Round Table Discussion

1
Colloque Evolution des publications
scientifiquesAcadémie des sciences, 14-15 mai
2007Pierre BraunsteinAcadémie des
sciencesInstitut de Chimie (UMR 7177
CNRS)Université Louis Pasteur - Strasbourg

• Round Table Discussion
• Bibliometric Indicators in Research Evaluation
  and Policy

2

• CHEMISTRY
• Core chemistry
• Numerous interfaces
• chemistry/mathematics
• chemistry/biology
• Increasing number of international collaborations
• The increasing number of journals leads to their
  selection according to some criteria (tradition,
  userfriendly submission process, rapidity, etc )
• The highest Impact Factor in chemistry is 20
  (review articles)
• The ranking of authors on a publication is highly
  variable

3

• Do impact factors reflect what is important or
  what is fashionable ?
• Areas in which much is published generate more
  citations and therefore higher impact factors
• The period covered (2 years) is too short to be
  really significant (compare e.g. an article
  published in December of year N vs. January of
  year N1.
• The ISI classification  General  vs.
   Specialized  journals is very far from being
  satisfactory. Any ranking based on this splitting
  is rather useless.
• Total citations counts (or H factor number of
  articles cited more than H times) provide a
  usefull image of the impact of the scientific
  contributions of an author, in particular a
  senior one (but remember the  cold fusion 
  effect!). However, since practices/numbers vary
  considerably from one discipline to another,
  comparisons become difficult/impossible at
  scientific interfaces.

4

• Indicators can be useful, they will be used
  anyway, we better understand their meaning,
  strengths and weaknesses in order to improve the
  system. A European model?
• A calibration system is required and various
  benchmarking procedures will minimize the
  misinterpretations
• Indicators can assist in identifying performances
  within a similar area, at the national and
  international levels, particularly for more
  senior scientists (ex. 1000 most cited chemists).
  They have no absolute value.
• Other criteria are needed to identify the
  younger, most promising scientists key role of
  the community!
• There must be feedback between the evaluators and
  the individuals or labs or institutions being
  evaluated to explain and communicate the source
  of indicators.
• This must be a multicriteria and transparent
  mechanism

5

• BIBLIOMETRICS
• BIBLIO MIXI CATORS