Impact Factor and Journal Metrics: What They Mean and How They Are Calculated

A single number printed next to a journal's name can determine whether a researcher's career advances, whether a grant application succeeds, or whether a paper gets read at all. Journal metrics — led by the Impact Factor but extending into a constellation of related scores — function as the credit ratings of academic publishing. This page explains how the major metrics are calculated, what causal forces shape them, where they diverge from each other, and where they quietly mislead the people who rely on them most.

• Definition and scope
• Core mechanics or structure
• Causal relationships or drivers
• Classification boundaries
• Tradeoffs and tensions
• Common misconceptions
• Checklist or steps
• Reference table or matrix

Definition and scope

The Journal Impact Factor (JIF) is a ratio: citations received by a journal in a given year divided by the number of citable items the journal published in the prior two years. Clarivate Analytics — which owns and publishes the annual Journal Citation Reports — is the sole commercial source of official Impact Factor values. The metric was developed by bibliometrician Eugene Garfield in the 1950s and 1960s as an internal tool for selecting journals to include in the Science Citation Index, not as a universal measure of quality.

Impact Factor covers only journals indexed in Clarivate's Web of Science database. Journals outside that index receive no JIF regardless of their scientific merit. Alongside JIF, three other metrics now hold significant institutional weight: the SCImago Journal Rank (SJR), the Eigenfactor Score, and the Article Influence Score — each drawing on different underlying citation databases and methodological assumptions. The Scimago Journal Rank and the Eigenfactor and Article Influence Score are covered in dedicated reference pages; this page treats them comparatively within a broader metrics framework.

Core mechanics or structure

Impact Factor calculation. The 2-year JIF for year Y is:

JIF(Y) = Citations in Y to articles published in (Y-1) and (Y-2) ÷ Citable items published in (Y-1) and (Y-2)

"Citable items" includes original research articles and review articles, but excludes editorials, letters, news items, and errata — even though those excluded items still attract citations that land in the numerator. This asymmetry inflates the ratio and is a documented structural artifact, not an accident.

Clarivate also publishes a 5-year Impact Factor using the same logic across a longer citation window. The 5-year variant is more stable year-to-year and tends to favor fields where citation accumulation is slow, such as mathematics or the humanities, where the 2-year window captures only a fraction of a paper's citation lifetime.

SCImago Journal Rank (SJR). SJR is calculated from Elsevier's Scopus database and applies a weighted citation model: citations from high-prestige journals count more than citations from low-prestige journals — similar in logic to Google's PageRank algorithm. A citation from Nature carries more SJR weight than a citation from a regional specialty journal with low visibility.

Eigenfactor Score. Developed by Jevin West and Carl Bergstrom at the University of Washington and freely available at eigenfactor.org, the Eigenfactor Score uses all five years of citation data from Web of Science and weights citations by journal prestige in a manner similar to SJR. Crucially, it excludes self-citations — citations a journal makes to its own prior articles — which SJR and JIF do not exclude by default.

Article Influence Score. The Article Influence Score normalizes the Eigenfactor by dividing it by the journal's share of total articles across the database, making it a per-article prestige estimate. A score of 1.0 represents the field-wide average; scores above 1.0 indicate above-average influence per published article.

h-index at the journal level. The h-index — most familiar as an author-level metric — can be applied to journals. A journal with an h-index of 150 has published at least 150 articles each cited at least 150 times. The h-index and citation metrics page covers author-level application in detail.

Causal relationships or drivers

Several structural features of a journal's operation reliably push its Impact Factor upward or downward, independent of raw scientific quality.

Review articles. Reviews synthesize existing literature and attract citation rates far higher than original research. A journal that publishes a high proportion of review articles will accumulate more citations in its numerator without increasing its denominator proportionally. Chemical Reviews, which publishes almost exclusively review articles, carries an Impact Factor consistently above 50 — a level no primary-research chemistry journal approaches.

Field size and citation norms. Fields with large active researcher populations and high publication rates generate more citations in absolute terms. A journal in molecular biology competes in a citation-rich environment; a journal in Antarctic geology does not. Raw JIF comparisons across fields are methodologically incoherent for this reason.

Publication frequency and article volume. A journal that publishes 50 articles per year has 50 chances to attract citations to its denominator. A journal that publishes 5,000 articles per year faces a statistical averaging effect — a handful of highly cited papers cannot move the overall ratio as dramatically.

Self-citation practices. Journals can artificially inflate JIF by encouraging (or requiring) authors to cite prior articles from the same journal. Clarivate has suppressed Impact Factors for journals found to engage in citation stacking, as documented in the annual Journal Citation Reports since 2017.

