Researchers choosing where to publish, librarians managing subscriptions and evaluators reviewing CVs all run into the same three numbers: the Journal Impact Factor, CiteScore and the SCImago Journal Rank (SJR). They look similar but measure slightly different things, and they can rank the same journals differently.
At a glance
| Journal Impact Factor (JIF) | CiteScore | SJR | |
|---|---|---|---|
| Provider | Clarivate (Journal Citation Reports) | Elsevier (Scopus) | SCImago, based on Scopus data |
| Database | Web of Science Core Collection | Scopus | Scopus |
| Citation window | Citations in year Y to items from the previous 2 years | Citations in a 4-year window to items from the same 4 years | Citations in year Y to items from the previous 3 years |
| Documents counted | Citations to all items, divided by “citable items” (articles and reviews) | All document types in both numerator and denominator | Citations weighted by prestige |
| Weighting | All citations equal | All citations equal | Citations from high-prestige journals count more (PageRank-like) |
| Access | Subscription | Free to view | Free |
Journal Impact Factor
The JIF for year Y is the number of citations received in year Y by items the journal published in Y−1 and Y−2, divided by the number of articles and reviews it published in those two years.
Strengths: long history, widely recognised, available for decades of data.
Weaknesses:
- The numerator/denominator mismatch: citations to editorials, letters and news items count in the numerator, but those items do not count in the denominator. This can inflate the JIF of journals that publish a lot of front matter.
- A two-year window suits fast-moving fields such as molecular biology. It penalises fields where citations accumulate slowly, such as mathematics and social sciences. Clarivate also publishes a five-year JIF for this reason.
CiteScore
CiteScore for year Y counts citations received during a four-year window (Y−3 to Y) to documents published in the same four years, divided by the number of those documents. All document types count on both sides.
Strengths: consistent numerator and denominator, longer window, free to view, and Scopus covers more journals than Web of Science, including more non-English and regional titles.
Weaknesses: journals that publish many editorials and letters (which are rarely cited) get lower scores than under the JIF. Like the JIF, it treats every citation equally.
SCImago Journal Rank (SJR)
SJR applies an idea similar to Google’s PageRank to citations. A citation from a highly cited, prestigious journal is worth more than a citation from a little-cited one. It also limits the effect of journal self-citation.
Strengths: weights citation quality, not just quantity; free; includes country and subject rankings and quartiles.
Weaknesses: harder to explain and reproduce, and the weighting reinforces the position of journals that are already prestigious.
Why the rankings differ
The same journal can rank differently on the three metrics because of:
- Coverage: citations from journals indexed only in Scopus count for CiteScore and SJR, but not for the JIF.
- Window length: fast-citing fields look better with short windows.
- Document mix: journals with many non-research items behave differently.
- Weighting: SJR rewards being cited by central journals.
How to use journal metrics sensibly
- Compare within a field only. A JIF of 3 can be excellent in one field and modest in another. Use quartiles or percentile ranks within subject categories.
- Look at several metrics and years. A single year can be distorted by one highly cited paper.
- Never judge an individual paper or researcher by the journal’s metric. This is the core recommendation of DORA and the Leiden Manifesto. Citation distributions are skewed, so most papers in any journal are cited less than the journal’s average.
- Check for warning signs: unusually high self-citation, sudden jumps, or a journal that is listed in a citation index but has questionable peer review.
- Consider what you actually need. For choosing where to publish, audience, scope, review speed, open-access options and editorial quality often matter more than a decimal point of impact.
Journal metrics are useful for questions about journals, such as which journals in a field are most cited or how a journal’s influence is changing. They are poor answers to questions about the quality of an individual piece of research. That still requires reading it.