1. Introduction
For many researchers, journal quartiles become consequential at specific points in a career: when a promotion form requests them, when a supervisor asks whether a target journal is ranked Q1, or when a funding application requires the indexing status of each output. In these settings, an indicator designed to describe journals is frequently used, in effect, to describe the researchers who publish in them.
Given this use, it is helpful to understand how the indicator is constructed. This article outlines what SJR calculates, how quartiles are derived from it, which reporting conventions can mislead, and where the indicator provides little or no information.
2. What SJR calculates
SCImago Journal Rank is a prestige-weighted measure of citations per document (González-Pereira et al., 2010; Guerrero-Bote & Moya-Anegón, 2012). Two features of this definition are particularly relevant to its interpretation.
It is an average per document. SJR considers citations received in a given year by documents that a journal published in the preceding three years, relative to the number of those documents. Journal size is therefore largely normalised: a journal publishing 60 articles a year and one publishing 600 are placed on a comparable scale. This helps to explain why a small, specialised journal may rank above a large flagship title, and why publication volume alone does not raise the indicator.
It is prestige-weighted. This feature distinguishes SJR from a simple citation ratio and draws on reasoning similar to that underlying Google's PageRank algorithm. Citations are not weighted equally: a citation from a journal that is itself highly cited by influential journals contributes more than a citation from a journal that receives few citations. The calculation is iterative, so a journal's score depends on the standing of the journals that cite it, which in turn depends on the standing of the journals that cite them.
Two implications follow, and neither is apparent from a quartile badge alone.
First, SJR reflects a journal's position in a citation network rather than a count of citations. A journal may increase its raw citation numbers with relatively little change in its SJR if the additional citations come from peripheral venues. Conversely, a journal that is cited less often, but by central journals in its field, may score higher than its raw citation count would suggest.
Second, SJR values are strongly field-dependent. Citation practices vary considerably across disciplines. Researchers in molecular biology, for example, tend to cite recent work quickly and in long reference lists, whereas historians and applied linguists typically cite more slowly, more sparsely, and more often to books that citation databases index less completely. An SJR of 1.5 may therefore indicate a strong position in education research but a modest one in genomics. For this reason, a single threshold for a "good" SJR is unlikely to be meaningful across disciplines.
3. How quartiles are derived
Quartiles are intended to address this problem of comparability. Within each subject category, SCImago ranks journals by SJR and divides the list into four groups of equal size: Q1 contains the top 25 per cent of journals in that category, Q2 the next 25 per cent, and so on.
Because the division is made within a category, quartiles are considerably more comparable across disciplines than raw SJR values. A Q1 journal in applied linguistics and a Q1 journal in immunology each sit in the top quarter of their respective lists, whatever their absolute scores. This is a genuine advantage, and it may explain why many institutions refer to quartiles rather than to raw scores. However, the approach also has a structural feature that is often overlooked.
4. The best-quartile convention
A hypothetical example illustrates the consequence. A journal on language teaching might be classified in Linguistics and Language, a large and competitive category; in Education; and in a smaller arts and humanities category such as Language and Linguistics. It could rank Q2 in the first two categories and Q1 in the third, because the third category is small and its median journal receives fewer citations.
In such a case, the displayed badge would read Q1. This is not an error, but it is accurate for only one of the three categories, and that category may not be the one that assessors have in mind.
A practical recommendation follows: for consequential decisions, it is advisable to consult the per-category breakdown and to use the category that best describes the work in question. In the Scholar Metrics extension, this breakdown appears when hovering over the badge; on scimagojr.com, it is shown on each journal's page. The check takes little time relative to the months a submission may spend in review.
5. Limitations of the indicator
The preceding sections concern what SJR measures. It is equally important to recognise what it does not measure.
- It provides no information about individual articles. This is arguably the most consequential source of confusion. SJR is an average across a journal's output, and citation distributions within journals are typically highly skewed: a small proportion of papers attracts most citations, while many receive few (Seglen, 1997). A quartile describes the venue. It is not evidence that a particular article is correct, well designed, or important.
- It provides no information about peer review. No component of the calculation observes whether review at a journal was rigorous, cursory, or absent.
- It is distinct from the Journal Impact Factor. The Journal Impact Factor (JIF) is a Clarivate metric calculated from Web of Science data over a two-year window without prestige weighting. SJR is derived from Scopus data, uses a three-year window, and is weighted. The two indicators can rank journals differently, and treating one as the other is a common error.
- It is published with a time lag. Values are released annually and rest on a three-year citation window. A journal whose standing has recently declined may continue to show a strong score for some time, while a rapidly improving journal may be undervalued.
- It covers books and conference proceedings less completely. In fields where monographs or refereed conference papers are primary publication venues, including much of the humanities and computer science, a journal-based citation network may not capture much of the most important work.
6. Recommendations for use
These limitations do not make the indicator uninformative. They suggest that it is best treated as an aid to triage, which is a useful function in its own right.
Use it to prioritise rather than to judge. When faced with a long list of search results, indexing and quartile information offers a defensible way to decide what to read first. This is arguably its strongest legitimate use.
Consider subject fit before quartile. A Q2 journal that is widely read in a researcher's subfield may be a more suitable venue than a Q1 journal in an adjacent field where the intended readership is unlikely to encounter the work. Quartiles reflect citation position. They do not indicate whether the intended readers will see the work.
Check the coverage years. A journal whose Scopus coverage ended in a previous year has, in most cases, been discontinued from the database, and new articles published there are unlikely to be indexed, whatever the historical quartile indicates. This check is quick and may prevent costly errors.
Avoid aggregating quartiles. A statement such as "twelve Q1 papers" conveys little without information on authorship position, field, and contribution. Experienced assessment panels are likely to recognise this, and a CV that relies on such totals may give the impression that the indicator has been mistaken for the achievement.
Verify information before official use. Tools, including those on this site, draw on annual snapshots, and indexing status can change during the year. Before a submission, promotion application, or funding form, it is advisable to confirm details at the source: scimagojr.com for quartiles and SJR, and mjl.clarivate.com for Web of Science indexing.
7. Conclusion
A quartile condenses a three-year, weighted citation network into a single label. This degree of compression is useful precisely because it discards information, and it carries risks for the same reason. A quartile is a reasonable starting point for questions about a journal, but a weak basis for final judgements about it, and an inappropriate basis for judgements about an individual piece of research.
The Scholar Metrics extension displays quartiles and indexing information in Google Scholar search results, with the per-category breakdown available on hover for the reasons outlined above.
References
- González-Pereira, B., Guerrero-Bote, V. P., & Moya-Anegón, F. (2010). A new approach to the metric of journals' scientific prestige: The SJR indicator. Journal of Informetrics, 4(3), 379–391. https://doi.org/10.1016/j.joi.2010.03.002
- Guerrero-Bote, V. P., & Moya-Anegón, F. (2012). A further step forward in measuring journals' scientific prestige: The SJR2 indicator. Journal of Informetrics, 6(4), 674–688. https://doi.org/10.1016/j.joi.2012.07.001
- Seglen, P. O. (1997). Why the impact factor of journals should not be used for evaluating research. BMJ, 314(7079), 497. https://doi.org/10.1136/bmj.314.7079.497