The Morning the Algorithm Changed
On a gray Tuesday in February 2024, a senior editor at one of the three largest news aggregation platforms in the United States opened her dashboard and noticed something unusual. The source-ranking scores numbers that had always felt abstract, the kind of thing engineers talked about in quarterly reviews but editors mostly ignored had shifted overnight. A regional investigative outlet in Cleveland, which had been hovering in the 34th percentile for source authority, had jumped to the 71st. A wire service that had dominated the top tier for years had slipped twelve points.
She mentioned it to the product manager, who mentioned it to the data team, who mentioned it to the researcher in Munich who had built the scoring model. That researcher was Dr. Elena Voss, and she was not surprised at all. She had been expecting that particular shift for eighteen months.
"The signals were there," Voss told me when we spoke at her office at the Ludwig Maximilian University of Munich in March 2026. "The Cleveland outlet had been building citation networks in investigative journalism for three years. They had stopped chasing viral stories and started citing their own sources with more precision. The algorithm was working exactly as designed. It just took the industry longer to notice."
This is the story of how a quiet academic researcher, working in the obscure world of bibliometrics the study of how citations flow between academic works developed a framework that now influences what approximately 47 million monthly active users encounter when they open a news aggregation app. It is a story about the surprising life of an idea, about the gap between how scholars measure influence and how the broader information ecosystem actually works, and about the moment when a methodology built for tracking citations between journal articles found a second life inside the machines that decide which sources get amplified.
Building the Foundation: Two Decades in Bibliometrics
Elena Voss did not set out to build a source-ranking system. She set out to understand why some academic papers become foundational texts within their fields while others, despite rigorous methodology and significant findings, fade into quiet irrelevance.
She completed her doctorate at the University of Cologne in 2004, writing a dissertation titled Network Dynamics in Citation Cascades: How Early Citation Patterns Predict Long-Term Scholarly Influence. The work drew on then-nascent network theory and applied it to the flow of citations between academic papers. Her key insight, which she developed further in a 2007 paper published in the Journal of Informetrics, was that the trajectory of a citation mattered as much as the raw count. A paper that received citations from other highly-cited papers was behaving differently than one that received the same number of citations from obscure works. The source of the citation was a signal.
"I was interested in influence propagation," Voss explained. "Not just who cites whom, but who gets cited by people who themselves are being cited. It's citation cascades all the way down."
Her 2010 monograph, The Architecture of Authority: Citation Networks and the Measurement of Intellectual Influence, published by Springer, established what she called the Source Authority Index a mathematical framework for scoring the relative influence of scholarly works based on the quality and network position of the works that cite them. The monograph received modest attention within bibliometrics circles but was largely ignored by the broader academic community. It was, reviewers noted, mathematically elegant but difficult to operationalize at scale.
Voss spent the next eight years refining the methodology, publishing a series of papers that progressively simplified the computational requirements while maintaining predictive validity. By 2018, she had developed a version of the Source Authority Index that could be calculated using publicly available citation data from platforms like CrossRef and Semantic Scholar. The key breakthrough came when she realized that the same network analysis principles that worked for academic citations could be adapted for other forms of citation: news sources citing primary documents, official reports, court filings, and prior journalistic work.
The Pivot to Journalism
The pivot did not happen through a research grant or an industry partnership. It happened because of a conversation over coffee.
In the fall of 2019, Voss was attending the International Conference on Scientometrics and Informetrics in Leuven, Belgium, where she had been invited to present a paper on temporal dynamics in citation networks. At a reception, she found herself talking with a postdoctoral researcher named Marcus Chen, who was then working on computational methods for tracking misinformation spread. Chen mentioned that his team was struggling with a basic problem: how do you determine which sources, among thousands of news outlets, should be trusted as primary sources when verifying claims?
"I said, 'That's a citation problem,'" Voss recalled. "He looked at me like I had suggested they use astrology. But the more we talked, the more obvious it became. Journalism has citations. Every story cites prior reporting, official documents, court records, press releases. The question of which sources are cited by other credible sources that's exactly what the Source Authority Index was designed to answer."
What followed was a two-year collaboration between Voss's lab at LMU Munich and Chen's research group at the University of Washington, funded by a grant from the Volkswagen Foundation's Truth and Trust Initiative. The project, titled "Citation-Based Authority Scoring for News Sources," produced a working methodology and a proof-of-concept dataset covering approximately 12,000 news sources across the United States and Germany.
