The study, led by Maria Flakus of the Institute of Philosophy and Sociology at the Polish Academy of Sciences and colleagues, examined two cross-sectional samples: 397 respondents surveyed in May 2022 and 1,066 respondents surveyed in August 2022. Participants indicated which sources they relied on most for information about the war, choosing up to three from a list that included broadcast and online TV or radio, printed press, online news portals, and social media platforms such as Facebook and Twitter/X. The researchers then measured belief in pro-Russian disinformation using questionnaires covering narratives that justified Russia’s attack, denied or minimized Russian responsibility for civilian destruction, and portrayed Ukrainian refugees as undeserving of support.

The statistical approach was a path analysis—a form of structural equation modeling with manifest variables—estimated separately for each study. Because the models included many predictors and interaction terms, the team applied Holm corrections for multiple testing, a procedure that adjusts p-values within predefined families of tests to keep the overall risk of false positives in check. Confidence intervals were estimated with 5,000 bootstrap replications, and all models were adjusted for age, gender, education, place of residence, and political conservatism.

The headline finding concerns what did predict disinformation beliefs. In both studies, endorsement of classic conspiracy theories—captured with brief items about the 9/11 inside-job narrative and the New World Order—was strongly and positively associated with belief in pro-Russian disinformation, with correlations of roughly 0.65 to 0.66. Political paranoia, measured in the first study with a ten-item scale assessing unfounded suspicion that political decisions are made in secret, correlated at 0.63 with disinformation beliefs. In the second study, COVID-19 conspiracy beliefs also correlated positively, at a more moderate 0.42. These belief-based indicators were descriptively stronger and more stable predictors than any of the media-source measures.

By contrast, the media-source associations were weak and inconsistent. Reliance on social media was not positively associated with disinformation beliefs in either study—in fact, the raw association trended slightly negative in the first survey. Broadcast media reliance showed weak negative associations that did not survive correction for multiple testing. The only media indicator that remained robustly associated after correction was reliance on online news portals in the second study, where it was negatively associated with disinformation beliefs, meaning that people who named portals among their top three sources were less likely to endorse pro-Russian narratives.

The moderation results were the most theoretically interesting part of the analysis. The researchers hypothesized that distrust-related dispositions—political paranoia and conspiracy endorsement—might condition how media reliance relates to disinformation beliefs, and they outlined two competing scenarios. In a susceptibility amplification pattern, high conspiracist dispositions would strengthen the link between media reliance and disinformation, consistent with confirmatory, threat-based processing. In a skepticism or attenuation pattern, the same dispositions would weaken or even reverse the link, because deeply distrustful individuals might reject narratives they perceive as emanating from hostile or manipulative actors, regardless of the channel.

In the first study, the interaction between broadcast media reliance and classic conspiracy endorsement was supported after correction: at higher levels of conspiracy endorsement, the association between broadcast reliance and disinformation beliefs became more negative. Political paranoia, by contrast, did not show independent moderating effects once conspiracy endorsement was modeled simultaneously, suggesting that the shared distrust core of the two constructs accounts for much of the apparent paranoia-specific conditionality. In the second study, the picture grew more differentiated. Classic conspiracy endorsement again strengthened negative associations for both broadcast and online portal reliance, but COVID-19 conspiracy beliefs shifted the broadcast association in the opposite direction, making it more positive at higher levels of endorsement.

The authors interpret this divergence through the lens of motivated skepticism and epistemic vigilance. Classic conspiracy beliefs appear to function as a broad, content-independent lens of suspicion—a general tendency to suspect covert malevolent agency behind major events. When such suspicion is high, individuals may scrutinize and reject disinformation narratives that seem institutionally driven or misaligned with their worldview, producing the attenuation pattern. Content-specific conspiracy beliefs, such as those about COVID-19, may instead operate as anti-institutional priors tied to distrust of governments, public-health authorities, and scientific institutions. Among people high in these beliefs, broadcast coverage may be interpreted not as corrective information but as part of a distrusted mainstream communication environment, consistent with the hostile media effect, in which attitude-incongruent content elicits counter-arguing rather than belief revision.

The findings challenge the widespread assumption that social media reliance necessarily heightens vulnerability to disinformation. In the highly salient, morally charged context of the Ukraine war, social media environments may host corrective content, fact-checking, community notes, and peer correction alongside misinformation, and aggregate measures of platform reliance may mask substantial heterogeneity across users and networks. The authors caution, however, that their social media measure captured only whether respondents selected Facebook or Twitter/X among their top three sources, not what content they actually encountered, so any protective interpretation remains context-dependent rather than a stable cross-study conclusion.

Several limitations temper the conclusions. Both studies were cross-sectional, precluding causal inference and leaving open the possibility of reverse causality, in which people less inclined to endorse disinformation preferentially choose particular sources. The media-reliance indicators were coarse binary measures of source selection, while the conspiracy and disinformation measures were fine-grained belief scales, so the stronger associations for belief-based predictors should be read as descriptive rather than as evidence that dispositions matter more than information environments. Sampling also varied: the first study recruited via Facebook advertisements and snowball sharing and overrepresented women and university-educated respondents, while the second used an online panel with demographic quotas but was still not probability-based. Nevertheless, the overall message is clear: understanding who endorses geopolitical disinformation requires attention not only to where people get their information but also to the distrust-related lenses through which they evaluate it.

Subject of Research: Conditional associations between media-source reliance, conspiracy beliefs, political paranoia, and pro-Russian disinformation beliefs during the war in Ukraine

Article Title: When conspiracist beliefs matter: conditional associations between media-source reliance and pro-Russian disinformation beliefs during the war in Ukraine

Article References: Flakus, M., Waszkowski, A., Plisiecki, H., Małyska, K., & Pokropek, A. (2026). When conspiracist beliefs matter: conditional associations between media-source reliance and pro-Russian disinformation beliefs during the war in Ukraine. Frontiers in Social Psychology, 4, Article 1886614. https://doi.org/10.3389/frsps.2026.1886614

Image Credits: AI Generated

DOI: 10.3389/frsps.2026.1886614

Keywords: disinformation, conspiracy beliefs, political paranoia, social media, broadcast media, Ukraine war, Poland, media reliance, motivated skepticism, epistemic vigilance, pro-Russian narratives, path analysis