The research, published in Frontiers in Psychiatry, examined a 45-minute online questionnaire completed in February 2022 by adults aged 18 to 65 recruited through the Respondi market-research panel. The final analytic sample comprised 278 participants who identified as female, 285 who identified as male and three who described themselves as diverse. Its mean age was 43.25 years, with a standard deviation of 13.88. Of the final sample, 142 participants, or 25.1%, reported having experienced a mental disorder during the previous year. The investigators excluded three participants from the structural equation model because only three participants selected the diverse gender category, which was too small for meaningful model estimation, although those participants remained in the descriptive statistics. Because the analysis reused data originally collected for a questionnaire-validation project, the researchers did not conduct a prospective power calculation for this particular structural equation model; a post hoc sensitivity analysis indicated more than 99% power for the estimated model.

Epistemic stance was measured with the German version of the 12-item Epistemic Trust, Mistrust and Credulity Questionnaire, which separates trust, mistrust and credulity as non-pathological but potentially vulnerability-related communication styles. Trust reflects willingness to consider socially transmitted information relevant and reliable, whereas mistrust involves heightened vigilance about interpersonal threats and credulity reflects indiscriminate acceptance of information or sources. Mentalizing deficits were assessed with the 15-item Mentalization Questionnaire, on which higher scores indicate less ability to reflect on internal states and hypothesize about the mental states of others. Its internal consistency in the study was α = .87. The researchers used established self-report measures for conspiracy mentality, violence-supporting attitudes and antisemitic attitudes, incorporating two antisemitism items from a German model questionnaire on right-wing extremism. Internal consistency was α = .88 for the conspiracy measure and α = .90 for the violence-supporting measure, while the two-item antisemitism scale had a Spearman-Brown coefficient of .87.

The investigators specified nine indirect pathways in a structural equation model, treating the three epistemic stances as predictors, mentalizing deficits as the proposed mediator and the three extremist-attitude outcomes as dependent variables. Gender and self-reported presence of a mental disorder were entered as control variables. The model fit was acceptable: the Comparative Fit Index was .944, the Root Mean Square Error of Approximation was .046 and the Standardized Root Mean Square Residual was .058. Mentalizing deficits had significant positive associations with violence-supporting attitudes, conspiracy mentality and antisemitic attitudes, with standardized coefficients of .175, .345 and .179, respectively. These are associations estimated in one cross-sectional dataset, not evidence that mentalizing deficits caused the attitudes measured. The strongest of these relationships was the model’s association between mentalizing deficits and conspiracy mentality.

Epistemic mistrust showed the clearest relationship with mentalizing. Participants reporting greater mistrust also reported substantially greater mentalizing deficits, with a standardized coefficient of .704 and p ≤ .001. The corresponding indirect associations were significant for all three extremist-attitude measures: the coefficient was .123 for violence-supporting attitudes, .243 for conspiracy mentality and .128 for antisemitic attitudes. The conspiracy-mentality pathway had the largest estimated indirect effect. Epistemic trust showed only a marginal negative association with mentalizing deficits, at a coefficient of −.103 and p-value of .055, and its indirect association with conspiracy mentality through mentalizing was also marginal, with a coefficient of −.035 and p-value of .056. The indirect paths from trust to violence-supporting and antisemitic attitudes were not significant, with both p-values reported as .095. No indirect pathway involving epistemic credulity reached statistical significance.

Reporting a mental disorder was associated with epistemic credulity, epistemic mistrust and greater mentalizing deficits, but it was not associated with any of the three extremist-attitude measures after the model accounted for the other variables. Female gender was associated with higher epistemic trust, lower antisemitic attitudes and lower support for violence. These control-variable findings do not mean that mental illness either causes or protects against extremist attitudes; the study found no corresponding association between the self-reported mental-disorder variable and the measured attitudes. The statistical model also did not reveal a simple pattern in which greater epistemic trust uniformly predicted greater conspiracy mentality. The marginally significant association involving trust was indirect through mentalizing deficits, not a direct pathway, and most other direct and indirect routes did not reach conventional significance thresholds.

The pattern is compatible with the idea that difficulty interpreting other people as intentional agents can accompany both interpersonal hypervigilance and rigid or distorted beliefs. Under stress, the authors note, mentalizing capacity may deteriorate, increasing the risk of interpreting ambiguous social information in inflexible ways. Effective mentalizing is associated in prior clinical theory with better self-regulation and adaptation, whereas mistrust and credulity combined with mentalizing impairments are associated with interpersonal and self-regulation difficulties. The present results distinguish among forms of distrust. The data did not support a simple continuum in which greater trust automatically produced more extreme beliefs, and credulity did not show significant indirect effects through mentalizing. Instead, the strongest statistical evidence concerned epistemic mistrust and the three measured forms of extremist attitudes, with mentalizing deficits supplying the model’s proposed mediating route.

Several limitations constrain what can be concluded. The design was exploratory and cross-sectional, so it cannot establish temporal order or causation. Reciprocal relationships and shared underlying vulnerabilities remain possible, and other research models treat epistemic stance and mentalizing as mutually related, sequentially ordered or jointly developed. The specified model treated epistemic stance as the predictor and mentalizing as the mediator, drawing on a theoretical distinction between relatively stable communicative stances and a more dynamic, situation-sensitive mentalizing capacity. The self-report instruments, however, measured relatively stable tendencies rather than momentary fluctuations, so that distinction was not directly tested. The three-item mistrust scale had internal consistency of α = .60, below the commonly recommended threshold of .70, introducing measurement error and requiring cautious interpretation, particularly because the analyses were exploratory.

The sample was also not a probability-based national sample: all participants belonged to the same online panel, creating potential self-selection bias, and mean scores for the extremist-attitude measures were generally low. Generalization to groups with higher levels of extremist attitudes is therefore not warranted. The conspiracy instrument assessed a nonspecific tendency toward conspiracy thinking, the violence scale concerned endorsement in hypothetical cases rather than actual behavior, and the two antisemitism items could not capture the full range of antisemitic beliefs. The study therefore cannot establish that participants endorsed specific conspiracy theories, engaged in antisemitic conduct or were prone to violence. It also did not examine postcolonial antisemitism, proneness to violent behavior, situational stress, world events or differences between state and trait factors that might alter the measured relationships.

The authors call for longitudinal or experimental research to examine whether mentalizing could serve as a potential target in efforts intended to strengthen people’s ability to cope with aversive experiences. They also propose further investigation of gender-specific mechanisms because women in this sample reported greater epistemic trust and lower endorsement of violence and antisemitic attitudes. Mentalization-based therapy is designed to improve mentalizing within clinical treatment, but whether targeting mentalizing in the general population would translate into reduced extremist attitudes remains untested. The present study instead maps where associations appeared, where they did not and which uncertainties must be addressed before any intervention claim can be made. Its clearest result is a cross-sectional link among epistemic mistrust, mentalizing deficits and three distinct measures of extremist attitudes; it does not show that mental illness, mistrust or impaired mentalizing causes such attitudes.

Subject of Research: Exploratory associations among epistemic mistrust, mentalizing deficits and extremist attitudes in a German community sample.

Article Title: Epistemic trust, mentalizing and extremist attitudes in a German community sample

Article References: Hauschild, S., Berning, A., Heimert, K., Kasper, L. A., & Taubner, S. (2026). Epistemic trust, mentalizing and extremist attitudes in a German community sample. Frontiers in Psychiatry, 17, Article 1934135. https://doi.org/10.3389/fpsyt.2026.1934135

Image Credits: AI Generated

DOI: 10.3389/fpsyt.2026.1934135

Keywords: epistemic mistrust, mentalizing, conspiracy mentality, antisemitic attitudes, violent extremism, cross-sectional study, structural equation modeling, German community sample, gender differences, mental health, Epistemic, trust