Rossini, P., Mont’Alverne, C., & Kalogeropoulos, A. (2026). Disrupting democracy: how informational pathways and misinformation shape trust in election results. Journal of Elections, Public Opinion and Parties, 1–23. https://doi.org/10.1080/17457289.2026.2701896
Summary
This paper investigates how citizens’ informational pathways and their belief in electoral misinformation shaped trust in the results of the 2022 Brazilian Presidential Election — a contest defined by incumbent Jair Bolsonaro’s systematic disinformation attacks on the integrity of the electronic voting system. Using a three-wave panel survey fielded around the two rounds of voting and results certification, the authors show that mainstream news use bolsters electoral trust over time (even among Bolsonaro voters), while belief in electoral misinformation and voting for Bolsonaro predict lower trust. The core argument reframes disinformation not as isolated false content but as a strategic elite practice embedded in evolving, platform-driven media systems, positioning professional journalism as a source of democratic resilience against elite-driven delegitimization channeled through digital and messaging platforms.
Key Contributions
- First over-time (panel) evidence on how informational pathways and misinformation beliefs shape electoral trust in a major Global South, weak-partisanship, multi-party democracy — extending a US-dominated literature.
- Demonstrates the stabilizing role of mainstream journalism as a counterweight to elite disinformation, even among electoral losers.
- Disaggregates messaging-app use into news consumption versus discussion across strong, weak, and no-tie networks, revealing heterogeneous effects on trust.
- Reframes disinformation as strategic elite behavior embedded in media systems, informing theories of democratic backsliding and resilience.
Methods
- Original three-wave panel survey of Brazilian internet users (W1 = 1,600; W2 = 1,328; W3 = 1,034), fielded October, November, and December 2022 around the two voting rounds and certification.
- Opt-in panels (Ipec Inteligência) with quotas for age, gender, social class, and region.
- Dependent variable: single-item trust in the 2022 election results (4-point scale, adapted from the World Values Survey).
- Independent variables include mainstream news use, social media news use, WhatsApp news use, WhatsApp/Telegram group talk across strong/weak/no-tie networks, and an eight-point “misinformed score” derived from twelve fact-checked statements (measured in Wave 2 only).
- Random effects within-between models separate within-person change from between-person differences; a Bolsonaro-voter subset analysis and Bonferroni correction are applied.
Findings
- Belief in electoral misinformation was negatively associated with trust (between-person β = −.157).
- Increases in mainstream news use predicted increases in trust over time (within-person β = .069).
- Social media news use showed null effects; messaging news use had no over-time effect but a negative between-person association (β = −.129).
- Voting for Bolsonaro strongly predicted lower trust (β = −.946); right-wing ideology remained negative even controlling for vote choice.
- WhatsApp group discussion effects depended on network composition: weak-tie weakly positive, no-tie (stranger) negative, strong-tie non-significant in the full sample.
- Among Bolsonaro voters, mainstream news still raised trust, but messaging news use lowered trust over time (β = −.069), no-tie talk lowered trust, and strong-tie talk raised trust (β = .165).
Connections
This study extends the growing non-US and Brazilian body of work on electoral misinformation and platform ecosystems, connecting closely to Rossini2026-jn and to work on messaging apps and disinformation in the Brazilian context such as Inacio-da-Silva2026-zf and Ramos2026-qo. Its treatment of misinformation belief and resistance links to debates on misperception and inoculation in van-der-Linden2026-jt and Hameleers2026-mc, while its framing of disinformation as strategic elite practice speaks to accounts of platform-embedded influence in Marwick2026-qd.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify