CoChat
Sign Up
Cited by 0Opinion Dynamics and Social InfluenceComputational and Text Analysis MethodsSocial Media and Politics

Computational Social Science of Social Cohesion and Polarization

Open access
Keijzer, Marijn A., Lorenz, Jan, Bojanowski, Michał·2026·Computational Social Sciences
DOI 10.1007/978-3-032-01373-6_1
Open this paper in CoChat — free accountOpen PDF

Abstract

The emergence of Computational Social Science (CSS) marks a transformative shift in the study of social phenomena, triggered by advancements in computational technology and large-scale collection of (digital) behavioral trace data. CSS promises to provide an accurate and deep understanding of social behavior through the analysis of social interactions, movement patterns, and communication networks at scale. However, the lack of a common theoretical language and clearly defined scientific agenda complicates CSS’ ability to build a common understanding of social phenomena. In this chapter, we establish CSS as a scientific paradigm, emphasizing the integration of computational tools with rigorous theoretical modeling. Specifically, we position CSS within the trajectory of a maturing science and emphasize the synthesis of data-driven CSS with theoretically-informed modeling for understanding social phenomena like social cohesion and polarization. To illustrate how CSS can advance these fields specifically, we unpack a range of conceptualizations of political polarization and social cohesion, and demonstrate how they have evolved in Europe over recent decades. Finally, we discuss how the subsequent chapters in this book aim to bridge the gap between theoretical and empirical CSS for social cohesion and polarization.

AI snapshot

Written from the paper’s full text.

TL;DR

This introductory chapter argues that Computational Social Science (CSS) must move beyond descriptive, data-driven analysis toward a paradigm that tightly integrates formal theory with computational and empirical work, using social cohesion and polarization as the illustrative domain. It reviews competing conceptualizations of both phenomena and presents European trend analyses showing rising issue alignment but no uniform rise in affective polarization.

Key findings

  • Issue alignment in Europe has increased: using PCA on five ESS attitude items (left-right self-identification, European unification, homosexuality, immigration, redistribution) for 212,387 individuals across 15 countries over nine years, the variance explained by the first principal component rose from 29.2% to 35.1% on average between 2004 and 2023, with the trend holding for almost all 15 countries.
  • The structure of beliefs was strikingly stable: PC loadings from 2004 and 2023 were highly similar, and similar when PCAs were computed independently per country-year.
  • Affective polarization trends diverge by region: using CSES data (waves 1–5, 25 years), the stark increase seen in North America does not hold in Europe, where affective polarization scores have on average decreased over the last 30 years.
  • Social cohesion (ESS waves 2–11, 15 European nation-states, ~20 years) shows mostly stable or slightly upward trends in both interpersonal and institutional trust, with a notable dip in interpersonal trust in 2020; institutional trust was modeled as a common factor across five items with average loading .78 (range .58–.89) and 62% explained variance.

Methods

  • Conceptual/theoretical review chapter: defines CSS as "a scientific approach to understanding social phenomena using computational tools," positions it within a three-component model of modern science (theoretical, empirical, computational), and reviews conceptualizations of social cohesion (Durkheimian solidarity, social capital, Schiefer and Noll's six dimensions, Green and Janmaat's four cohesion regimes) and of polarization (issue-based, attitude consolidation, affective).
  • Empirical trend analyses: PCA on ESS attitude items to measure issue alignment; affective polarization scores computed from CSES as the difference between feelings toward one's own party's voters and the vote-share-weighted average of feelings toward other parties' voters; confirmatory factor analysis for institutional trust.
  • The chapter also previews the book's eleven chapters, grouped by method: social network analysis (Chaps. 3–5), NLP (Chaps. 6–8), and agent-based modeling (Chaps. 9–11).

Limitations

  • Stated: an exhaustive review of social cohesion conceptualizations is "beyond the scope of this introduction"; globally there is a lack of research on polarization in general and affective polarization in particular; affective polarization is harder to measure in multiparty systems as in many European countries; and the authors caution that issue alignment does not itself mean polarization—attitudes may align while remaining consensual, spread out, or bipolarized.
  • Evident from the text: the trend analyses are descriptive and cover only 15 European countries (plus North America for the CSES comparison), with the authors noting considerable variance across countries; the chapter is an introduction to an edited volume rather than a standalone empirical study, and its empirical results are presented as illustrative "to pave the way for the remaining chapters."

