Törnberg, P., & Chueri, J. (2026). When do parties lie? Misinformation and radical-right populism across 26 countries. The International Journal of Press/Politics, 19401612241311886. https://doi.org/10.1177/19401612241311886
Summary
This paper asks which political parties actually spread misinformation, and adjudicates between three competing accounts: that misinformation is driven by populism in general, by right-wing ideology, or specifically by the combination — radical-right populism. Using 32 million tweets from parliamentarians across 26 countries (2017–2022), the authors link shared URLs to media factuality ratings and merge these with Parlgov and V-Dem party data. They find that neither populism nor right-wing ideology alone predicts misinformation sharing, but their interaction does: radical-right populist parties are the dominant elite-level vector of low-factuality content. The paper reframes misinformation as a strategic feature of comparative party politics rather than a generic platform or populist pathology.
Key Contributions
- First large-scale cross-national comparative analysis of elite misinformation sharing, covering 26 countries and 177 parties.
- A publicly released dataset linking parliamentarian tweets, URL-level factuality scores, and party-level political variables.
- Theoretical adjudication supporting a “disinformation order” account (radical-right populism) over generic populism or right-ideology explanations.
- A reusable methodology for party-level factuality measurement built on URL sharing plus media factuality databases (MBFC, Wikipedia Fake News list).
- A reframing of misinformation as a problem of comparative party politics rather than primarily of platforms or individuals.
Methods
The authors compile 32M tweets from 8,198 parliamentarians via the Twitter Parliamentarian Database, extract ~18M URLs, and assign factuality scores (0–1) to 582,148 URLs using Media Bias/Fact Check and Wikipedia’s Fake News list. Scores are aggregated to a party-level factuality indicator and merged with Parlgov (party family, government status) and V-Dem (populism index, left-right position, personalization, cohesion). They validate the measure via manual coding of 250 stratified-sampled articles. Generalized linear mixed-effects beta regressions with random country intercepts test a populism × left-right interaction, plus a party-family specification.
Findings
- Bivariate associations between populism, right-wing ideology, and low factuality exist but are weak, with high variance among populist parties.
- Neither populism nor left-right position has a significant main effect in regression; their interaction is significant.
- Among non-populist parties, economic ideology does not predict misinformation sharing.
- Among populist parties, moving rightward sharply raises the probability of sharing low-factuality sources.
- Radical-right parties have the highest predicted misinformation levels, significantly above all party families except Conservative and Single-issue.
- Conservative parties’ elevated levels likely reflect blurring boundaries with the radical right.
- Socialist/left, Christian Democratic, and Liberal parties cluster at comparably high factuality.
- Left-wing populists target economic rather than cultural/media elites, making disinformation less central to their repertoire.
Connections
This paper complements work locating misinformation dynamics in elite and partisan production rather than diffuse user behavior, resonating with Allen2025-ot on the outsized role of a small set of partisan actors and with Lewandowsky2026-ob and Hameleers2026-mc on the politics of post-truth communication. It also speaks to platform-level studies of radical-right ecosystems and amplification such as Tornberg2026-lc and Pierri2025-hm, and to broader debates on how to conceptualize and measure misinformation at scale exemplified by Bak-Coleman2026-mk and Efstratiou2025-gs.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify · Apple Podcasts