News Consumption and Partisan Attention
Mapping the Terrain: Sources, Stories, and the Limits of Domain-Level Measurement
A consistent thread running through this collection is dissatisfaction with how scholars have historically measured partisan news consumption. Gaisbauer2025-by frames the problem programmatically: existing work oversimplifies political leaning, fixates on the outlet level, and overweights a narrow set of national contexts. Green2025-ap makes the empirical case for moving beyond domains, showing that within-outlet story-level heterogeneity is substantial — especially for moderate-scoring sources like the Wall Street Journal, whose audience score conceals starkly divergent partisan publics for different articles. Goel2025-iq extends this critique to misinformation research: domain-list approaches miss the prevalent phenomenon of users repurposing factually accurate mainstream content to advance misleading narratives, with co-shared mainstream articles reaching roughly twice the audience of fake news. Hartmann2025-px generalizes the methodological lesson across 129 echo chamber studies, arguing that much of the field’s empirical disagreement is an artifact of conceptual and operational heterogeneity rather than a feature of the world. Together, these papers reframe partisan attention as a phenomenon that lives at the intersection of story, source, and user — and warn that infrastructural choices in measurement materially shape conclusions. Ulloa2024-jm reinforces this at the data-collection layer: ex-situ scraping introduces systematic, category-biased distortion (~34%) that swamps the time-delay effects prior work has emphasized.
Partisan Insularity and the Multi-Party Media Partisanship Attention Score
Giglietto2019-882f1900 develops the Multi-Party Media Partisanship Attention Score logic in the 2018 Italian election, showing that populist-aligned sources (M5S, the League) attract more insular Twitter audiences and that insularity correlates with a distinctive Facebook engagement signature — more shares per comment for insular sources, more comments per share for cross-partisan ones. Rossi2023-847d5a9f carries the European newsfeed agenda forward with Meta’s URL Shares Dataset, finding that untrustworthy URL shares rose in Germany and Italy during election years while view-level exposure stayed comparatively flat, with older cohorts driving most untrustworthy sharing. Giglietto2026-632ef967 then closes the loop between sharing and viewership at scale: highly partisan audiences impose a measurable reach penalty on Facebook (≈2.3M fewer views per SD of partisan alignment), journalistic quality earns a reach reward, and both effects spike around the 2020 election — evidence that the platform is an active curator whose calibrations track political crises.
Algorithmic Curation as Strategic Intervention
A second cluster reframes algorithms not as inscrutable black boxes but as governable sociotechnical systems. Bastos2025-ya, McNally2025-dn, and Bastos2025-ol — three closely linked analyses of Guardian engagement against 52 documented News Feed updates — converge on the finding that hard news is differentially treated, with Granger-causal effects detectable at a 19–24 day lag consistent with phased software rollouts. They argue this demonstrates the feasibility of independent auditing under DSA Article 40(4). Gauthier2026-iq complements this from the X side with a field experiment showing the algorithmic “For You” feed shifts users rightward on policy, Trump investigations, and Ukraine, and — crucially — that effects persist via algorithm-induced following choices even after the algorithm is switched off, helping reconcile the null results from Meta’s 2020 collaboration. Efstratiou2025-gs complicates the partisan-amplification story: right-leaning visibility gains on post-acquisition Twitter are largely mediated by agitating content and proximity to Elon Musk himself, suggesting platform-owner attention and engagement-bait are the underlying mechanisms rather than ideology per se.
Cross-Partisan Exposure, Echo Chambers, and the Persistence Question
The collection’s stance on echo chambers is notably ambivalent. Bakshy2015-rn remains the foundational counterweight: Facebook networks are homophilous but contain meaningful cross-cutting ties, and individual click choices suppress diverse exposure more than the News Feed algorithm does. Green2025-ap reinterprets this through “curation bubbles” — networked unbundling that makes consumption more partisan than domain measures suggest, though framed as democratic participation rather than pathology. Hartmann2025-px warns that whether a study “finds” an echo chamber depends largely on how it defines one. Gonzalez-Bailon2024-rq, analyzing roughly a billion 2020 election reshares, locates misinformation’s distinctive virality in a small set of highly amplifying actors rather than in diffuse exposure patterns. Mosleh2024-op then pushes against the field’s Twitter-Facebook duopoly, showing engagement with partisan and low-quality news diverges meaningfully across seven platforms — a reminder that “social media” is not a single object.
Affect, Reactions, and Populist Mobilization
Several papers center the emotional architecture of platform engagement. Anwar2024-34dba628 synthesizes 64 studies of Facebook Reactions, establishing that sociopolitical content draws a broader, more negative emotional spectrum and that “Angry” is a reliable driver of engagement with populist content. Kalsnes2025-zb operationalizes this in the Scandinavian newsfeed: populist parties (Sweden Democrats, Progress Party, Hard Line) attract disproportionate Angry reactions, and Angry reactions positively predict shares while Love reactions suppress them. Giglietto2019-882f1900’s share/comment asymmetry adds a structural interpretation: affective amplification through sharing concentrates among insular populist publics, while contestation flows through commenting on cross-partisan content.
Transnational Right-Wing Ecosystems and Epistemic Authority
A final arc connects partisan attention to the production side. Copland2025-em traces how Sky News Australia became a node in Benkler-style network propaganda, deploying parochial Facebook content domestically while pushing US-centric YouTube videos into the American right-wing pipeline. Rodarte2026-dk examines the inverse — how Brazilian parliamentarians during the Manaus crisis bypassed or mimicked journalism to claim epistemic authority directly, with opposition deputies engaging in “derivative reporting” that borrows journalistic credibility while pro-Bolsonaro deputies severed ties with mainstream media altogether. Ghezzi2023-8bebc91f documents the same partisan sorting in the more rarefied register of scientific expertise: UK newspapers cite Great Barrington Declaration vs. John Snow Memorandum signatories along left-right lines, with the segregation reproduced in academic co-authorship and Twitter sharing. Across these cases, partisan attention is not merely a downstream consumption pattern but a contested process in which political actors, platforms, and sources jointly construct the boundary between credible and discreditable information.