38978 research outputs found
Sort by
The Role of Support Networks in Promoting Sense of Belonging and Mental Health Well-being among International Students: A Scoping Review Protocol
No full textTo map the existing literature on how social support and sense of belonging influence the mental health and well-being of international students
How Do We Learn What to Believe and Who to Trust?
No full textIn everyday life, people constantly revise beliefs, for example about policies and events, based on new information, often from news sources. This process requires not only updating beliefs about the topic (“Is this policy good?”) but also judging the credibility of the source (“Can I trust this news outlet?”). In such cases, it is likely that people are updating both their beliefs based on one piece of information. As a result, people might judge the credibility of the source based on how much it agrees with their existing beliefs, irrespective of actual credibility.
While rational from a probabilistic perspective, this inference has the potential to form a positive feedback loop reinforcing initial beliefs where 1) false beliefs are reinforced by reliance on untrustworthy sources that agree with initial beliefs, and 2) disconfirming evidence is dismissed by judging opposing sources as less credible. This cycle has the potential to perpetuate misinformation and erode the impact of evidence-based reporting.
However, the underlying computational mechanisms of how belief updatings take place in real-world contexts is relatively unknown. While past research has examined belief-updating in simplified tasks, it's unclear whether similar mechanisms operate in real-world contexts, where beliefs are strong and shaped by broader belief systems.
Current research on belief updating under misinformation has primarily followed two distinct paths. Active psychology research leverages real-world media content and news-consumption data to uncover the prevalence of fake news (Allen et al., 2020), external factors that determine how misinformation spreads (Guess et al., 2020), behavioral and intrinsic factors that drive believing in and sharing of misinformation (Pennycook et al., 2018; Pennycook et al., 2019; Chen et al., 2023; Fazio, 2022), and its broader societal impact (Vosoughi et al., 2018). However, uncovering the exact mechanisms by which people deal with and learn about misinformation in the real world is an outstanding question. In parallel, cognitive neuroscience research has shed light on the computational (Kording & Wolpert, 2004, Ma et al., 2006, Behrens et al., 2007) and neural (Behrens et al., 2007; Nassar et al., 2010; McGuire et al., 2014; Gershman et al., 2015, Bruckner et al., 2025) mechanisms of belief updating under uncertainty. By leveraging abstract tasks that combine single-trial measures of learning, biological markers with computational frameworks such as Bayesian inference and reinforcement learning, these studies demonstrate how the brain integrates new information to update existing beliefs. However, because such tasks are intentionally abstract, they fall short of explaining how belief updating unfolds in the complex, information-rich environments of the real world. Unlike in the lab, where participants typically learn about outcomes of little personal relevance and with simplified, one-dimensional feedback, real-world belief updating often concerns topics tied to identity and values. Moreover, real-world evidence is multidimensional, ambiguous, and open to interpretation, meaning that belief updating outside the lab is not just about outcome tracking but also about negotiating how new information fits within one’s broader belief system and sense of self.
Here we aim to combine the two common approaches to studying belief updating in order to gain mechanistic understanding of how individuals update real-world beliefs. To do so, we first developed a computational model that uses Bayesian inference to jointly learn about a source’s credibility and the topic of interest. The key insight from model simulations is that these processes are interconnected: when a source provides information that aligns with one's prior beliefs about a topic, it reinforces credibility and increases trust in that source on other topics. Conversely, information that contradicts expectations reduces trust across all topics from that source. The model tracks how credibility beliefs transfer between topics, creating realistic learning dynamics where early experiences with a source shape subsequent information processing and belief formation. Simulations yielded two key predictions: 1) when a source provides information aligning with prior beliefs, its perceived credibility increases, regardless of its true reliability and vice versa, 2) this heightened credibility, in turn, amplifies belief-updating from that source whereas perceived reduced credibility does not allow learning. Critically, the implication is that even rational Bayesian inference can create a loop that drives people toward increasingly biased information. As a source gains credibility through initial belief confirmation, individuals are more likely to accept its claims on unrelated topics, allowing agreement on one issue to bootstrap acceptance of a broader worldview. This provides a computational account for behavior that is akin to that of real-world echo chambers.
To test these predictions, we have developed and piloted a news-based inference task that uses authentic media content to track how participants update their topic-related beliefs and perceptions of source credibility on a trial-by-trial basis. Each block of the task consists of three phases. 1) Initial belief phase: Participants first report their current beliefs on four real-world topics. These baseline responses are used to curate news snippets that either align or misalign with their initial beliefs. 2) Source Credibility Learning: Participants are then introduced to a specific news source and instructed to evaluate its credibility. They are presented with 20 curated news snippets (covering the same four topics in an interleaved order). After every three snippets, participants rate the perceived credibility of the source using a continuous slider, based on the snippets that they are presented with. 3) Belief Updating: In the final phase, participants first report their beliefs on 20 new topics and are then shown new snippets from the same news source. After each snippet, they indicate how much their belief about that topic changed based on the presented information. Each participant completes two blocks and we manipulate alignment across blocks. On one of the blocks, the news source presents snippets that align on 75% of the topics in the source credibility learning phase (high alignment condition) while on the other block, the news source presents snippets that align only on 25% of the topics in the source credibility learning phase (low alignment condition). We randomize the order of these blocks in a way that half of the participants do the high alignment condition before the low alignment condition and vice versa. We also manipulate the actual credibility of the news source across participants, where half of the participants are presented with news snippets that are evidence-based and fact-checked on 90 % of the snippets, while the other half are presented with snippets that are evidence-based only on 10 % of the snippets
Fear-Based Decision-Making in Public Administration: A Conceptual Framework and Five-Type Typology
No full textPublic administration systems have intensified accountability and oversight to address corruption, inefficiency, and misuse of public resources. These developments have reinforced an audit society in which verification and performance control increasingly shape everyday administrative routines. Although scholarship widely documents proceduralism, risk aversion, and compliance-oriented behavior under audit-intensive regimes, it has paid limited attention to the behavioral mechanism through which these outcomes are systematically produced.
