Fletcher, R., & Nielsen, R. K. (2026). What has happened to trust in news in the long run globally?. Mainstream News Media Trust, 55–76. https://doi.org/10.4324/9781003618041-6
Summary
This chapter offers the most empirically expansive descriptive account to date of long-term global trends in public trust in news. Combining two major survey programmes — the WVS/EVS (1981–2022) and the Reuters Institute Digital News Report (2015–2025) — the authors assemble nearly 1.5 million responses across 115 countries and use multi-level modelling to trace how trust in news has evolved over almost half a century. Their central finding is a clear but non-universal decline that began in the 1980s and accelerated around 2000 as digital and platform media displaced television. Crucially, the trend diverges by press-freedom context, with trust actually rising in countries lacking press freedom. The authors are deliberate about their scope, framing the work as rigorous description rather than causal explanation or theory-building.
Key Contributions
- Expands longitudinal trust-in-news research from a prior benchmark of 53 countries/40 years to 115 countries over nearly 50 years.
- Provides a harmonisation strategy for combining conceptually distinct trust measures (WVS/EVS “confidence in the press” and DNR “trust most news most of the time”) into a single longitudinal analysis.
- Addresses a noted gap: trust in news has been studied heavily at the individual level but little at the societal level.
- Situates recent DNR-observed declines within a longer downward arc and links acceleration to the digital, platform-dominated media shift.
- Introduces press-freedom trajectory as an analytic lens revealing divergent trust dynamics.
- Offers competing interpretive frames (optimistic “critical citizens” vs. pessimistic “cynical citizenry”; acquiescence and response bias in unfree contexts) to guide future qualitative work.
Methods
The authors pooled two datasets — WVS/EVS (632,711 responses, 8 waves) and DNR (858,668 responses, 11 waves) — and harmonised divergent question wordings by dichotomising each into “Trust” versus “Lack of trust.” Adapting an approach from work on political trust, they fit linear multi-level models with year predicting trust and random intercepts for country (and dataset in pooled models), estimated non-linearity with a five-degree-of-freedom B-spline, and ran 140 separate OLS models across 94 countries to map country-level trajectories. Countries were classified using Freedom House press-freedom data into consistently Free, Not Free, and changing (Improving, Deteriorating, Fluctuating) categories, then interacted with year. The stance is explicitly descriptive rather than causal.
Findings
- WVS/EVS trust fell ~1 point every 5 years (b=-.002, p<.001), from ~49% in the 1980s to ~39% in the 2020s; DNR declined slightly faster (~2 points/5 years, b=-.004).
- A B-spline shows a slight decline from 1981 accelerating around 2000, coinciding with the mainstreaming of digital media.
- Of 140 country-dataset combinations, 78 showed significant negative trends, 34 positive, and 28 non-significant — confirming a general but not universal decline.
- Between-country and between-dataset variance dwarfs the time trend (low marginal R², e.g. .003 in the pooled model).
- Consistently “Free” countries saw small declines while consistently “Not Free” countries saw trust rise since the 1980s; improving, fluctuating, and deteriorating contexts all declined.
- Even 50 years ago, fewer than half the public trusted the news — majority trust is the exception, not the rule.
- Question wording matters: “confidence in the press” diverges from “trust in news” (e.g. higher in Japan, lower in the UK).
- The US is an outlier — sharply declining in WVS/EVS (50% in 1981 to 23% in 2011) but stable in DNR — not representative of global trends.
Connections
No other papers were provided under this note’s topic(s), so there are no sibling notes to link here. The work extends comparative scholarship on institutional and media trust and debates about “critical citizens” and democratic recession, and would connect naturally to future notes on press-freedom effects and survey harmonisation methodology.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify