Fake News on Twitter during the 2016 U.S. Presidential Election
Science, 2019 · paper · Media studies
Examines exposure to and sharing of fake news by registered voters on Twitter during the 2016 election. Engagement was highly concentrated: 1% of individuals accounted for 80% of exposures. Heavy engagers tended to be conservative, older and politically engaged, and most news exposure was still mainstream.
Checked: PDF from web.archive.org, fetched 5 Oct 2026. Page numbers are those of that PDF.
0 findings
None that passed a second check.
4 minor (typos and the like; most checked by the judge only)
Wording Journal summary on the appended page: "Fake news accounted for nearly 6% of all news consumption"; the article reports "5.0% of aggregate exposures to political URLs" and "6.7% of political URLs shared".
- p. 6 Journal summary 'Finding facts about fake news'
Fake news accounted for nearly 6% of all
- p. 1 Results, Prevalence and concentration
5.0% of ag-
- p. 2 Results, Prevalence and concentration
6.7% of po-
Judge's note: Page 6 image shows the summary sentence as quoted. The article's exposure share is 5.0% (page 1), matching the category breakdown 2.4, 1.9 and 0.7% on page 2; 6.7% is the sharing share; the Discussion's 6% is the share of people who shared any fake news content. None is a figure of nearly 6% of consumption, and the article's denominators are political URLs, not all news. The text is the journal's editor summary on the cover page, not the authors' article, so it is recorded as an observation.
Second check: another model agreed
Report a problem with this finding
Wording Discussion: "content from fake news sources constituted only 1.18% of political exposures, or about 10 URLs during the last month of the election campaign"; the Results give 204 potential exposures, and about 10 only "If 5% of potential exposures were actually seen (16)", excluding outliers.
- p. 4 Discussion, 1st paragraph
1.18% of political exposures, or about 10 URLs
- p. 2 Results, Who was exposed to fake news sources?
Excluding outliers, panel members averaged 204
- p. 2 Results, same paragraph
If 5% of potential ex-
Judge's note: Compared page 4 Discussion with page 2 Results on the page images. The Results state 204 potential exposures (95% CI 185 to 224) for non-outlier panel members and about 10 after assuming 5% were seen; 1.18% is the average share of fake news sources among political URLs in a feed. The Discussion joins the share and the count with "or" and omits the 5% assumption and the outlier exclusion. A summary that drops qualifiers stated earlier, not a wrong number.
Second check: another model agreed
Report a problem with this finding
Typo "only 1% of content posted by nonsupersharers. (SM S.15). Finally": a full stop precedes the parenthetical pointer, which then stands as a fragment with its own full stop.
- p. 4 Discussion, final paragraph, middle column
only 1% of content posted by nonsupersharers. (SM S.15).
- p. 4 Discussion, 3rd paragraph
only 1% of content posted by nonsupersharers. (SM S.15). Finally
Judge's note: Page 4 image, middle column: "nonsupersharers." ends one line and "(SM S.15)." begins the next, both full stops printed. Elsewhere the pointer sits inside its sentence, e.g. "(SM S.15)." on page 2. Evident fix: "nonsupersharers (SM S.15)."
Second check: none (judge only)
Report a problem with this finding
Typo Funding: "support by the ESRC ES N012283/1" prints the grant reference with a space where a slash would be expected (ES/N012283/1).
- p. 5 Acknowledgments, Funding
ESRC ES N012283/1
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: ESRC grant references take the form ES/N012283/1. The page 5 image shows "ESRC ES N012283/1" with a space after ES and the second slash kept. The format of the grant number cannot be checked from the files.
Second check: none (judge only)