Guides · Publishing and impact

Journal metrics explained: impact factor, SJR, SNIP and h-index

The impact factor divides the citations a journal receives in one year to items from the previous two years by the number of citable items it published in those years. Alternatives normalise for field (SNIP), weight citations by their source (SJR, Eigenfactor) or lengthen the window to five years. All of them depend on field, database and time, and because citations are highly skewed, a journal's average is a poor guide to any single article in it.

From retrieval tool to ranking tool

Citation indexing was proposed in 1955 as a way to link papers through the ideas they share, so that readers could follow how a work was used by later authors. That proposal already used the phrase impact factor for the influence of a single article. The Science Citation Index followed in 1964, and in 1972 Garfield showed how citation data could be used to rank journals.^[1][2][3]

Looking back in 2006, Garfield noted that his 1972 paper on journal evaluation had drawn far more attention than the 1955 proposal, and that citation indexing had moved from a system for finding literature to a tool for evaluating research, a use he suggested had come to overshadow the original purpose.^[4]

How the impact factor is calculated

Garfield's starting point was that a journal's total citations depend on how much it publishes as well as on the significance of what it publishes. To discount size, he divided citations by the number of articles published, giving an average citation rate per article. He chose a two-year publication window because the typical article in his data was cited most heavily in the two years after it appeared.^[3]

In its standard form, a journal that published 100 items over two years, which were then cited 200 times in the following year, has an impact factor of 2. The figure is not exactly an average, because the numerator counts citations to all document types while the denominator counts only so-called citable items. A five-year impact factor exists because two years is too short in some fields, and the Journal Citation Reports also give an immediacy index for citations received in the year of publication.^[5]

Size and average give different pictures. When Garfield ranked journals by impact factor instead of total citations, only 75 of the 152 most frequently cited journals appeared on both lists.^[3]

Other journal indicators you will see

Many alternatives have been proposed as complements to the impact factor. Those most often shown in citation databases fall into a few families.^[5]

• Rank-based indicators place a journal by the position of its impact factor within its subject category, so a journal ranked 10th of 200 scores about 0.95.^[5]
• Field-normalised indicators compare a journal's citations with what is expected in its field. SNIP, shown in Scopus, uses a citing-side approach to this normalisation.^[5]
• Recursive indicators give more weight to citations from high-impact sources. Article Influence and Eigenfactor appear in the Journal Citation Reports and SJR in Scopus. Eigenfactor grows with journal size, both Eigenfactor and Article Influence ignore a journal's citations to itself, and the current SJR also weights citations by how close the citing journal is in subject.^[5]
• A journal h-index is sometimes reported, but unlike most journal indicators it rises with the number of articles a journal publishes.^[5]
• Structural influence, developed for marketing journals, measures how much other journals in a citation network depend on a journal for knowledge, taking into account the influence of the citing journals and indirect links.^[6]
• Counting the distinct authors who cite a work, rather than the citations, has been proposed as a way to capture how widely its influence spreads.^[7]

Why scores differ by field, database and time

Basic indicators such as the impact factor do not correct for differences in citation density between fields, which is why normalised indicators exist. Scores also depend on the database. Scopus covers more journals than Web of Science and tends to report higher citation counts. Google Scholar has much broader coverage of business and management than Web of Science, but studies report gaps, inaccurate counts and a lack of transparency about what it covers.^[5]

Time matters too. Older articles have had longer to gather citations, and as a field splits into more subareas, each topic has fewer scholars to cite it. Percentile indicators carry their own risk: because citation distributions are skewed, a top-percentile score can mislead when the least-cited papers of a country or journal deviate from the expected pattern, which happens in all journals with high impact factors. Comparing the share of uncited papers helps reveal these cases.^[8][9]

A journal's score is not an article's score

The most criticised use of the impact factor is to judge an article or its authors by the journal it appeared in. Citations within a journal are highly skewed, so that, for example, a fifth of its papers may receive 60% of its citations. The average is driven by a few highly cited papers, and most articles fall well below it. Garfield warned against this use, and the San Francisco Declaration on Research Assessment argues against it. As early as 1972 he noted that one extraordinarily cited article could distort a journal's ranking.^[5][3]

Articles in higher impact factor journals are cited more on average, which a review of citation predictors calls almost a tautology, since the impact factor is built from those citations and counts only short-term ones. It says less about quality: in the behavioural and brain sciences, impact factors were weak and inconsistent predictors of statistical accuracy and replicability, and sometimes negatively related to them. Some researchers argue that a journal's indicator adds information for very recent papers that have not yet had time to be cited.^[10][11][5]

What marketing studies of journal influence show

In a network of 49 marketing and marketing-related journals, a few journals held a disproportionate share of influence: in 1996 and 1997 the top five accounted for more than 60% of it. Different journals led different subareas of marketing, and the Journal of Marketing had the broadest span across them.^[6]

