization of power in the hands of the executive and the party, and limiting the independence of the judiciary, the media and civil society. This has brought both governments into direct confrontation with the European Commission. However, the EU’s track record in managing backsliding crises is at best mixed. This comes down to a combination of limited tools and lack of political will. Ordinary infringement procedures offer a limited toolbox, and the Commission has proven reluctant to use even these tools fully. At the same time, party groups in the European Parliament and many member state governments have been reluctant to criticize one of their own, let alone go down the path of suspending aspect of a states’ EU membership. Hence the EU’s dilemma: it is caught between undermining its own values and cohesion through inaction on one hand, and relegating one or more member states it to a second tier—or even pushing them out altogether—on the other.

Scandinavian party competition has incorporated divisions over European integration to a greater degree than most West European party systems, but with considerable variation in Norway, Sweden and Denmark. From a comparative politics perspective this raises questions about the relatively high salience of Euro-scepticism in Scandinavian politics, the differences between the three cases and changes over time. The central argument in this article is that Europeanisation of party politics—the translation of issues related to European integration into domestic party politics—is driven by the dynamics of long- and short-term government-opposition competition, and the key driver of change is party strategy. Whether at the centre or extremes of the party system, Euro-scepticism is a product of party competition—and is, both in its origins and development, ‘the politics of opposition’.

We investigate how respectful leadership can help overcome the challenges for follower performance that female leaders face when working (especially with male) followers. First, based on role congruity theory, we illustrate the biases faced by female leaders. Second, based on research on gender (dis-)similarity, we propose that these biases should be particularly pronounced when working with a male follower. Finally, we propose that respectful leadership is most conducive to performance in female leader–male follower dyads compared with all other gender configurations. A multi-source field study ( N = 214) provides partial support for our hypothesis. While our hypothesized effect was confirmed, respectful leadership seems to be generally effective for female leaders irrespective of follower gender, thus lending greater support in this context to the arguments of role congruity rather than gender dissimilarity.

Advances in AI offer the prospect of manipulating beliefs and behaviors on a population-wide level. Large language models and autonomous agents now let influence campaigns reach unprecedented scale and precision. Generative tools can expand propaganda output without sacrificing credibility and inexpensively create falsehoods that are rated as more human-like than those written by humans. Techniques meant to refine AI reasoning, such as chain-of-thought prompting, can just as effectively be used to generate more convincing falsehoods. Enabled by these capabilities, a disruptive threat is emerging: swarms of collaborative, malicious AI agents. Fusing LLM reasoning with multi-agent architectures, these systems are capable of coordinating autonomously, infiltrating communities, and fabricating consensus efficiently. By adaptively mimicking human social dynamics, they threaten democracy. Because the resulting harms stem from design, commercial incentives, and governance, interventions are needed at multiple leverage points, focusing on pragmatic mechanisms over voluntary compliance.

The COVID-19 pandemic has been accompanied by a surge of misinformation on social media which covered a wide range of different topics and contained many competing narratives, including conspiracy theories. To study such conspiracy theories, we created a dataset of 3495 tweets with manual labeling of the stance of each tweet w.r.t. 12 different conspiracy topics. The dataset thus contains almost 42,000 labels, each of which determined by majority among three expert annotators. The dataset was selected from COVID-19 related Twitter data spanning from January 2020 to June 2021 using a list of 54 keywords. The dataset can be used to train machine learning based classifiers for both stance and topic detection, either individually or simultaneously. BERT was used successfully for the combined task. The dataset can also be used to further study the prevalence of different conspiracy narratives. To this end we qualitatively analyze the tweets, discussing the structure of conspiracy narratives that are frequently found in the dataset. Furthermore, we illustrate the interconnection between the conspiracy categories as well as the keywords.