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1007/978-3-032-01373-6_1

Cites

72 of 72 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

14 of 25 cited sources checked · 8 supported · 0 flagged · accuracy not yet measured

  • Premisecitation

    Predictive models tend to be 'black boxes' that do not contribute to a deeper understanding of social phenomena at the level of the explanation.

    “As a consequence, predictive models tend to be 'black boxes' that do not contribute to a deeper understanding of social phenomena at the level of the explanation (Molina and Garip 2019).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Molina and Garip 2019 source

      The “Classical Statistics Versus Machine Learning” passage says predictive modeling can produce black-box results that offer little insight into mechanisms. But the “Predictions for Policy and Theory Development” section also describes predictions as a starting point for new insights and theory, so the claim overstates the paper by saying predictive models do not contribute to explanation. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
  • Premisecitation

    To move from description to explanation, we cannot rely on empirical observation alone, but have to understand the dynamics of a system that led to the particular observed outcome.

    “To move from description to explanation, however, we cannot rely on empirical observation alone, but have to understand the dynamics of a system that led to the particular observed outcome (Keuschnigg et al. 2017;Flache et al. 2022).”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Keuschnigg et al. 2017

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Flache et al. 2022 source
  • Premisecitation

    Complex data analysis has become feasible to the point that our predictive models frequently outperform theoretical ones.

    “First, complex data analysis has become feasible to the point that our predictive models frequently outperform theoretical ones (Salganik 2020).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Salganik 2020 source

      The Results describe teams using varied data-processing and statistical-learning techniques, supporting the feasibility of complex predictive analysis. The Discussion does not compare those models with theoretical models or establish that they frequently outperform them; instead, it says a simple benchmark outperformed many submissions and was only slightly worse than the best. (LLM, conf=0.83)

      Experimental — accuracy not yet measured
  • Premisecitation

    Digital trace data-from social media posts to behavioral logs-provide an empirical foundation for tracking the evolution of social alignments at scale.

    “Digital trace data-from social media posts to behavioral logs-provide an empirical foundation for tracking the evolution of social alignments at scale (Lazer et al. 2009).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Lazer et al. 2009 source
  • Interpretationcitation

    The variance across countries is partly explained by the social cohesion regimes by Green and Janmaat (2011).

    “At face value, that variance is partly explained by the social cohesion regimes by Green and Janmaat (2011), discussed in Sect.1.2.1.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Green and Janmaat (2011) source
  • Classified as probably not a substantial claim. 32
  • May combine two claims. 12

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Theory-driven agent-based modeling emphasize the generative role of formal models in hypothesis development, allowing researchers to explore how micro-level assumptions-such as bounded confidence or social influence-can produce unexpected macro-level dynamics like polarization.

    “Theory-driven agent-based modeling emphasize the generative role of formal models in hypothesis development (Flache et al. 2022), allowing researchers to explore how micro-level assumptions-such as bounded confidence or social influence-can produce unexpected macro-level dynamics like polarization (e.g., see Chap.empirical data to calibrate agent behavior (e.g., see Chap.9: Edmonds et al. 2025) or”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Flache et al. 2022 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    In quantitative sociology, CSS has had a major impact across the board, overtaking fields like social network analysis and group formation, collective behavior and political sociology, the sociology of knowledge, cultural sociology, social psychology, and emotions, the production of culture, economic sociology and organizations, and demography and population studies.

    “In quantitative sociology, CSS has had a major impact across the board, overtaking fields like social network analysis and group formation, collective behavior and political sociology, the sociology of knowledge, cultural sociology, social psychology, and emotions, the production of culture, economic sociology and organizations, and demography and population studies (Edelmann et al. 2020).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Edelmann et al. 2020 source

      The Introduction says CSS is spreading rapidly and identifies the seven listed areas as subfields with at least five qualifying studies. It does not say CSS is 'overtaking' those fields or has had an impact 'across the board'; the Conclusion says it has not yet been broadly integrated into many other large sociology subfields. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

  • Classified as probably not a substantial claim.

    Premisecitation

    The emergence of Computational Social Science (CSS) marks a transformative shift in the study of social phenomena, triggered by advancements in computational technology and large-scale collection of digital behavioral trace data.

