CoChat

Algorithmic bias amplifies opinion fragmentation and polarization: A bounded confidence model

Open access

The DOI matches its Crossref record; 57 of the 57 references it cites were found in a registry or an index.

Abstract

The flow of information reaching us via the online media platforms is optimized not by the information content or relevance but by popularity and proximity to the target. This is typically performed in order to maximise platform usage. As a side effect, this introduces an algorithmic bias that is believed to enhance fragmentation and polarization of the societal debate. To study this phenomenon, we modify the well-known continuous opinion dynamics model of bounded confidence in order to account for the algorithmic bias and investigate its consequences. In the simplest version of the original model the pairs of discussion participants are chosen at random and their opinions get closer to each other if they are within a fixed tolerance level. We modify the selection rule of the discussion partners: there is an enhanced probability to choose individuals whose opinions are already close to each other, thus mimicking the behavior of online media which suggest interaction with similar peers. As a result we observe: a) an increased tendency towards opinion fragmentation, which emerges also in conditions where the original model would predict consensus, b) increased polarisation of opinions and c) a dramatic slowing down of the speed at which the convergence at the asymptotic state is reached, which makes the system highly unstable. Fragmentation and polarization are augmented by a fragmented initial population.

Overview

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

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

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

Integrity signals

Registry

Crossref · 10.1371/journal.pone.0213246

Cites

57 of 57 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig 1 Number of clusters obtained for various ε and γ. For ε > 0.4 the population always converges to 1 cluster, so we omitted the range from the plot, for better visualisation. Values are averaged over 100 runs.
Fig 1 Number of clusters obtained for various ε and γ. For ε > 0.4 the population always converges to 1 cluster, so we omitted the range from the plot, for better visualisation. Values are averaged over 100 runs.CC BY · PubMed Central
Fig 2 Mean opinion distance obtained for various ε and γ. Values are averaged over 100 runs.
Fig 2 Mean opinion distance obtained for various ε and γ. Values are averaged over 100 runs.CC BY · PubMed Central
Fig 3 Time to convergence. Total number of interactions required for convergence normalized by the number of individuals, averaged over 100 runs.
Fig 3 Time to convergence. Total number of interactions required for convergence normalized by the number of individuals, averaged over 100 runs.CC BY · PubMed Central
Fig 4 Time to convergence. Normalized total, non-null difference and active number of interactions required for convergence for ε = 0.4. The reference 35exp(γ3.4) is shown as a visual aid only.
Fig 4 Time to convergence. Normalized total, non-null difference and active number of interactions required for convergence for ε = 0.4. The reference 35exp(γ3.4) is shown as a visual aid only.CC BY · PubMed Central
Fig 5 Evolution of the population of opinions for various γ and ε values. The first row corresponds to the case where ε = 1, the second row corresponds to ε = 0.32 while the last row corresponds to ε = 0.2. In all cases γ ∈ {0, 1, 1.1, 1.6} (left to right).
Fig 5 Evolution of the population of opinions for various γ and ε values. The first row corresponds to the case where ε = 1, the second row corresponds to ε = 0.32 while the last row corresponds to ε = 0.2. In all cases γ ∈ {0, 1, 1.1, 1.6} (left to right).CC BY · PubMed Central
Fig 6 Finite size effects: Number of clusters without algorithmic bias. Effective number of clusters obtained for various ε and N, when γ = 0. Values are averaged over 200, 150, 100, 100 and 100 runs for N ∈ {100, 250, 500, 750, 1000}, respectively. Error bars show one standard deviation from the mean.
Fig 6 Finite size effects: Number of clusters without algorithmic bias. Effective number of clusters obtained for various ε and N, when γ = 0. Values are averaged over 200, 150, 100, 100 and 100 runs for N ∈ {100, 250, 500, 750, 1000}, respectively. Error bars show one standard deviation from the mean.CC BY · PubMed Central
Fig 7 Finite size effects: Number of clusters with algorithmic bias. Effective number of clusters obtained for various ε, γ and N. Values are averaged over 200, 150, 100, 100 and 100 runs for N ∈ {100, 250, 500, 750, 1000}, respectively. Error bars show one standard deviation from the mean.
Fig 7 Finite size effects: Number of clusters with algorithmic bias. Effective number of clusters obtained for various ε, γ and N. Values are averaged over 200, 150, 100, 100 and 100 runs for N ∈ {100, 250, 500, 750, 1000}, respectively. Error bars show one standard deviation from the mean.CC BY · PubMed Central
Fig 8 Finite size effects: Opinion distance with algorithmic bias. Mean opinion distance obtained for various ε, γ and N. Values are averaged over 200, 150, 100, 100 and 100 runs for N ∈ {100, 250, 500, 750, 1000}, respectively. Error bars show one standard deviation from the mean.
Fig 8 Finite size effects: Opinion distance with algorithmic bias. Mean opinion distance obtained for various ε, γ and N. Values are averaged over 200, 150, 100, 100 and 100 runs for N ∈ {100, 250, 500, 750, 1000}, respectively. Error bars show one standard deviation from the mean.CC BY · PubMed Central
Fig 9 Initial condition: Number of clusters. Effect of the initial condition on the effective number of clusters (averages over 100 runs).
Fig 9 Initial condition: Number of clusters. Effect of the initial condition on the effective number of clusters (averages over 100 runs).CC BY · PubMed Central
Fig 10 Initial condition: Opinion distance. Effect of the initial condition on the average opinion distance (averages over 100 runs).
Fig 10 Initial condition: Opinion distance. Effect of the initial condition on the average opinion distance (averages over 100 runs).CC BY · PubMed Central
Fig 11 Initial condition: Time to convergence. Effect of the initial condition on the convergence time measured in number of active interactions (averages over 100 runs).
Fig 11 Initial condition: Time to convergence. Effect of the initial condition on the convergence time measured in number of active interactions (averages over 100 runs).CC BY · PubMed Central

