Elsevier's Scopus.com upgraded the Journal Analyzer with Source Normalized Impact per Paper (SNIP), which measures a source's contextual impact, and SCImago Journal Rank (SJR), which measures the scientific prestige of scholarly sources.
These indicators will be applied to all journals indexed by Scopus and will be freely available to both subscribers and non-subscribers @ scopus.com and www.journalmetrics.com
2. “Using the Impact Factor alone to judge a journal is
like using weight alone to judge a person’s health.”
Source: The Joint Committee on Quantitative Assessment of Research: “Citation
Statistics”, a report from the International Mathematical Union
3. “There is no single ‘best’ indicator that could
accommodate all facets of the new reality of
Journal Metrics.”
- Wolfgang Glänzel, Centre for R&D Monitoring, ECOOM
Leuven, Belgium
, g
Bibliometrics – A discipline that uses statistical methods to analyze content and measure
research performance.
4. Users demand alternative to Impact Factor
Assign value to Scopus- - “How do impact factors of journals in Scopus, especially
ditor
unique content, especially
q , p y those not covered by Web of Science, compare to those
y , p
Ed
regional content indexed in WoS?” Lopez-Illescas et al, J Informetrics, 2008, p304-316
Journal quality is a key aspect - “Basically, I measure good research by papers. The better
of performance evaluation
f f l ti the
th papers, the b tt my research. So the I
th better h S th Impact F t
t Factor.
Scientist
Quite simple.” Group Leader, Netherlands
- “Impact Factor is considered one of the crucial citation
indicators widely used in research assessment and science
S
administration.” Pislyakov, 2008, Scientometrics p1-10
ublisher Librarian
Facilitates Scopus adoption if
p p - “It is a no-brainer – the only thing missing from Scopus to
y g g p
faced with switch from WoS really take over from ISI is legitimate journal metrics.” Inn Beng
Lee, Head of Customer Marketing, Asia Pacific
Increases willingness of other -“This will have a very positive effect on our relationship with
publishers to cooperate other publishers and spur collaboration to ensure full
coverage.” Ove Kähler, Senior Product Manager for Scopus Content,
Netherlands
N th l d
Pu
5. How would you know that all of these journals are
the best?
Journal Impact Factor 2008*
Lancet Infectious Diseases 13.165
Social St di
S i l Studies of S i
f Science 1.343
1 343
Dyes & Pigments 2.507
Expert Systems with Applications 2.596
2 596
Progress in Nuclear Magnetic Resonance
6.162
Spectroscopy
Communications on Pure & Applied Mathematics 3.806
Citation behaviour is different per field
• Frequency of publication
• Number ofCitation Reports 2009
*Journal
authors
•LLength of reference li
h f f list
How do you compare with confidence?
6. Our approach
Not one-dimensional, but multi-dimensional approach
Make metrics freely available
y
Don’t invent metrics ourselves – ask the external experts
Survey what’s out there
Meet experts to understand how metrics meet user needs
Calculate test data for metric shortlist
Decide with advice from the internal experts, Academic Relations
Result – endorse SJR and SNIP
Neither are ‘leading’ – both are given equal importance
Secure support from publishing divisions
7. 3-year citation window = fairest compromise
To which years are citations made in 2007 directed?