    “The emergence of Computational Social Science (CSS) marks a transformative shift in the study of social phenomena, triggered by advancements in computational technology and large-scale collection of (digital) behavioral trace data.”
    Quote in the paperSubstantial claim2 of 3 sources checked
    • cites Lazer et al. 2009 source

      In “Life in the network” and “From individuals to societies” (page not provided), the paper links emerging computational social science to new technologies and large-scale digital records of behavior. It describes a transformation in understanding as a potential outcome, however, and says barriers may limit progress; the claim presents the shift more definitively than the source does. (LLM, conf=0.82)

      Experimental — accuracy not yet measured
    • cites Conte et al. 2012 source
    • cites Lazer et al. 2020 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    The new millennium has witnessed the Computational Social Science (CSS) revolution.

    “The new millennium has witnessed the Computational Social Science (CSS) revolution (Lazer et al. 2009;Conte et al. 2012;Lazer et al. 2020).”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites Lazer et al. 2009 source

      In “The coming age of computational social science,” the paper says computational social science is emerging and occurring on a large scale in some organizations. But it describes its emergence in academic social science as slow and much of its transformative potential as prospective, rather than saying a CSS revolution had already occurred. (LLM, conf=0.82)

      Experimental — accuracy not yet measured
    • cites Conte et al. 2012 source
    • cites Lazer et al. 2020 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Former results showed differences in profiles and level of social cohesion between countries, but rather small changes over time.

    “former results showed differences in profiles and level of social cohesion between countries, but rather small changes over time (Dragolov et al. 2016;Delhey et al. 2018).”
    Quote in the paperNot a substantial claim1 of 2 sources checked
    • cites Dragolov et al. 2016 source
    • cites Delhey et al. 2018 source

      The figure captions in the indexed copy (listed under References; no page available) rank countries by overall social-cohesion score and group them into levels, supporting the claim of differences between countries. They do not establish differences in cohesion profiles or that changes over time were small; searches for temporal stability and trends found no matching passage. (LLM, conf=0.78)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingcitation

    The Nordic regimes plus Switzerland and The Netherlands score on the higher side of both trust scales, whereas the Continental regimes in central and southern Europe group together in the middle and lower end of the trust distributions.

    “The Nordic regimes plus Switzerland and The Netherlands score on the higher side of both scales, whereas the Continental regimes in central and southern Europe group together in the middle and lower end of the trust distributions.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    PC1 has a negative loading for being politically right and a positive loading for all other attitude dimensions.

    “PC1 has a negative loading for being politically right and a positive loading for all other attitude dimensions.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    Data-driven CSS appears to have taken an inductive approach; learning from real-world observations alone.

    “It may be implicit, but data-driven CSS appears to have taken an inductive approach; learning from real-world observations alone.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisenone given

    Digitization in many parts of society, including the emergence of the Internet as a public sphere for many social interactions, has created enormous amounts of behavioral data.

    “Concurrently, digitization in many parts of society, including the emergence of the Internet as a public sphere for many social interactions, has created enormous amounts of behavioral data.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

    Interpretationreasoning

    As a consequence, a significant amount of work has been descriptive in nature, or has tested rudimentary ideas that are perhaps more developed in the longer theoretical literature of the disciplinary social sciences.

    “As a consequence, a significant amount of work has been descriptive in nature, or has tested rudimentary ideas that are perhaps more developed in the longer theoretical literature of the disciplinary social sciences.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given

    As a scientific discipline, CSS is concerned with the production of generalizable knowledge about social phenomena.

    “As a scientific discipline, CSS is concerned with the production of generalizable knowledge about social phenomena.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    None of that exists in CSS, which has yet to develop its proprietary language, methods, and paradigms.

    “None of that exists in CSS, which has yet to develop its proprietary language, methods, and paradigms.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    This makes CSS almost exclusively a quantitative science.

    “This makes CSS almost exclusively a quantitative science.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationreasoning

    Since CSS researchers draw from a variety of disciplines, there is a lack of common understanding about what has and ought to be done.

    “Since CSS researchers draw from a variety of disciplines, there is a lack of common understanding about what has and ought to be done.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Natural language processing (NLP) enables the extraction of ideological positions, sentiment, and framing from large-scale textual data, offering a lens through which to study discourse and alignment.

    “Natural language processing (NLP), which builds on traditional text analysis and increasingly leverages transformer-based models, enables the extraction of ideological positions, sentiment, and framing from large-scale textual data, offering a lens through which to study discourse and alignment (e.g., see Chaps.6-8: Brockmann et al. 2025; Lee et al. 2025; Yuan 2025).”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Brockmann et al. 2025

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Lee et al. 2025

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Yuan 2025

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given

    Simulation models have the advantage over empirical models in that they are better at making generalizations, cutting out what is less important, and focus on mechanisms in isolation.

