Rossini, P., Kalogeropoulos, A., & Mont’Alverne, C. (2026). Amplifying division: Electoral misinformation and political intolerance in Brazil. Social Media + Society. https://doi.org/10.1177/20563051261419393
Summary
This paper investigates the relationship between belief in electoral misinformation and political intolerance during Brazil’s contested 2022 presidential election. Using a three-wave panel survey of Brazilian internet users, the authors show that citizens who believed false claims of electoral fraud grew more intolerant of political opponents over time. They further identify messaging apps (particularly WhatsApp) as an indirect pathway to intolerance: using these apps for news strengthened misinformation beliefs, which in turn eroded tolerance. The core argument reframes the democratic stakes of misinformation — it does not merely confuse citizens or undermine institutional trust, but corrodes commitment to democratic norms like tolerance toward disliked groups, amplifying pre-existing cleavages such as antipetismo rather than creating division from scratch.
Key Contributions
- First empirical evidence linking electoral misinformation beliefs to political intolerance, moving beyond effects on knowledge and institutional trust.
- Distinguishes messaging apps (WhatsApp) from general social media as an indirect informational route to intolerance via heightened misinformation belief.
- A longitudinal panel design in a major non-Western democracy, addressing a literature skewed toward authoritarian/hybrid regimes and cross-sectional data.
- Reframes misinformation’s democratic threat as the erosion of tolerance norms, not just epistemic confusion.
- Argues fact-checking alone is insufficient for misinformation interventions and electoral regulation.
Methods
- Three-wave nationally representative panel of Brazilian internet users (N=1600 W1, 1328 W2, 1034 W3), fielded by Ipec Inteligência with demographic quotas.
- Waves timed to the 2022 election: W1 after first round, W2 after second round, W3 during the anti-democratic protests.
- Political intolerance measured via a “least-liked” GSS-style approach (extending civil rights to disliked political groups: petistas, bolsonaristas, communists, dictatorship advocates).
- Electoral misinformation belief measured in W2 only (for ethical reasons) using eight fact-checked false statements plus true and placebo controls.
- Autoregressive OLS models predicting W2/W3 intolerance controlling for prior levels, plus SEM mediation models (lavaan, bootstrapped 95% CIs, 5000 samples).
Findings
- Electoral misinformation beliefs significantly predicted increases in intolerance in both W2 (β = −.09) and W3 (β = −.10).
- Social media news use had no significant effect (only marginally positive for tolerance in W3); messaging app news use had no significant direct effect.
- SEM confirmed an indirect effect: W1 messaging app news use → higher W2 misinformation belief → lower W3 tolerance (β = −.009); no equivalent path for social media.
- Engagement with online disagreement was negatively associated with tolerance (W1→W2, β = −.07), contrary to classical deliberation theory.
- 46% of respondents held zero misinformed beliefs; mean score was 1.58 of 8.
Connections
This work sits alongside other studies of the Brazilian information environment and its democratic stakes, most directly Ventura2025-sw, Ventura2026-yc, and Inacio-da-Silva2026-zf, as well as work on WhatsApp and messaging-app misinformation such as Bosch2024-hj. Its concern with attitudinal downstream effects of misinformation and platform use also connects to broader debates on misinformation susceptibility and intervention efficacy, e.g. van-der-Linden2026-jt on inoculation and fact-checking limits, and Arceneaux2026-xk on the affective/attitudinal consequences of exposure.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify · Apple Podcasts