Exposure to Ideologically Diverse News and Opinion on Facebook
Science, 2015 · paper · Media studies
Analyzes how 10.1 million U.S. Facebook users encounter socially shared news across ideological lines, covering friend networks, News Feed ranking and click choices. Individuals' own choices limited exposure to cross-cutting content more than algorithmic ranking did.
1 finding
Substance The editor's summary (PDF page 4) says "People encountered roughly 15% less cross-cutting content in news feeds due to algorithmic ranking and clicked through to 70% less of this cross-cutting content". The article reports 5%/8% for ranking and 17%/6% for clicks.
- p. 4 Editor's summary 'Not getting all sides of the news?'
15% less cross-cutting content in news feeds due to algorithmic ranking and clicked through to 70% less of this cross-
- p. 2 Main text, col 2, ranking effect
is 5% for conservatives and 8% for liberals
- p. 2 Main text, col 3, click effect
to be 17% for conservatives and 6% for liberals
- p. 1 Abstract, last sentence
Compared with algorithmic ranking, individuals’ choices played a stronger role in limiting exposure to cross-cutting content.
Only the authors can say what is right here.
Judge's note: Page 4 is in scope. The 15% and 70% figures appear nowhere in the article. Page 2 gives 5% (conservatives) and 8% (liberals) for ranking, and 17% and 6% for clicks. Fig. 3B shows exposed-to-selected drops of roughly 13% and 9% in relative terms, not 70%. The second reviewer note (the abstract's claim that choice outweighs ranking, while liberals show 8% > 6%) is not adopted. In the abstract and discussion, "individuals' choices" plausibly includes the choice of friends, and the discussion names friend-network composition as the most important factor (Fig. 3B: random to potential drops of about 45%→24% for liberals). So that part is explained in the text and is a matter of interpretation.
Second check: another model agreed
Report a problem with this finding
6 minor (typos and the like; most checked by the judge only)
Reference or label Ref. 5 reads "Proc. 21st Int. Conf. World Wide Web Pages 1201.4145 (2012)". "1201.4145" is not a page range.
- p. 3 References, ref. 5
World Wide Web Pages 1201.4145 (2012)
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: that 1201.4145 is the arXiv identifier of Bakshy et al.'s WWW 2012 paper. On the page 3 image, ref. 5 shows "1201.4145" set in bold where a volume or page range would normally go. Read as printed it is not a valid page range, so the citation is malformed whatever its source.
Second check: another model agreed
Report a problem with this finding
Reference or label Ref. 4 reads "S. Messing, S. J. Westwood, Communic. Res. 41, 1042–1063 (2012)". Volume 41 with these pages appeared in 2014; 2012 is the online-first year.
- p. 3 References, ref. 4
Communic. Res. 41, 1042–1063 (2012).
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: that Communication Research volume 41, pp. 1042–1063, is dated 2014. The page 3 image confirms the printed text. Citing the online-first year is a common practice, so this is minor.
Second check: another model agreed
Report a problem with this finding
Wording Discussion: "on average more than 20% of an individual’s Facebook friends who report an ideological affiliation are from the opposing party". The only reported statistics are medians of 0.20 (liberals) and 0.18 (conservatives).
- p. 2 Discussion, para 2
than 20% of an individual’s Facebook friends
- p. 1 Results, homophily paragraph
ships that conservatives maintain with liberals is 0.18
Only the authors can say what is right here.
Judge's note: Page 1 reports median proportions of 0.20 (IQR [0.09, 0.36]) and 0.18 (IQR [0.09, 0.30]); Fig. 2 shows medians and percentile ranges, not means. No mean is reported in the article. Means of right-skewed proportions could exceed 20%, so this may be true but cannot be checked from the files. The supplement (not provided) might report the mean.
Second check: another model agreed
Report a problem with this finding
Wording "The risk ratio comparing the probability of seeing cross-cutting content relative to ideologically consistent content is 5% for conservatives and 8% for liberals" (likewise "risk ratio ... to be 17% for conservatives and 6% for liberals"). The values read as percentage reductions (1 − RR), not as risk ratios.
- p. 2 Main text, col 2
The risk ratio comparing the probability of seeing cross-cutting content relative to ideologically consistent content is 5% for conservatives and 8% for liberals
- p. 2 Main text, col 3
we estimated the risk ratio comparing the likelihood that an individual clicks on a cross-cutting content relative to a consistent content to be 17% for conservatives
- p. 2 Results, para 1
tive to ideologically consistent content is 5% for
- p. 2 Results, para 2
a consistent content to be 17% for conservatives
Only the authors can say what is right here.
Judge's note: On page 2, columns 2–3, the text calls these values risk ratios but says the result means "slightly less" cross-cutting content. Fig. 3B shows small relative drops (conservatives about 34.5% potential to about 33.5% exposed), which fits RR ≈ 0.95, not RR = 0.05. The wording therefore misdescribes the quantity. The supplement (section S1.7), where it may be defined, is not among the files, so the authors should confirm the intended wording.
Second check: another model agreed
Report a problem with this finding
Typo Alignment scores are formatted inconsistently within one sentence: "(As = +.80)" has no leading zero, while "(As = –0.65)" has one.
- p. 1 p1130, right column
As = +.80
Judge's note: The page 1 image, right column, shows "+.80" and "–0.65" in the same sentence. This is a minor formatting inconsistency only.
Second check: none (judge only)
Report a problem with this finding
Typo Misspelling: "Individual choice futher limits exposure" should read "further".
- p. 2 Main text, col 2, start of paragraph
Individual choice futher limits exposure
- p. 2 Results, para 2, first sentence
Individual choice futher limits exposure to
Judge's note: The page 2 image, column 2, shows "futher" as printed. It is not an extraction artefact.
Second check: none (judge only)