    “Simulation models have the advantage over empirical models in that they are better at making generalizations, cutting out what is less important, and focus on mechanisms in isolation.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given

    Social network modeling is typically computationally expensive, and hence benefited greatly from the computational shift.

    “Social network modeling is typically computationally expensive, and hence benefited greatly from the computational shift.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given

    What sets computational sciences apart from fields without an explicit computational component is that they use cutting-edge computational technologies to solve problems.

    “At the end of the day, what sets computational sciences apart from fields without an explicit computational component is that they use cutting-edge computational technologies to solve problems.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    CSS has primarily focused on analyzing pre-existing, digital, behavioral trace data.

    “CSS has primarily focused on analyzing pre-existing, digital, behavioral trace data.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationnone given

    The lack of a common theoretical language and clearly defined scientific agenda complicates CSS' ability to build a common understanding of social phenomena.

    “However, the lack of a common theoretical language and clearly defined scientific agenda complicates CSS' ability to build a common understanding of social phenomena.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    With the addition of a computational component, two new scientific activities become part of CSS' practice.

    “With the addition of a computational component, two new scientific activities become part of CSS' practice.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationnone given

    Various disciplines have integrated CSS into the production of theory and generalizable knowledge to different degrees.

    “Various disciplines have integrated CSS into the production of theory and generalizable knowledge to different degrees.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Social cohesion is under pressure, and this pressure is felt strongly in the political domain, manifesting as polarization.

    “Social cohesion is under pressure (see Chap. 2: Boehnke et al. 2025), and this pressure is felt strongly in the political domain, manifesting as polarization.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Boehnke et al. 2025

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    Social cohesion appears to have decayed in recent years.

    “Yet, social cohesion appears to have decayed in recent years.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisereasoning

    NLP methods enable the extraction of patterns, trends, and structures in language use across diverse contexts.

    “These methods enable the extraction of patterns, trends, and structures in language use across diverse contexts such as traditional and social media, political discourse, or online communities.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    CSS has enabled us to access and gauge belief segregation using ecologically valid data at an unprecedented scale.

    “CSS has enabled us to access and gauge belief segregation using ecologically valid data at an unprecedented scale.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Both generalized and institutional trust strongly relate to feelings of cooperation, unity, and identification, and are considered prerequisites of economic prosperity and growth.

    “Both concepts strongly relate to feelings of cooperation, unity, and identification, and are considered prerequisites of economic prosperity and growth (Schiefer and Noll 2017).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Schiefer and Noll 2017 source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisereasoning

    By leveraging NLP, researchers can gain insights into public opinion, collective identity, polarization, and the diffusion of ideas, often at scales and levels of granularity that would be impossible with manual analysis alone.

    “By leveraging NLP, researchers can gain insights into public opinion, collective identity, polarization, and the diffusion of ideas, often at scales and levels of granularity that would be impossible with manual analysis alone.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationreasoning

    Conceptual clarity is not social cohesion's strong suit.

    “As signaled above, conceptual clarity is not social cohesion's strong suit.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Social cohesion and polarization are two sides of the same coin, covering some degree of social or political distance between individuals.

    “It is there that we can see how social cohesion and polarization are two sides of the same coin, covering some degree of social or political distance between individuals.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationreasoning

    The more a single component can explain the shape of the data, the more individuals use the same mental model to structure their beliefs.

    “The more a single component can explain the shape of the data, the more individuals use the same mental model to structure their beliefs.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given

    Political polarization is less ambiguous than social cohesion, but still, there is considerable variance in what is typically understood when discussed in news media, by politicians, and across social sciences.

    “Political polarization is less ambiguous than social cohesion, but still, there is considerable variance in what is typically understood when discussed in news media, by politicians, and across social sciences.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given

    Both social cohesion and polarization are examples of elusive concepts contested in the public debate, discussed across (social) sciences, and are therefore difficult to define.

    “Both social cohesion and polarization are examples of elusive concepts contested in the public debate, discussed across (social) sciences, and are therefore difficult to define.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Social cohesion has been described as a quasi-concept, a construct that politicians like because it sounds like a scientifically solid societal goal, yet remains vague enough to be adapted for different interpretations in different situations.

    “Social cohesion has been described as a quasi-concept (Bernard 1999), a construct that politicians like because it sounds like a scientifically solid societal goal, yet remains vague enough to be adapted for different interpretations in different situations.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Bernard 1999

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisereasoning

    Belief network analysis (BNA) can capture which beliefs are good predictors of other beliefs, and to what extent we can capture polarization as belief clustering across time, topics, and societies.