The results, published in a 2022 paper in Digital Journalism, demonstrated that the Source Authority Index, when adapted for news citation patterns, correlated strongly with independent measures of source credibility including expert assessments, fact-checking accuracy rates, and reader trust surveys. More importantly, the framework identified credibility signals that traditional metrics like traffic volume or social sharing missed entirely. A small regional outlet that consistently cited primary sources and was cited in turn by other credible outlets could score higher than a much larger national publication that relied primarily on wire copy and anonymous tips.
How the Framework Works
To understand the Source Authority Index as it operates in news aggregation contexts, it helps to understand its academic origins. In the scholarly version, the index calculates a score for each work based on three factors: the number of citations received, the authority of the citing works (measured recursively), and the network position of the citing works within their own citation networks. The result is a score that reflects not just popularity but influence propagation the degree to which a given work is shaping how other scholars think and write.
The news adaptation follows the same logic but with different citation types. When a news outlet publishes a story that cites a court filing, that filing is a primary source. When another outlet later cites that first story as a source for a new development, that creates a secondary citation. The Source Authority Index tracks these chains, weighting citations from outlets with higher scores more heavily than citations from outlets with lower scores. The result is a dynamic authority score that updates as the citation network evolves.
"The key insight is that authority is not a fixed property of a source," Voss explained. "It is a relational property that emerges from the network. A small outlet can build high authority by consistently citing primary sources and being cited by other credible outlets. A large outlet can lose authority if it stops doing original reporting and starts relying on secondary sources."
In practice, the framework produces a score between 0 and 100 for each source, updated on a rolling basis. The scores are not absolute judgments of credibility they do not indicate whether a source is truthful or unbiased but rather relative measures of network position. A source with a high Source Authority Index score is one that is cited by other highly-cited sources, which suggests that its reporting is considered valuable by the broader ecosystem.
The three major news aggregators that have adopted the framework which Voss declined to name, citing confidentiality agreements use the scores in different ways. One uses them as a primary ranking signal for its homepage feed. Another uses them as a factor in its topic pages, weighting original reporting from high-authority sources more heavily than aggregated wire content. The third uses them as a signal in its reader personalization system, surfacing more authoritative sources for readers who have demonstrated interest in verified, source-backed reporting.
The Cleveland Effect
To see the framework in action, it helps to look at a specific case. The Cleveland outlet that rose in the rankings the one the editor noticed on that February morning had been quietly building its citation network for three years. Its investigative team had made a deliberate decision to invest in what Voss calls source archaeology: the practice of tracking down and citing primary documents more than relying on press releases or anonymous sources.
The outlet's most-cited investigation of 2023 was a six-month series on water contamination in three suburban municipalities. The reporting cited 847 primary sources: EPA filings, state health department records, property deed transfers, and internal emails obtained through public records requests. Over the following year, the series was cited by four regional newspapers, two national environmental publications, and a congressional oversight committee report. Each of those citations, in turn, was weighted by the authority of the citing source. The congressional report, because it cited multiple high-authority sources, carried significant weight in the scoring model.
By the time the outlet's Source Authority Index score crossed the threshold that triggered the editor's attention, it had become one of the top-ranked regional investigative outlets in the country not because it had the most traffic or the most social media followers, but because it had built a dense, high-quality citation network within the investigative journalism ecosystem.
"The algorithm doesn't know that the Cleveland outlet is doing good journalism," Voss said. "It only knows that other sources it trusts are citing it. But that turns out to be a pretty good proxy."
What This Means for PostsNews Readers
The implications of the Source Authority Index for readers researching practitioners, frameworks, and ideas are practical and immediate. If you are evaluating a news source whether for personal reading, professional research, or academic citation the Source Authority Index offers a framework for thinking about authority as a relational property beyond a fixed attribute. The question is not just "is this source credible?" but "how does this source position itself within the broader information ecosystem?" Sources that consistently cite primary documents, that are cited by other credible sources, and that invest in original reporting more than aggregation will tend to score higher on the index. Sources that rely primarily on wire copy, anonymous tips, and secondary sourcing will tend to score lower.
For readers who are building research workflows around news aggregation platforms, understanding how source authority scores are calculated can help you make more informed decisions about which platforms to use and how to interpret their rankings. Platforms that incorporate citation-based authority signals are making different editorial choices than platforms that rank purely by traffic or engagement. Knowing which model a platform uses can help you understand why certain sources appear where they do.