10
Immunology and Microbiology (67,9%)
9 Biochemistry, Genetics and Molecular Biology (65,3%)
Pharmacology, Toxicology and Pharmaceutics (63,6%)
General (58,3%)
8 Medicine (60,8%)
M di i (60 8%)
Neuroscience (64,8%)
Chemistry (55,4%)
7 Health Professions (56,4%)
Nursing (53,5%)
Chemical Engineering (51,6%)
6
Materials Science (52,6%)
Dentistry (55,9%)
5 Energy (42,1%)
Physics and Astronomy (47,3%)
Veterinary (47,9%)
4 Agricultural and Biological Sciences (45,0%)
Environmental Science (43,5%)
Engineering (38,7%)
3 Psychology (44,8%)
Earth and Planetary Sciences (41,3%)
Computer Science (33,7%)
2 Mathematics (32,5%)
Decision Sciences (32,4%)
Economics, Econometrics and Finance (29,0%)
1
Business, Management and Accounting (27,7%)
, g g( , )
Social Sciences (20,1%)
0 Arts and Humanities (10,1%)
2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996
Source: Felix de Moya
8. SJR and SNIP snippets…
SCImago Journal Rank – SJR
• Developed by SCImago – Felix de Moya SCImago research group
Moya,
Consejo Superior de Investigaciones
Científicas (CSIC), University of Granada,
• Main feature: prestige metric – similar to Google PageRank
• Citations are weighted depending on the status of the source they come from
• Secondary feature : different citation behaviour addressed
Source-Normalized
So rce Normali ed Impact per Paper – SNIP
• Developed by Henk Moed, CWTS, University of Leiden
• Main feature: contextual citation impact
• Part 1 – evens out differences in likelihood of citation (i.e. different behaviours)
• Part 2 – evens out differences in database coverage in subject fields
9. SCImago Journal Rank - SJR
SJR is a prestige metric – citations weighted depending on where they come from
- A journal transfers its prestige by citing
- Prestige transferred = journal’s SJR
e.g.
e g Lancet SJR 2007 = 1.541 – high prestige
1 541
e.g. Scandinavian Journal of Medicine and Science in Sports SJR 2007 = 0.153 – lower prestige
A journal’s prestige is shared equally between its citations
Life Sciences journal Arts & Humanities journal
Normalize for
differences in
• Hi h impact, llots of citations
High i t t f it ti citation behaviour
it ti b h i • Low impact, few of citations
• One citation = low value between subject fields • One citation = high value
10. Source Normalized Impact per Paper - SNIP
Calculate ‘Citation Potential’ for 2009
1. Collect papers citing 1-3 year old papers 2. Collect reference lists of citing papers
in target journal in 2009
Etc.
3. Count number of references in citing papers
to any (in any journal) 1-3 year old papers
4. Citation Potential = average number
of references to any 1-3 year old papers
Etc.
Normalize for
SNIP = 2009 Impact / 2009 Citation Potential differences in
• Life Sciences – high impact high Citation Potential
impact, citation behaviour
it ti b h i
• Arts & Humanities – low impact, low Citation Potential between subject fields
11. Metrics to Support User Needs
Need
N d SJR SNIP
Manipulation-proof
All journals have a metric
Fair to database coverage
Metrics
M i are current
Database transparency
Rank journals without Impact Factors
R kj l ith t I tF t
Metric transparency
Fair to ll bj t fields
F i t all subject fi ld
Flexible ranking
12. First Scopus, and then…
SJR and SNIP will be launched in Scopus on 23 January 2010
All 18,000 journals, proceedings and book series receive SJR and SNIP values
SJR and SNIP for 2009 go back to 1999
Refreshed twice per y
p year: April and September
p p
New titles will have SJR and SNIP in their second year of coverage
Freely available via the Scopus Info site
15. We are asking users to change their mindset
Launch 23 January 2010 All journals, book series, proceedings
All years, 1999-2009
y ,
Note – 2009 Impact Factor will be published June 2010
Metrics refresh 2010 September
2011 and ongoing April, September
Refresh what? All years recalculated to reflect retrospective changes in
dynamic database
Historical values archived on Info site
New titles rule: if data are available, metric will be calculated
If title covered from 2010, first metrics published in April 2011
Note – first Impact Factor for new launch would be 2011, published June 2012,
2011 2012
Note – first Impact Factor for newly indexed but established title would be 2012,
published June 2013
16. Information available
Access all SNIP & SJR values @
www.journalmetrics.com
www journalmetrics com
Additional Resources:
• Research Trends Special Issue on Journal
performance Indicators [www.researchtrends.com]
(p t and online
(print a d o e at )
• Press Release
• Scopus.com
• Info Site Scopus [ http://info.scopus.com ]