CoChat
Sign Up
Cited by 204Complex Network Analysis TechniquesRecommender Systems and TechniquesOpinion Dynamics and Social Influence

The DOI matches its Crossref record; 31 of the 32 references it cites were found in a registry or an index.

How algorithmic popularity bias hinders or promotes quality

Open access
Ciampaglia, Giovanni Luca, Nematzadeh, Azadeh, Menczer, Filippo, Flammini, Alessandro·2018·Scientific Reports
DOI 10.1038/s41598-018-34203-2PMC PMC6206065
Open this paper in CoChat — free accountOpen PDF

Abstract

Algorithms that favor popular items are used to help us select among many choices, from top-ranked search engine results to highly-cited scientific papers. The goal of these algorithms is to identify high-quality items such as reliable news, credible information sources, and important discoveries–in short, high-quality content should rank at the top. Prior work has shown that choosing what is popular may amplify random fluctuations and lead to sub-optimal rankings. Nonetheless, it is often assumed that recommending what is popular will help high-quality content “bubble up” in practice. Here we identify the conditions in which popularity may be a viable proxy for quality content by studying a simple model of a cultural market endowed with an intrinsic notion of quality. A parameter representing the cognitive cost of exploration controls the trade-off between quality and popularity. Below and above a critical exploration cost, popularity bias is more likely to hinder quality. But we find a narrow intermediate regime of user attention where an optimal balance exists: choosing what is popular can help promote high-quality items to the top. These findings clarify the effects of algorithmic popularity bias on quality outcomes, and may inform the design of more principled mechanisms for techno-social cultural markets.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41598-018-34203-2

Cites

31 of 31 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Predicted average quality at equilibrium. Heatmap of average quality q¯ as a function of α and β, based on numerical solution of equilibrium condition derived in the text.
Figure 1 Predicted average quality at equilibrium. Heatmap of average quality q¯ as a function of α and β, based on numerical solution of equilibrium condition derived in the text.CC BY · PubMed Central
Figure 2 Effects of popularity bias on average quality and faithfulness from model simulations. (a) Heatmap of average quality q¯ as a function of α and β, showing that popularity bias can either hinder or promote average quality depending on exploration cost. (b) The location of the maximum q¯ as a function of β depends on α, here shown for α = 0, 1, 2. When α = 1 an intermediate amount of popularity bias is optimal. (c) Faithfulness τ of the algorithm as a function of α and β. (d) τ as a function of β for the same three values of α. Standard errors are shown in panels (b,d) and are smaller than the markers if not visible.
Figure 2 Effects of popularity bias on average quality and faithfulness from model simulations. (a) Heatmap of average quality q¯ as a function of α and β, showing that popularity bias can either hinder or promote average quality depending on exploration cost. (b) The location of the maximum q¯ as a function of β depends on α, here shown for α = 0, 1, 2. When α = 1 an intermediate amount of popularity bias is optimal. (c) Faithfulness τ of the algorithm as a function of α and β. (d) τ as a function of β for the same three values of α. Standard errors are shown in panels (b,d) and are smaller than the markers if not visible.CC BY · PubMed Central
Figure 3 Temporal evolution of average quality. Average quality q¯ is traced over time for β = 0.6 and different values of exploration cost. Error bars represent standard errors across runs. Compared to the optimal case α = 1, with more exploration (α = 0) the popularity bias just adds noise, and with less exploration (α = 2) it causes the system to converge prematurely to sub-optimal quality.
Figure 3 Temporal evolution of average quality. Average quality q¯ is traced over time for β = 0.6 and different values of exploration cost. Error bars represent standard errors across runs. Compared to the optimal case α = 1, with more exploration (α = 0) the popularity bias just adds noise, and with less exploration (α = 2) it causes the system to converge prematurely to sub-optimal quality.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

1 retracted

Research works printed but not linked: 5

  1. Sunstein, C. R. Republic.com 2.0. Princeton university press (2009).
  2. Pariser, E. The filter bubble: How the new personalized Web is changing what we read and how we think. Penguin (2011).
  3. Adler, M. Stardom and talent. American economic review 75(1), 208–12 (1985).
  4. Dawkins, R. The selfish gene. Oxford university press (1989).
  5. Bell, E J., Owen, T., Brown, P D., Hauka, C. & Rashidian, N. The platform press: how Silicon Valley reengineered journalism. Technical report, Tow Center for Digital Journalism (2017).

Cited by

Sorted by citations

See all 204

Related

See all

Metadata

Authors
Ciampaglia, G. L., Nematzadeh, A., Menczer, F., & Flammini, A.
Affiliations
Indiana University Bloomington (US)
Corresponding author
Ciampaglia, Giovanni Luca
ORCID
Venue
Scientific Reports
Publisher
Springer Science and Business Media LLC
Published
2018-10-29
Volume
8
Issue
1
Pages
15951-15951
ISSN
2045-2322 (ISSN-L)
DOI
10.1038/s41598-018-34203-2
PMC
PMC6206065
OpenAlex
W2732691751
Cited by
204
References
32
Citations by year
2026: 13 · 2025: 23 · 2024: 41 · 2023: 25 · 2022: 25 · 2021: 33 · 2020: 27 · 2019: 8 · 2018: 5
FWCI
9.32
Type
journal article
Language
en
Open access
Licence
Locations
  • Scientific Reports (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, other-oa)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • IUScholarWorks Open (Indiana University) (submittedVersion)
  • arXiv (Cornell University)
Funders
  • James S. McDonnell Foundation (220020274)
  • National Science Foundation (CCF-1101743)
Topics
  • Physical Sciences › Physics and Astronomy › Statistical and Nonlinear Physics › Complex Network Analysis Techniques
  • Physical Sciences › Computer Science › Information Systems › Recommender Systems and Techniques
  • Physical Sciences › Physics and Astronomy › Statistical and Nonlinear Physics › Opinion Dynamics and Social Influence
Keywords
popularity bias, algorithmic ranking, cultural market, exploration costs, user attention, recommendation algorithms, information quality, search engine ranking, citation ranking
MeSH terms
Algorithms, Quality Control, Social Media

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review