Last Updated on July 16, 2026 by Dr. Bhagat
Journal Metrics·Updated June 2026
What Is the Eigenfactor Score?
Explore the Eigenfactor Score: a network-based metric that measures a journal's total influence on the scientific citation ecosystem.
SectionWhat the Eigenfactor Score is
The Eigenfactor Score was introduced in 2007 by Jevin West and Carl T. Bergstrom, a biologist and information scientist at the University of Washington. Their goal was to create a metric that measured total influence rather than average performance.
The core insight behind Eigenfactor is simple: the scientific literature is not a collection of isolated journals but a network where ideas flow through citations. When a highly influential journal like Nature or Science cites a paper, that citation carries more weight than a citation from a rarely-read journal. Eigenfactor captures this by using an iterative network algorithm that assigns each journal a score based on how much “influence” it receives from — and passes to — other journals in the network.
### Key features of Eigenfactor – Network-based: It analyzes the entire citation network of thousands of journals simultaneously. – Prestige-weighted: Citations from influential journals count more than citations from obscure ones. – No self-citation inflation: Journal self-citations are excluded from the calculation entirely.
– Total influence, not average: It measures a journal’s overall contribution to the network, not citations per article. – Freely available: Eigenfactor scores are published in JCR and on the free Eigenfactor.org website. ### Why Eigenfactor matters for researchers The Impact Factor tells you the average citation rate of a journal’s articles.
Eigenfactor tells you the journal’s total footprint on science. A journal with a high Impact Factor but low Eigenfactor might publish few, highly cited articles but have limited overall influence. A journal with a lower Impact Factor but high Eigenfactor might publish many widely read articles that shape research across multiple fields.
SectionHow Eigenfactor is calculated
Eigenfactor uses the same mathematical framework as Google’s PageRank algorithm: an eigenvector centrality calculation on a directed network. ### The Eigenfactor algorithm in plain terms 1. Build the citation network. Every journal in the Web of Science Core Collection is a node.
Every citation from one journal to another is a directed edge (arrow) with a weight. 2. Exclude self-citations. Citations from a journal to itself are removed from the network.
This prevents journals from artificially inflating their scores. 3. Assign random surfer probabilities. The algorithm includes a “random walk” component (similar to PageRank’s damping factor): a small probability that a researcher moves to any random journal in the network rather than following a citation.
This prevents isolated clusters from dominating the results. 4. Iterate until convergence. The algorithm repeatedly calculates how much influence each journal receives from the journals that cite it.
If Journal A (highly influential) cites Journal B, Journal B gains influence. If Journal B then cites Journal C, some of Journal A’s influence flows through to C. This process repeats until the scores stabilize.
5. The final scores are the Eigenfactor scores. These sum to 100 across all journals in the network, so each score can be read as a percentage of total scientific influence. ### The Eigenfactor formula (conceptual) The exact computation uses matrix algebra and iterative convergence, but the conceptual formula is: Eigenfactor = f(network position, citation volume, prestige of citing journals) More specifically, for journal j: > EF(j) = α × Σᵢ (EF(i) × P(i→j)) + (1–α) × (1/N) Where: – EF(i) = Eigenfactor score of the citing journal i – P(i→j) = Probability of moving from journal i to journal j (based on citation patterns) – α = Damping factor (typically ~0.85, similar to PageRank) – N = Total number of journals in the network – (1–α) × (1/N) = Random jump probability (prevents dead ends and manipulation) ### Key parameters of Eigenfactor | Parameter | Value | |———–|——-| | Source database | Web of Science Core Collection (Clarivate) | | Citation network | All journals in the database simultaneously | | Self-citations | Excluded entirely | | Citation weighting | Yes — prestige-weighted (PageRank-style) | | Field normalization | No (raw network scores) | | Score scale | Sum to 100 across all journals (percentage of total influence) | | Cost to access | Free (Eigenfactor.org) or via JCR subscription | | Release frequency | Annually (with JCR) |
SectionEigenfactor vs Impact Factor