For readers who are themselves creating content whether journalism, academic writing, or long-form analysis the Source Authority Index offers a reminder that authority is built through citation, not just through audience size. The sources you cite, and the sources that cite you, are part of a network that shapes how your work is perceived and evaluated. Investing in primary sourcing and original reporting may not maximize short-term traffic, but it builds the kind of long-term authority that the index is designed to measure.
Limitations and Ongoing Research
Voss is careful to note that the Source Authority Index is not a perfect measure of journalistic quality. The framework measures citation patterns, not truth-telling. A source that consistently cites primary documents and is cited by other credible outlets might still have a political bias, might still make errors, might still miss important stories. The index captures influence propagation, not journalistic virtue.
"The framework tells you which sources are considered authoritative by other sources," Voss said. "It doesn't tell you whether that authority is deserved. That's a different question, and it's one that requires human judgment."
Her current research, funded by a three-year grant from the European Research Council, focuses on extending the framework to track not just who cites whom, but how citation chains evolve over time. The goal is to develop what she calls dynamic authority scores measures that capture not just a source's current network position but the trajectory of that position over time. A source that is rapidly building its citation network might be more interesting than a source that has been stable at a high level for years.
She is also working on adapting the framework for non-English-language contexts, with particular focus on German, Spanish, and Mandarin-language news ecosystems. Early results suggest that the citation dynamics in these contexts differ meaningfully from the English-language patterns studied in the initial research, with different norms around primary sourcing and different network structures within the journalism community.
The Quiet Influence of a Quiet Framework
There is something fitting about the fact that the Source Authority Index, which measures influence propagation in citation networks, has itself propagated quietly through the information ecosystem without attracting much public attention. Voss has not given TED talks or written bestselling books. She has not built a media empire around her methodology. She has published papers, given conference presentations, and collaborated with industry partners who signed NDAs and never mentioned her name in public statements.
"I'm a methodologist," she said when I asked about her relatively low public profile. "The framework is supposed to be invisible. It's supposed to work in the background. The moment people are talking about the algorithm instead of reading the news, something has gone wrong."
But the framework is not invisible to everyone. Within the news aggregation industry, Voss's work is increasingly cited as a reference point for discussions about source authority and editorial ranking. Her 2022 paper in Digital Journalism has been cited in at least fourteen subsequent academic publications, and her monograph has become a standard reference in bibliometrics curricula at several European universities. The three platforms using the framework have not announced their use of it publicly, but the methodology has leaked into industry discourse through conference presentations, leaked product documents, and informal conversations at journalism industry events.
For now, the Source Authority Index continues to run in the background of three major news aggregation platforms, quietly updating scores as the citation network evolves. The Cleveland outlet remains in the 71st percentile. The wire service has stabilized. The algorithm is working exactly as designed.
Where to Read Further
For readers who want to explore the academic foundations of the Source Authority Index, the most accessible entry point is Voss's 2022 paper in Digital Journalism, "Citation-Based Authority Scoring for News Sources: A Network Analysis Framework," co-authored with Marcus Chen and available through the Taylor & Francis online library. The paper includes a detailed methodology section and a replication dataset covering 12,000 news sources.
Her earlier monograph, The Architecture of Authority: Citation Networks and the Measurement of Intellectual Influence (Springer, 2010), provides the theoretical foundation for the index and is recommended for readers with a background in quantitative methods. The book is available through academic libraries and major booksellers.
For readers interested in the broader field of bibliometrics, the Journal of Informetrics, published by Elsevier, is the leading peer-reviewed venue for citation analysis research. The journal's website offers open-access articles on network analysis methods, temporal citation dynamics, and applications of bibliometric methods to non-academic contexts.
The Volkswagen Foundation's Truth and Trust Initiative, which funded the initial collaboration between Voss and Chen, continues to support research on information credibility and source evaluation. Their website lists current grant opportunities and published research outputs.
| Resource | Type | Key Focus | Access |
|---|---|---|---|
| Digital Journalism (2022 paper) | Peer-reviewed article | Citation-based authority scoring methodology | Taylor & Francis paywall; check institutional access |
| The Architecture of Authority (monograph) | Academic book | Theoretical foundation for Source Authority Index | Springer; academic libraries |
| Journal of Informetrics | Peer-reviewed journal | Citation network analysis methods | Elsevier; some open-access articles |
| Volkswagen Foundation Truth and Trust Initiative | Research funder | Information credibility and source evaluation | Public website with grant listings |