Fake news prevalence¶
Fake news prevalence refers to measures of how widespread exposure to false or misleading content is within a population or across platforms. Prevalence is a distinct concept from impact (whether fake news changes beliefs or behavior) and from production (how much fake news is created).
Prevalence can be measured in several ways: the fraction of people exposed, the fraction of content consumed, the frequency of exposure, or the proportion of shares/engagement relative to other content. Different metrics yield different conclusions about how serious the fake news problem is.
Key observations¶
Prevalence is rare in absolute terms:
When measured as a fraction of total media consumption, fake news comprises only 0.15% of Americans' daily media diet Allen et al. (2020).
Prevalence on social media is higher than in population aggregates but still small:
On Twitter, fake news comprises roughly 1% of news consumption among active users Grinberg et al. (2019), but Twitter users are unrepresentative of the broader population Allen et al. (2020).
Exposure is concentrated among certain demographics:
Older Americans (55+) consume more fake news than younger groups, but even they spend less than 1 minute per day on it; very few people (1–2%) consume more fake news than mainstream news Allen et al. (2020).
Platform selection biases research conclusions:
Studies relying exclusively on social media data dramatically overestimate fake news prevalence compared to the population-representative measure Allen et al. (2020).
Related concepts¶
- Fake news — the content type whose prevalence is being measured
- Media consumption — the denominator against which prevalence is assessed
- Information ecosystems — the larger context in which prevalence is evaluated
- Misinformation — a broader category than fake news; prevalence varies by definition
Key papers in this wiki¶
- Allen et al. (2020) — Evaluating the fake news problem at the scale of the information ecosystem — definitive ecosystem-scale measurement showing fake news is rare (0.15% of total media diet)
- Allcott & Gentzkov (2017) — Social Media and Fake News in the 2016 Election — early empirical measurement showing exposure is rare compared to other content
- Grinberg et al. (2019) — Fake News on Twitter during the 2016 U.S. Presidential Election — platform-specific prevalence measurement
Open challenges¶
- How has fake news prevalence changed over time since 2016–2018?
- How does prevalence differ across countries and media systems?
- How do platform algorithm changes affect measured prevalence?
- Should prevalence be weighted by engagement, time spent, or other factors beyond simple reach?
- How do different definitions of "fake news" affect prevalence estimates?