This paper develops the Fear-Based Decision-Making Framework, positioning fear as an institutionalized mechanism linking audit culture and asymmetric accountability to conservative administrative behavior. Using theory-driven conceptual synthesis across Audit Culture Theory, Risk Aversion Theory, and Behavioral Public Administration, the analysis integrates evidence from peer-reviewed literature and institutional reports, with comparative attention to Western and ASEAN governance contexts.
The synthesis shows that expanding audit regimes tend to redefine accountability around documentation and procedural defensibility, encouraging anticipatory compliance and narrowing discretionary space. Conservative administrative behavior commonly manifests as rigid rule adherence, documentation inflation, defensive standardization, and discretion avoidance, reflecting loss-avoidance incentives under conditions of retrospective scrutiny.
The framework specifies five-part typology of fear: anticipatory, sanction, reputational, interpretive, and institutionalized, and explains how ASEAN scholarship often masks fear-driven behavior through proxy framings such as capacity and compliance narratives. By making fear analytically explicit, the paper reframes conservative administration as an accountability design problem and provides a basis for developing testable propositions and governance reforms that protect integrity without systematically constraining discretion and judgment
Behaviour, low basic skills, and the rationale for restorative practice in schools: A scoping review
No full textSchools in Wales are experiencing increasing challenges related to learner behaviour and mental health, with reports indicating rising levels of disruptive and violent behaviour alongside heightened anxiety and emotional distress following the COVID-19 pandemic (Welsh Government, 2025; NASUWT Cymru, 2025). Similar trends have been identified internationally, with evidence of post-pandemic declines in children’s educational attainment and wellbeing across high-income countries (UNICEF, 2025). In response, policy and inspection guidance in Wales increasingly emphasise the importance of universal, whole-school approaches that promote positive relationships, emotional wellbeing, and inclusive school environments (Estyn, 2024).
Restorative practice is one approach that has gained policy support in this context, focusing on repairing harm, strengthening relationships, and supporting accountability following behavioural incidents (Estyn, 2024). While international evidence suggests that restorative approaches may reduce bullying and improve wellbeing (Bonell et al., 2018), the Welsh Government has highlighted that the overall research base remains limited and that findings should be interpreted with caution (Welsh Government, 2025). As restorative practice is increasingly promoted within Welsh schools, there is a clear need to systematically map how it is defined, implemented, and evaluated across educational settings
[Core] Defining the Epistemic Activation Problem: Interim Findings (v0.2.0)
No full textFormal definition of the Epistemic Activation Problem (EAP) — the phenomenon whereby knowledge stored in accessible systems does not participate in reasoning unless explicitly activated.
EAP definition is now integrated as Section 1 of the v0.3 compiled paper.
Key files:
- interim_findings.md (v0.2 interim)
- Epistemic_Activation_Problem_Interim_v0.2.0.pd
Napping in Adults of Latin American Origin: A Systematic Review of Napping Patterns and Health Associations
No full textWe conducted a systematic review to examine: (1) differences in napping behaviors between Hispanic/Latino adults and other racial/ethnic groups; (2) predisposing and precipitating factors influencing napping behaviors in Hispanic/Latino adults; (3) associations between napping and other sleep characteristics in Hispanic/Latino adults; and (4) associations between napping and long-term health outcomes in Hispanic/Latino adults
Effects of Low-Level Laser Therapy on Facial Nerve Function and Disability in Bell’s Palsy: A Systematic Review and Meta-Analysis
No full textThis systematic review aims to summarize the effectiveness of low-intensity laser therapy in the treatment of Bell's palsy. Only randomized clinical trials will be included in this study
Gender/Sex Differences in the Impact of Salient Financial Cues on Investment Decisions
No full textThis study aims to investigate the effect of salient financial cues on investors' (1) return expectations, (2) risk expectations and (3) investment propensities, between gender identities measured using the Multidimensional Gender and Sexuality Inventory (MGSI)
Medidas de avaliação do desengajamento moral no esporte: protocolo de uma revisão de escopo
No full textEste protocolo descreve a metodologia para uma revisão de escopo que busca mapear a literatura existente sobre as medidas de avaliação do desengajamento moral no contexto do esporte. Serão investigadas as características dos instrumentos utilizados, contexto de aplicação e propriedades psicométricas reportadas nos estudos analisados. A revisão seguirá a metodologia proposta pelo JBI (Joanna Briggs Institute; Peters et al., 2020) para revisões de escopo e será relatada de acordo com as diretrizes PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews; Tricco et al., 2018). As bases de dados PubMed, SPORTdiscus, PsycArticles e Embase serão sistematicamente pesquisadas. Os critérios de elegibilidade incluirão estudos envolvendo adultos no contexto esportivo, mensuração de desengajamento moral, publicados em inglês, português ou espanhol. Dois revisores independentes realizarão o processo de seleção e extração dos dados, com resolução de conflitos por discussão. Os dados serão sintetizados narrativamente e apresentados em tabelas descritivas