Verification

Claims the paper makes

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

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

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

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

Research works printed but not linked: 2

  1. K Purcell, 2010, Understanding the participatory news consumer, Pew Internet and American Life Project, 1, 19
  2. MW Macy, 2003, Polarization in dynamic networks: A Hopfield model of emergent structure, Dynamic social network modeling and analysis, 162

Cited by

See all 135

Metadata

Authors
Sîrbu, A., Pedreschi, D., Giannotti, F., & Kertész, J.
Affiliations
  • University of Pisa (IT)
  • Istituto di Scienza e Tecnologie dell'Informazione "Alessandro Faedo" (IT)
  • Budapest University of Technology and Economics (HU)
Corresponding author
Sîrbu, Alina
Venue
PLOS ONE
Published
2019-03-05
Volume
14
Issue
3
Pages
e0213246
ISSN
1932-6203 (ISSN-L)
OpenAlex
W2790854217
Cited by
112
References
56
Citations by year
2026: 11 · 2025: 14 · 2024: 19 · 2023: 18 · 2022: 22 · 2021: 13 · 2020: 11 · 2019: 3
FWCI
5.67
Type
journal article
Language
en
Open access
Locations
  • PLoS ONE (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • Scuola Normale Superiore di Pisa (submittedVersion, cc-by)
  • RePEc: Research Papers in Economics (submittedVersion)
  • CINECA IRIS Institutial research information system (University of Pisa) (submittedVersion, cc-by)
  • Archivio istituzionale della ricerca (Alma Mater Studiorum Università di Bologna) (submittedVersion)
  • ISTI Open Portal (submittedVersion, other-oa)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Open Research (Auckland University of Technology) (submittedVersion, cc-by)
  • arXiv (Cornell University)
Funders
H2020 Future and Emerging Technologies (641191)
Topics
  • Physical Sciences › Physics and Astronomy › Statistical and Nonlinear Physics › Opinion Dynamics and Social Influence
  • Social Sciences › Communication › Social Media and Politics
  • Social Sciences › Sociology and Political Science › Media Influence and Politics
Keywords
algorithmic bias, opinion dynamics, bounded confidence model, opinion fragmentation, political polarization, social media, information filtering, convergence rate, consensus formation
MeSH terms
Algorithms, Computer Simulation, Humans, Models, Theoretical, Probability, Public Opinion

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

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.