Skip to Main Content

Evaluating Journal Quality — Metrics

A guide to the tools and methods for evaluating the quality of scholarly journals. Intended for authors, editors and reviewers.

Overview of Metrics Tools

Journal metrics illustrate the impact of a journal on its field. Metrics do not translate into "quality" but they are indicative of importance. They should always be used in conjuction with other evaluative tools to determine a journal's quality.

Journal metrics have been around since the arrival of the Citation Indexes in the 1970's. Currently there are three recognized and respected ranking tools for journals: the Journal Citation Reports from ISI which includes the impact factor; Scimago from Elsevier which includes the Scientific Journal Rank (SJR); the CWTS Journal Indicators includes the SNIP, and the Eigenfactor Metrics which include both the Eigenfactor and the Article Influence number. Each metrics tool is described in this page.

Journal Citation Reports

Journal Citation Reports (JCR) is a subscription database from the makers of Web of Science, the Institute for Scientific Information (ISI). The JCR provides rankings for journals included in the Web of Science database and includes journals in science and social science areas, and to a lesser degree in Arts & Humanities. The University of Connecticut subscribes to this database.


JCR offers subject rankings (ecology, political science, women's studies, etc.) The rankings can be sorted by total cites, impact factor and Eigenfactor. It can be customized to include cited half-life, immediacy index and other indicators as well. Individual journals can also be profiled by a variety of data between 1997 and the present.

  • The impact factor is a measure of how often the average article in a journal is cited by another article, based on the previous 2 years of data.   A 2015 journal impact factor of 1.863 means that the average article in this journal was cited an average of 1.863 times in 2014 and 2013 combined, by other articles. It measures an approximation of the mean citation rate per citable item.
  • The half-life states the point at which half the publications (either citing or cited) were published more recently than that date. Therefore a cited half-life of 10 means half the publications cited by articles in the journal were published more recently than 10 years ago. It measures the "shelf life" of articles - how long they tend to be cited. A low cited half-life suggests that citation activity peaks and drops off quickly.
  • The immediacy index states the average number of times an article is cited in the year it is published. Therefore an immediacy index of 0.344 means that in the year 2015 the articles in that journal were cited an average of 0.344 times in 2015. It measures the speed of information citing for a title. It favors journals which publish regularly or early in the year.

CWTS Journal Indicators

Information is available by discipline, with journal titles ranked within  that discipline. Searching by individual titles is also an option  but the context is best seen discipline-wide. CWTS Journal Indicators offers SNIP indicators (which stands for "source normalized impact per paper") which correct for differences in citation rates between scholarly disciplines and journal sizes. For instance, although the field of Mathematics has much lower citation rates than the field of Molecular Biology, the importance of specific journals in different fields can be compared because the SNIP adjusts according to citation practices in each field. 

CWTS Journal Indicators is freely available from Leiden University's Centre for Science & Technology Studies (CWTS). The statistics are based on Scopus database citation information.  It provides rankings for science, social science, and arts & humanities titles, but journal title coverage is weighted towards the sciences, as is Scopus.

  • Also offered are search parameters by year, source type (journals, book series, conference proceedings) and journal size
  • The stability of the indicator over time is also available, represented by a horizontal bar - the wider the bar, the less stable the indicator
  • Rankings can be sorted by number of publications (P), raw impact per publication (IPP), SNIP, and % self citing
  • Also includes books, conference proceedings and trade journals, but scholarly journals can be sorted specifically


Scimago Journal & Country Rank

Scimago is a freely available journal ranking product from the University of Granada in  the publisher Elsevier. It provides rankings for journals included in Elsevier's Scopus database and includes science, social science, and arts & humanities titles. Journal title coverage is weighted towards the sciences.


Scimago offers subject rankings  at two levels of specificity: the broader areas (chemistry, social sciences) and the more specific categories (ceramics & composites or education.) The rankings can be sorted by a wide variety of statistics, including Scimago Journal Ranking (SJR), H-index, and citable documents. Individual journals can also be profiled by a variety of data between 1999 and the present.

  • The Scimago Journal Ranking (SJR) measure the influence of a journal and is analogous to a journal impact factor. The SJR algorithm includes both the number of citations and the prestige of the journals from which the citations come. So of two journals holding the same number of citations from other journals, the journal containing citations coming from more highly ranked journals will have a higher SJR. The intent is to look at both the number and the quality of citations. It is very similar to the Eigenfactor scoring model.
  • The H-Index ranking is available from many sources, including Scimago. The H-index measures the number of citations against the number of articles (of a journal, an author, etc.) Thus a journal with an H-index of 100 contains at least 100 articles which have each been cited at least one times. It is intended to be a measure of both productivity and citation impact.
  • Citable documents are the number of documents that can be cited in a journal within a given time period (generally 3 years.) Information is also avaiable for total documents, and for self cites and references per document.

Scimago also has visualization options in the Viz Tools segment and country ranking information as well.



Eigenfactor is a freely available journal ranking product from individuals at the University of Washington., beginning in 2007. The Eigenfactor site was established to evaluate the influence of scholarly journals and map the structure of academic research. It covers the journals in the Web of Science database and so includes science, social science and to a lesser degree arts & humanities journals. It seems to lag about a year behind the other two journal ranking products so that, in September 2017 it is reporting 2015 statistics rather than 2016 statistics.


The Eigenfactor subject categories are drawn from the JCR subject categories but do not exactly duplicate them, as each journal belongs to only one category. Rankings of journals within each category are available. Journal specific rankings are also available. Three ranking scores are offered at this site: the Eigenfactor score, the Normalized Eigenfactor Score and the Article Influence score.

  • The Eigenfactor (EF) score is intended to measure a journal's total importance.  The scores are scaled so that the rankings of all journals listed will equal 100. If a journals has an EF of 1 then it has 1% of the total influence of all the indexed publications for a given year. That would be a very high score. The size of a journal will influence its score, with larger journals having higher scores.
  • The Normalized EF (NEF) score rescales the score so that it is easier to interpret. In this score the average journal has a score of 1, so that a journal with an NEF of 3 has three times the influence of a journal with a NEF of 1.
  • The Article Influence score measures tte average influence of the articles in a journal over 5 years. It is considered comparable to the JCR's Impact Factor. This score has also been normalized so that the mean AI score is 1.0.