Tagging tagging. Analysing user keywords in scientific bibliography management systems


  • Markus Heckner University of Regensburg
  • Susanne Mühlbacher University of Regensburg
  • Christian Wolff University of Regensburg


In this paper, an empirical study of tagging behaviour in web-based bibliographic annotation systems is presented. Starting from an initial category finding phase in which tags attributed to selected articles from Connotea were classified we have set up a category model for linguistic and functional aspects of tag usage as well as for the relationship between tags and document full text. In a second phase this model is applied to approx. 500 tagged articles from the information and computer technology domain randomly selected from Connotea. Our findings show significant differences to other tagging research which was primarily conducted using popular (non-scientific) tagging platforms like Flickr or Delicious. We observe a great overlap of tag material and document text and rather few non-content related tags. The comparison of user tags with author keywords shows that users tend to use less and more general tags. Finally, system functionality seems to play a role for users’ tagging behaviour.

Author Biographies

Markus Heckner, University of Regensburg

Doctoral student Media Educational Science University of Regensburg 93040 Regensburg, Bavaria, Germany

Susanne Mühlbacher, University of Regensburg

Doctoral student Information Science University of Regensburg 93040 Regensburg, Bavaria, Germany

Christian Wolff, University of Regensburg

Associate professor for media computing Media Computing University of Regensburg 93040 Regensburg