Recent changes to SNIP & SJR metrics

Published Mar 19, 2013 in the Newsletter Issue: Research Data Management — 2013

The Scopus Journal Analyzer’s unique functionality provides you with five graphical representations of the journals.

The multidisciplinary landscape of journal evaluation

In recent years, computational advances have contributed to acceleration in the field of bibliometrics. While the journal evaluation landscape was previously somewhat characterized by a scarcity of measures, now many journal metrics are available, providing a varied and more integral picture of journal impact.1 Librarians may find these useful to compare journals in various systematic ways.

Scopus features two citation indicators to measure a journal’s impact: SNIP (Source Normalised Impact per Paper) and SJR (SCImago Journal Rank). These indicators use the citation data captured in the Scopus database to reveal two aspects of a journal’s impact:

  • SNIP takes into account the field in which a journal operates, smoothing differences between field-specific properties such as the number of citations per paper, the amount of indexed literature, and the speed of the publication process.
  • SJR takes into account the prestige of the citing journal: Citations are weighted depending on whether they come from a journal with a high or low SJR.


These two indicators use a three-year window, are freely available on the Web2 and are calculated for all journals indexed in the Scopus database. The metrics have article type consistency: Only citations to and from scholarly papers are considered.

In October 2012, changes were introduced to both metrics to make them more intuitive and easily understandable. Following these improvements, the values are now computed and released once a year in the summer.

SNIP: How does it work?

SNIP was developed by Henk Moed, who was then part of the CWTS bibliometrics group at the University of Leiden. It is a ratio, with a numerator and a denominator. SNIP’s numerator gives a journal’s raw impact per paper (RIP). This is simply the average number of citations received in a particular year by papers published in the journal during the three preceding years.

SNIP’s denominator is the Database Citation Potential (DCP). Because there are large differences in the frequency with which authors cite papers between various scientific subfields, the DCP indicates a journal’s citation potential in the subject field it covers.

SNIP is RIP divided by DCP.

As of October 2012, the following changes apply:

  • A different averaging procedure is used to calculate the denominator, to reduce the impact of outliers.
  • A correction factor now weights citations from journals with low numbers of references.
  • The new calculation results in a SNIP average score for all journals in Scopus to approximately equal one.


“SNIP allows the impact of journals to be compared across fields in a fair way,” comments Ludo Waltman, a researcher at the Centre for Science and Technology Studies of Leiden University, “and has been updated following the most recent insights in the fields of bibliometrics and scientometrics. The recent changes ensure the most balanced treatment of journals from different fields, with minimal implications for users.”

SJR: SCImago Journal Rank

SJR was developed by the SCImago research group from the University of Granada, dedicated to information analysis, representation and retrieval by means of visualization techniques.

SJR looks at the prestige of a journal, as indicated by considering the sources of citations to it, rather than its popularity as measured simply by counting all citations equally. Each citation received by a journal is assigned a weight based on the SJR of the citing journal.

As of October 2012, the following changes apply:

  • A heavier weighting of the more prestigious citations that come from within or closely related fields
  • A compensating factor to overcome the decrease of prestige scores over time as the number of journals increases
  • A more readily understandable scoring scale with an average of one
     

References

1 Bollen J, Van de Sompel H, Hagberg A, Chute R (2009) A Principal Component Analysis of 39 Scientific Impact Measures. PLoS ONE 4(6): e6022. doi:10.1371/journal.pone.0006022

2 http://www.journalmetrics.com/

About the Author

Sarah Huggett

Publishing Information Manager

Elsevier

Oxford, GB

Comments