The same study compared measures. Structural influence correlated strongly with a ranking based on marketing faculty's votes for their top journals, while the SSCI impact factor correlated less with structural influence and not significantly with the faculty ranking. The authors suggest that the impact factor captures the recent influence of an average article, whereas a journal's standing builds over a longer history and a larger body of articles.^[6]

Journals also publish their own figures. The Journal of Marketing reported a 2014 impact factor of 3.938 and 14,741 total citations, and the Journal of the Academy of Marketing Science reported ranking 24th among 45 Financial Times list journals on its 2014 five-year impact factor, with a self-citation rate of 6.5%. Read these as editors' self-reports for a single year.^[12][13]

The h-index and open access

A unit has an h-index of h when h of its publications each have at least h citations. Five papers cited 14, 12, 3, 1 and 0 times give an h-index of 3. The index grows with output and has no size-independent counterpart, and a review of citation predictors calls it a problematic hybrid because it mixes productivity, citation impact and career age.^[5][10]

Metrics used as thresholds invite gaming. After Italy tied academic qualification and committee membership to citation and h-index thresholds, researchers there report a rise in self-citation and citation exchanges within groups, and one business and management journal reportedly gave authors detailed instructions to cite it ahead of a journal rating based on the Google Scholar h-index.^[7]

On open access, the review of citation predictors finds the evidence inconclusive. Open-access articles may be cited more because they are easier to reach, but comparisons are confounded by the many forms of open access, by journal-level differences and by authors' choices about which papers to make open.^[10]

How to read journal metrics

• Compare journals within the same field and the same database, and prefer field-normalised indicators when comparing across fields.^[5]
• Check which citation window and year a figure uses; the five-year impact factor exists because two years is too short for some fields.^[5]
• Look at more than one indicator, because total citations, averages and network-based measures can rank the same journals differently.^[3][6]
• Never transfer a journal's average to a single article or author; article-level citations are only a weak guide to quality even for the article itself.^[5][14]
• Treat figures that journals report about themselves as self-reports for a stated year.^[12][13]
• Remember that every number here measures influence among citing authors, not whether the findings are correct.^[15][11]

The evidence behind this

• Citation analysis: what citation counts do and don't tell you
• Marketing knowledge: how it is made and when to trust it

References

1. Eugene Garfield (1955). Citation Indexes for Science: A New Dimension in Documentation Through Association of Ideas. Science, 122(3159).
2. Eugene Garfield (1964). 'Science Citation Index' — A New Dimension in Indexing. Science 144(3619); reprinted in Essays of an Information Scientist, Vol. 7.
3. Eugene Garfield (1972). Citation Analysis as a Tool in Journal Evaluation. Science.
4. Eugene Garfield (2006). Commentary: Fifty years of citation indexing. International Journal of Epidemiology.
5. Ludo Waltman (2016). A review of the literature on citation impact indicators. Journal of Informetrics.
6. Hans Baumgartner, Rik Pieters (2003). The Structural Influence of Marketing Journals: A Citation Analysis of the Discipline and its Subareas over Time. Journal of Marketing.
7. Beniamino Cappelletti–Montano, Silvia Columbu, Stefano Montaldo, Monica Musio (2021). New perspectives in bibliometric indicators: Moving from citations to citing authors. Journal of Informetrics.
8. Joseph A. Cote, Siew Meng Leong, Jane Cote (1991). Assessing the Influence of Journal of Consumer Research: A Citation Analysis. Journal of Consumer Research.
9. Alonso Rodríguez-Navarro (2024). Distinct citation distributions complicate research evaluations. A single indicator that universally reveals research efficiency cannot be formulated. arXiv preprint 2407.09138.
10. Kayvan Kousha, Mike Thelwall (2023). Factors associating with or predicting more cited or higher quality journal articles: An Annual Review of Information Science and Technology (ARIST) paper. Journal of the Association for Information Science and Technology.
11. Michael R. Dougherty, Zachary Horne (2022). Citation counts and journal impact factors do not capture some indicators of research quality in the behavioural and brain sciences. Royal Society Open Science.
12. V. Kumar (2016). My Reflections on Publishing in Journal of Marketing. Journal of Marketing.
13. Robert W. Palmatier (2016). Editorial: the past, present, and future of JAMS. Journal of the Academy of Marketing Science.
14. Dag W. Aksnes, Fredrik Niclas Piro, Lone Wanderås Fossum (2023). Citation metrics covary with researchers’ assessments of the quality of their works. Quantitative Science Studies.
15. Mike Thelwall, Kayvan Kousha, Emma Stuart, Meiko Makita, Mahshid Abdoli, Paul Wilson, Jonathan Levitt (2023). In which fields are citations indicators of research quality?. Journal of the Association for Information Science and Technology.

Bibliographic details only. Follow the DOI to read each study at its publisher.