Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

The Dissemination of Scientific Fake News : On the Ranking of Retracted Articles in Google

Genot, Emmanuel LU and Olsson, Erik J LU (2021) p.228-242
Abstract
Fake news can originate from an ordinary person carelessly posting what turns out to be false information orfrom the intentional actions of fake news factory workers,but broadly speaking it can also originate from scientific fraud. In the latter case, the article can be retracted upon discovery of the fraud. A case study shows, however, that such fake sciencecan be visible in Google even after the article was retracted, in fact more visible thanthe retraction notice. We hypothesize that the reason for this lies in the popularity-based logic governing Google, in particular its foundational PageRank algorithm,in conjunction with a psychological law which we refer to as the “law of retraction”: a retraction notice is typically taken to be... (More)
Fake news can originate from an ordinary person carelessly posting what turns out to be false information orfrom the intentional actions of fake news factory workers,but broadly speaking it can also originate from scientific fraud. In the latter case, the article can be retracted upon discovery of the fraud. A case study shows, however, that such fake sciencecan be visible in Google even after the article was retracted, in fact more visible thanthe retraction notice. We hypothesize that the reason for this lies in the popularity-based logic governing Google, in particular its foundational PageRank algorithm,in conjunction with a psychological law which we refer to as the “law of retraction”: a retraction notice is typically taken to be less interestingand therefore less popular with internet users than the original content retracted. We conduct anempiricalstudy drawing on records of articles retracted due to fraud (fabrication of data) in the Retraction Watch public database. The study tests the extent to which such retracted scientific articles are still highly ranked in Google –and more so than information about the retraction. We find, among other things, thatboth Google Search and Google Scholar more often than not rankeda link to the original article higher than a link indicating that the article has been retracted.Surprisingly, Google Scholar did not perform better in this regard than Google Search.We also foundcases in which Google didnot track the retraction of anarticle on the first result page at all.We conclude thatboth Google Search and Google Scholar runthe risk of disseminating fake science through theirranking algorithms. (Less)
Please use this url to cite or link to this publication:
author
and
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
keywords
fake news, retraction, “law of retraction”
host publication
The Epistemology of Fake News
editor
Bernecker, Sven ; Flowerree, Amy K. and Grundmann, Thomas
pages
228 - 242
publisher
Oxford University Press
external identifiers
  • scopus:85113545874
ISBN
9780198863977
DOI
10.1093/oso/9780198863977.003.0011
project
Filterbubblor och ideologisk segregering online: behövs reglering av sökmaskiner?
language
English
LU publication?
yes
id
c395d114-c3b2-486a-80a0-0c80c150adee
date added to LUP
2020-02-19 15:26:49
date last changed
2023-05-22 14:05:53
@inbook{c395d114-c3b2-486a-80a0-0c80c150adee,
  abstract     = {{Fake news can originate from an ordinary person carelessly posting what turns out to be false information orfrom the intentional actions of fake news factory workers,but broadly speaking it can also originate from scientific fraud. In the latter case, the article can be retracted upon discovery of the fraud. A case study shows, however, that such fake sciencecan be visible in Google even after the article was retracted, in fact more visible thanthe retraction notice. We hypothesize that the reason for this lies in the popularity-based logic governing Google, in particular its foundational PageRank algorithm,in conjunction with a psychological law which we refer to as the “law of retraction”: a retraction notice is typically taken to be less interestingand therefore less popular with internet users than the original content retracted. We conduct anempiricalstudy drawing on records of articles retracted due to fraud (fabrication of data) in the Retraction Watch public database. The study tests the extent to which such retracted scientific articles are still highly ranked in Google –and more so than information about the retraction. We find, among other things, thatboth Google Search and Google Scholar more often than not rankeda link to the original article higher than a link indicating that the article has been retracted.Surprisingly, Google Scholar did not perform better in this regard than Google Search.We also foundcases in which Google didnot track the retraction of anarticle on the first result page at all.We conclude thatboth Google Search and Google Scholar runthe risk of disseminating fake science through theirranking algorithms.}},
  author       = {{Genot, Emmanuel and Olsson, Erik J}},
  booktitle    = {{The Epistemology of Fake News}},
  editor       = {{Bernecker, Sven and Flowerree, Amy K. and Grundmann, Thomas}},
  isbn         = {{9780198863977}},
  keywords     = {{fake news; retraction; “law of retraction”}},
  language     = {{eng}},
  pages        = {{228--242}},
  publisher    = {{Oxford University Press}},
  title        = {{The Dissemination of Scientific Fake News : On the Ranking of Retracted Articles in Google}},
  url          = {{https://lup.lub.lu.se/search/files/94696954/Dissemination_of_Scientific_Fake_News.pdf}},
  doi          = {{10.1093/oso/9780198863977.003.0011}},
  year         = {{2021}},
}