    “Much like the method for studying belief alignment described in Sect.1.2.2,BNA can capture which beliefs are good predictors of other beliefs, and to what extent we can capture polarization as belief clustering across time, topics, and societies.”
    Quote in the paperSubstantial claim
35 of 35

  • The paper relies on this without support.

    Social cohesion is a desirable societal goal that should be promoted.

    “Social cohesion is the cornerstone of harmonious democratic societies.”

  • The paper relies on this without support.

    Political polarization is a problem that needs to be addressed.

    “Political polarization, and its consequences for the erosion of social cohesion, is named as one of the five greatest global risks of today by the World Economic Forum and the most severe societal challenge of tomorrow (World Economic Forum 2025).”
2 of 2

Self-check

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

Each contradiction between two passages listed here was confirmed by an independent check.

  • Affective-polarization trend exceeds the stated measurement period

    Minor

    The section says Figure 1.4 covers 25 years of measurement, but then describes a decline over the last 30 years. The stated measurement period cannot establish that 30-year trend.

1 of 1

Cites

Flagged first, then by citations

Research works printed but not linked: 12

  1. C. Anderson, The end of theory: the data deluge makes the scientific method obsolete. Wired Mag. 16(7), 16-07 (2008)
  2. Berger, (ed.), The Limits of Social Cohesion: Conflict and Mediation in Pluralist Societies; A Report of the Bertelsmann Foundation to the Club of Rome (Westview Press, Boulder, 1998)
  3. P. Bourdieu, The Forms of capital. Cult. Theor. Anthol. 1, 81–93 (2011)
  4. P.E. Converse, The nature of belief systems in mass publics, in Ideology and Its Discontents, ed. by Apter (The Free Press of Glencoe, New York, 1964)
  5. G. Esping-Andersen, The Three Worlds of Welfare Capitalism (Polity Press, Cambridge, UK, 1990)
  6. M. Hilbert et al., Computational communication science: a methodological catalyzer for a maturing discipline. Int. J. Commun. 13, 23 (2019)
  7. J. Jenson, Mapp. Soc. Cohes. State Canad. Res. (Family Network, CPRN, Ottawa, 1998)
  8. C. Read, Logic: Deductive and Inductive (Dodo Press, 1898–2007), 392pp
  9. M.J. Salganik, Bit by Bit: Social Research in the Digital Age (Princeton University Press, Princeton, NJ, 2018)
  10. D. Schiefer, J. Noll, The essentials of social cohesion: a literature review. Soc. Indic. Res. 132(2), 579–603 (2017). https://doi.org/10.1007/s11205-016-1314-5 · DOI · Paper page
  11. E. Ullmann-Margalit, The Emergence of Norms (Oxford University Press, Oxford, UK, 1977), p.224
  12. World Economic Forum, The Global Risks Report 2025 (Switzerland, Geneva, 2025)

Authors
Keijzer, M. A., Lorenz, J., & Bojanowski, M.
Affiliations
  • Toulouse School of Economics (FR)
  • Constructor University (DE)
  • Kozminski University (PL)
Corresponding author
Keijzer, Marijn A.
ORCID
Venue
Computational Social Sciences
Publisher
Springer Nature Switzerland
Published
2026
Pages
1-23
ISSN
  • 2509-9574 (ISSN-L)
  • 2509-9582
DOI
10.1007/978-3-032-01373-6_1
OpenAlex
W7128296238
Cited by
0
References
58
FWCI
0.00
Type
book chapter
Language
en
Open access
Licence
Locations
  • Computational social sciences (publishedVersion, cc-by)
  • Toulouse 1 Capitole Publications (Université Toulouse I Capitole) (acceptedVersion)
Funders
  • Agence Nationale de la Recherche (France2030, ANR-22-EXES-0015)
  • Région Occitanie Pyrénées-Méditerranée
  • HORIZON EUROPE Framework Programme
  • European Regional Development Fund
Topics
  • Physical Sciences › Physics and Astronomy › Statistical and Nonlinear Physics › Opinion Dynamics and Social Influence
  • Social Sciences › General Social Sciences › Computational and Text Analysis Methods
  • Social Sciences › Communication › Social Media and Politics
Keywords
computational social science, social cohesion, political polarization, scientific paradigms, theoretical modeling, digital behavioral data, social networks

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.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-13.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-13.

Cited sources are checked by gpt-6-sol.

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