Published in: P. Šipka (Ed.), Journal Publishing in Developing, Transition and Emerging Countries: Proceedings of the 5th Belgrade International Open Access Conference 2012, Belgrade, Serbia, May 18-19, 2012, pp. 25-35, Belgrade: Centre for Evaluation in Education and Science
Journal classifications based on citation data: The comparison and suitability of three distance measures
University of Novi Sad, Faculty of Philosophy, Dpt. of Psychology
Abstract: In the study presented in this paper, citation data from the Serbian Citation Index were used to calculate three types of proximity measures among Serbian scientific journals in the fields of social sciences and humanities. The measures were based on the frequency of intercitations among journals, journal co-citation counts, and bibliographic journal coupling. The clustering solutions derived from different distance matrices were compared and validated against the current subject classification of national journals. All three solutions were generally compatible with the list of major disciplines suggested by the Serbian Ministry of Education and Science, but the results indicated the need for a more precise subject classification. The most accurate journal classification has been achieved by using the bibliographic journal coupling method. Bibliographic journal coupling has primarily produced clusters of thematically similar journals, while intercitation and co-citation clusters also revealed some types of relationships not necessarily determined by subject similarity. The visualization of proximity data provided additional insight into the relations among disciplines and the status of several multidisciplinary journals. In general, the presented results indicate that the citation data are suitable not solely for journal classification tasks, but also as a mean of scientific domain analysis at the journal level. in certain aspects.
Keywords: journal mapping; journal clusters; co-citation analysis; bibliographic coupling