Eigenfactor and the Impact Factor are fundamentally different metrics, and comparing them reveals important truths about how journals shape scientific discourse. ### Comparison table: Eigenfactor vs Impact Factor | Feature | Eigenfactor Score | Impact Factor | |———|——————-|—————| | Developer | Jevin West & Carl Bergstrom (University of Washington) | Eugene Garfield (ISI, now Clarivate) | | Source database | Web of Science | Web of Science | | What it measures | Total influence on the citation network | Average citations per citable item | | Level of analysis | Journal’s position in the entire network | Per-article citation rate | | Citation weighting | Yes — citations from prestigious journals count more | No — all citations equal | | Self-citations | Excluded entirely | Included (suppressed if excessive) | | Field normalization | No | No | | Score scale | Percentage of total network influence (sum = 100) | Decimal (average citations per article) | | Best for | Measuring total scholarly footprint | Traditional reputation assessment | ### When Eigenfactor and Impact Factor diverge A journal with a high Impact Factor but low Eigenfactor might be: – A niche journal with a few blockbuster articles but limited overall reach – A review journal that publishes infrequently but each article is highly cited – A new journal that has not yet built a broad network of citing journals A journal with a lower Impact Factor but high Eigenfactor might be: – A large, multidisciplinary journal that publishes many articles shaping multiple fields – A foundational journal in a major discipline (e.g., PNAS, Physical Review) – A journal whose articles are consistently cited by many other journals across the network ### Examples of high Eigenfactor journals The journals with the highest Eigenfactor scores are typically the largest, most multidisciplinary journals that shape research across many fields. Based on recent JCR data, the top tier includes: | Journal | Approximate Eigenfactor | Notes | |———|————————|——-| | Nature | ~0.60 | Multidisciplinary; among the highest total influence | | Science | ~0.55 | Multidisciplinary; massive cross-field reach | | PNAS | ~0.35 | Large volume + broad disciplinary coverage | | Cell | ~0.25 | Top-tier life sciences with high per-article impact | | Journal of Biological Chemistry | ~0.15 | Very high volume; enormous total output | | Physical Review Letters | ~0.12 | Core physics journal with broad citing network | | IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) | ~0.10 | Major conference proceedings with network influence | | New England Journal of Medicine | ~0.08 | High per-article impact but lower volume than above | Note: Eigenfactor scores change annually and are proportional to the total network.
These figures are approximate for illustration.
SectionHow to interpret Eigenfactor scores
Because Eigenfactor scores represent a percentage of total influence, interpretation requires understanding the scale. ### How to read an Eigenfactor score – An Eigenfactor of 0.50 means the journal accounts for 0.50% of all scientific influence in the Web of Science network. This is enormous — only a handful of journals ever exceed this level.
– An Eigenfactor of 0.05 means the journal accounts for 0.05% of total influence. This is still highly significant; there are thousands of journals in the network. – An Eigenfactor of 0.001 means the journal accounts for 0.001% of total influence.
This is a small but non-zero contribution. ### Why Eigenfactor scores are not comparable to Impact Factors It is tempting to compare Eigenfactor scores directly to Impact Factors, but the scales are fundamentally different: – Impact Factor is a rate (citations per article, typically 0–50+). – Eigenfactor is a share of total influence (percentage of network, typically 0.001–0.60).
A journal with an Impact Factor of 30 and an Eigenfactor of 0.05 is not “better” or “worse” than a journal with an IF of 2.0 and an Eigenfactor of 0.15. The first has highly cited individual articles; the second has a broader total influence on the scientific ecosystem.
SectionThe relationship with Article Influence Score
Clarivate publishes a companion metric to Eigenfactor called the Article Influence Score (AIS). While Eigenfactor measures total influence, AIS measures average influence per article. ### How Article Influence Score is calculated AIS = Eigenfactor ÷ (Number of articles published over 5 years) × 0.01 The multiplication by 0.01 simply rescales the number for readability.
The result is a metric that is conceptually similar to the Impact Factor but prestige-weighted and self-citation-free. ### Article Influence Score vs Impact Factor | Feature | Article Influence Score (AIS) | Impact Factor | |———|——————————|—————| | What it measures | Average prestige-weighted influence per article | Average raw citations per article | | Self-citations | Excluded | Included (capped) | | Prestige weighting | Yes | No | | Scale | Similar to IF (typically 0–30+) | Typically 0–50+ | | Best for | Fairer per-article assessment | Traditional evaluation | Use AIS when you want a per-article metric that corrects for citation quality. Use Eigenfactor when you want to understand a journal’s total scholarly footprint.
Key Takeaways
- Eigenfactor measures total network influence.
- It uses a PageRank-like algorithm.
- Scores are in Clarivate JCR.
- Complements Impact Factor.
FAQPeople also ask
What is the Eigenfactor Score used for?
It measures a journal’s total influence on the scientific ecosystem.
Who created the Eigenfactor Score?
Jevin West and Carl Bergstrom at the University of Washington in 2007.
Is Eigenfactor similar to PageRank?
Yes. It uses an iterative network algorithm.
Where can I find Eigenfactor scores?
In Clarivate’s Journal Citation Reports.