The COVID-19 pandemic has severely affected the lives of people worldwide, and consequently, it has dominated world news since March 2020. Thus, it is no surprise that it has also been the topic of a massive amount of misinformation, which was most likely amplified by the fact that many details about the virus were not known at the start of the pandemic. While a large amount of this misinformation was harmless, some narratives spread quickly and had a dramatic real-world effect. Such events are called digital wildfires. In this paper we study a specific digital wildfire: the idea that the COVID-19 outbreak is somehow connected to the introduction of 5G wireless technology, which caused real-world harm in April 2020 and beyond. By analyzing early social media contents we investigate the origin of this digital wildfire and the developments that lead to its wide spread. We show how the initial idea was derived from existing opposition to wireless networks, how videos rather than tweets played a crucial role in its propagation, and how commercial interests can partially explain the wide distribution of this particular piece of misinformation. We then illustrate how the initial events in the UK were echoed several months later in different countries around the world.

Some disciplines in the social sciences rely heavily on collecting survey responses to detect empirical relationships among variables. We explored whether these relationships were a priori predictable from the semantic properties of the survey items, using language processing algorithms which are now available as new research methods. Language processing algorithms were used to calculate the semantic similarity among all items in state-of-the-art surveys from Organisational Behaviour research. These surveys covered areas such as transformational leadership, work motivation and work outcomes. This information was used to explain and predict the response patterns from real subjects. Semantic algorithms explained 60-86% of the variance in the response patterns and allowed remarkably precise prediction of survey responses from humans, except in a personality test. Even the relationships between independent and their purported dependent variables were accurately predicted. This raises concern about the empirical nature of data collected through some surveys if results are already given a priori through the way subjects are being asked. Survey response patterns seem heavily determined by semantics. Language algorithms may suggest these prior to administering a survey. This study suggests that semantic algorithms are becoming new tools for the social sciences, opening perspectives on survey responses that prevalent psychometric theory cannot explain.

This is a review of a range of empirical studies that use digital text algorithms to predict and model response patterns from humans to Likert-scale items, using texts only as inputs. The studies show that statistics used in construct validation is predictable on sample and individual levels, that this happens across languages and cultures, and that the relationship between variables are often semantic instead of empirical. That is, the relationships among variables are given a priori and evidently computable as such. We explain this by replacing the idea of “nomological networks” with “semantic networks” to designate computable relationships between abstract concepts. Understanding constructs as nodes in semantic networks makes it clear why psychological research has produced constant average explained variance at 42% since 1956. Together, these findings shed new light on the formidable capability of human minds to operate with fast and intersubjectively similar semantic processing. Our review identifies a categorical error present in much psychological research, measuring representations instead of the purportedly represented. We discuss how this has grave consequences for the empirical truth in research using traditional psychometric methods.

Abstract The rapid innovation in artificial intelligence (AI) is raising concerns regarding human autonomy, agency, fairness, and justice. While responsible stewardship of innovation calls for public engagement, inclusiveness, and informed discourse, AI seemingly challenges such informed discourse by way of its opacity (poor transparency, explainability, and accountability). We apply a deliberative approach to propose a framework for responsible innovation in AI. This framework foregrounds discourse principles geared to help offset these opacity challenges. To support better public governance, we consider the mutual roles and dependencies of organizations that develop and apply AI, as well as civil society actors, and investigative media in exploring pathways for responsible AI innovation.

While organizations today make extensive use of complex algorithms, the notion of algorithmic accountability remains an elusive ideal due to the opacity and fluidity of algorithms. In this article, we develop a framework for managing algorithmic accountability that highlights three interrelated dimensions: reputational concerns, engagement strategies, and discourse principles. The framework clarifies (a) that accountability processes for algorithms are driven by reputational concerns about the epistemic setup, opacity, and outcomes of algorithms; (b) that the way in which organizations practically engage with emergent expectations about algorithms may be manipulative, adaptive, or moral; and (c) that when accountability relationships are heavily burdened by the opacity and fluidity of complex algorithmic systems, the emphasis of engagement should shift to a rational communication process through which a continuous and tentative assessment of the development, workings, and consequences of algorithms can be achieved over time. The degree to which such engagement is, in fact, rational can be assessed based on four discourse-ethical principles of participation, comprehension, multivocality, and responsiveness. We conclude that the framework may help organizations and their environments to jointly work toward greater accountability for complex algorithms. It may further help organizations in reputational positioning surrounding accountability issues. The discourse-ethical principles introduced in this article are meant to elevate these positioning contests to extend beyond mere adaption or compliance and help guide organizations to find moral and forward-looking solutions to accountability issues.