Classification boundaries

Not all journals receive Impact Factors, and not all citation-based scores are equivalent in what they measure.

Journals indexed only in Scopus but not Web of Science receive SJR scores but no JIF. Open-access journals, including those that meet high scientific standards, may lack JIF values simply because they are newer and have not yet completed the Clarivate indexing review process. The journal indexing databases page covers which databases index which journals and on what criteria.

National or regional journals publishing in languages other than English are systematically underrepresented in both Web of Science and Scopus, which means their metrics consistently understate their actual citation footprint within their regional scientific communities.

The San Francisco Declaration on Research Assessment (DORA), published in 2012 and signed by thousands of researchers and institutions (DORA), formally identifies these boundaries as reasons to restrict the use of JIF in hiring and promotion decisions.

Tradeoffs and tensions

The tension at the center of journal metrics is the gap between what the numbers claim to measure and what they actually can measure.

JIF is a journal-level statistic applied — routinely and incorrectly — as though it were an article-level or author-level statistic. When a hiring committee evaluates a candidate partly on whether their papers appeared in high-JIF journals, they are treating a population average as an individual score. The actual citation distribution within any given journal is extremely skewed: in most journals, the top 20% of articles account for roughly 80% of all citations received, a distribution documented by Larivière and Gingras in a 2010 analysis published in the Journal of the American Society for Information Science and Technology.

Normalization by field is the most actively contested methodological divide. The Source Normalized Impact per Paper (SNIP), developed by Henk Moed at the Centre for Science and Technology Studies in Leiden and available through Elsevier's Journal Metrics platform, attempts to correct for field-specific citation density by dividing a journal's citations by the average number of references in the citing field. SNIP values are, in theory, comparable across disciplines; JIF values are not.

The open-access publishing in science landscape adds another layer of tension. Open-access articles demonstrably accumulate citations faster than paywalled counterparts in several empirical analyses — the "open-access citation advantage" — which can gradually shift Impact Factors for journals that convert to open-access models, independent of any change in editorial standards.

Common misconceptions

"A higher Impact Factor means the journal publishes better science." JIF measures citation frequency, not correctness, reproducibility, or methodological rigor. High-profile retractions have occurred in journals with Impact Factors above 30. The retractions and corrections in science record provides systematic evidence that citation frequency and scientific validity are not the same variable.

"JIF applies to individual papers." It does not. JIF is a ratio computed for the journal as a whole. A paper published in a journal with JIF 15 may itself receive zero citations; a paper in a journal with JIF 1.2 may receive 400.

"All metrics measure the same thing." SJR, Eigenfactor, SNIP, and JIF weight citations differently, draw on different databases, apply different normalization steps, and exclude different types of documents. A journal ranked 12th by JIF within its category may rank 45th by SNIP. These divergences are not errors — they reflect genuinely different questions being asked.

"Predatory journals have low Impact Factors." Some predatory journals have successfully obtained indexing and JIF values through strategic citation manipulation. Impact Factor does not reliably screen for predatory journals.

Checklist or steps

The following sequence describes how Clarivate processes a journal's annual Impact Factor calculation for publication in the Journal Citation Reports:

1. The citation counting period opens: all citations recorded in Web of Science for the target calendar year are compiled.
2. Citations pointing to articles published by the journal in the two prior calendar years are extracted and summed (numerator).
3. All original research articles and review articles published by the journal in those same two prior years are counted (denominator). Editorials, corrections, and news items are excluded.
4. The JIF is calculated as numerator ÷ denominator, rounded to three decimal places.
5. Clarivate applies anomaly detection for citation stacking — journals where more than 25% of citations come from a single citing journal are flagged and may be suppressed.
6. Final values are published in the annual Journal Citation Reports release, typically mid-year following the citation year.
7. The 5-year Impact Factor undergoes the same process using a five-year prior-publication window instead of two years.

The manuscript submission process and how to choose the right journal pages both address how authors navigate these metrics when selecting publication targets. The broader scientific journal authority index offers orientation across all related reference topics.

Reference table or matrix

References

• Clarivate Journal Citation Reports
• SCImago Journal & Country Rank
• Eigenfactor Project — University of Washington
• San Francisco Declaration on Research Assessment (DORA)
• Elsevier Journal Metrics (SNIP and SJR documentation)
• Leiden Manifesto for Research Metrics — Hicks, Wouters, Waltman, de Rijcke, Rafols (2015), Nature 520:429–431
• Larivière, V. & Gingras, Y. (2010). The impact factor's Matthew Effect. Journal of the American Society for Information Science and Technology, 61(2):424–427