Everyday encounters with misinformation online: examining sources, topics and modes

Summary

This Australian study challenges the dominant framing of misinformation research as a problem of “hot button” topics (health, politics, science) propagated by fringe actors. Using a week-long digital diary method that lets participants flag content themselves — without imposing a pre-set definition of misinformation — the authors find that everyday encounters with perceived misinformation span mundane domains like business/economics, celebrity, and crime; are overwhelmingly text-based (especially news headlines); and are most often attributed to mainstream domestic news outlets engaging in clickbait and sensationalism. The paper argues that misinformation perceptions are inseparable from broader crises of trust in journalism, shifting news logics, and platform dynamics, and that effective responses must therefore address the texture of everyday digital life rather than single-issue interventions.

Key Contributions

• Empirically maps the everyday character of misinformation encounters, moving beyond single-issue and big-data approaches.
• Introduces a participant-driven diary methodology that avoids pre-imposing what counts as misinformation, surfacing user perceptions rather than researcher categories.
• Documents the perception of mainstream news outlets — not fringe sources — as the leading propagators of misleading content, complicating the framing of journalism as a remediator of information disorder.
• Offers a three-dimensional analytic framework (topic × mode × source) for everyday misinformation.
• Derives policy and pedagogical implications for whole-of-society media literacy, journalistic transparency, and context-sensitive interventions.

Methods

A two-wave digital diary study (mid-2024) on the Indeemo platform with 55 purposively sampled Australian adults, balanced for demographics and self-reported misinformation exposure. Participants captured screenshots, photos, and recordings of online information across seven-day windows and narrated their trust judgments via video, yielding 1,564 examples and ~221,735 words of reflection. A 322-item subsample of user-flagged false/misleading/untrustworthy content was analyzed via deductive thematic analysis (Braun & Clarke) across topic, mode, and source dimensions, with strong intercoder reliability (Krippendorff’s α = 1.0 for topic/source; 0.91 for mode). An independent verification pass on 10% of flagged claims assessed objective veracity.

Findings

• Topics: Business/economics led (18%), often exploiting cost-of-living anxieties (e.g., dubious “tax hacks”); celebrity (16%) and crime/crisis (16%) followed.
• Modes: Text dominated at 68%, with news headlines making up 53% of text-based claims; multimodal 18%, video 11%, image 3%, audio 1% — undercutting alarm about audiovisual and AI-generated content.
• Sources: Mainstream/alternative news outlets accounted for 62% of flagged sources, 82% domestic (News Corp 25%, Seven West 17%, Nine 15%); social media was second at 19%, dominated by Meta platforms (Facebook 35%).
• “Clickbait Concern”: Untrustworthiness was most often attributed to clickbait headlines, sensationalism, perceived bias, and headline–article disconnect.
• Perception vs. veracity gap: Independent verification of flagged claims found only 9% objectively false/misleading, 28% indeterminate, 25% containing no verifiable claim — user perceptions of misinformation diverge substantially from objective falsity.

Connections

This paper sits within an audience-centered turn in misinformation research that interrogates how ordinary people perceive and label false content, resonating with work questioning the prevalence and definitional scope of misinformation (e.g., Budak2024-ef, Gonzalez-Bailon2024-rq). Its finding that mainstream news is widely perceived as a misinformation source connects to scholarship on declining institutional trust and journalistic credibility (Humprecht2025-ml, Cazzamatta2026-lo), while its emphasis on everyday text-based encounters over AI/deepfake alarmism complements skepticism about generative-AI panic (DeVerna2025-dl, Appel2026-qr). The methodological move toward situated, participant-driven accounts of news engagement also aligns with diary- and perception-based studies of information disorder (Rossini2026-jn, Hameleers2026-mc).

Podcast

A research-radio episode discusses this paper: Listen