Quality and Quantity

, Volume 26, Issue 4, pp 337–365 | Cite as

A methodology for cluster analysis of citation histories

  • Vesna Luzar
  • Vesna Dobrić
  • Siniša Maričić
  • Greta Pifat
  • Jagoda Spaventi


Citations to all the papers (558) published from 1955 to 1964 by a multidisciplinary (natural sciences) research institute within a ‘scientific periphery’ were collected for the 11-year period after a 10-year lapse since the publication years. All the papers were grouped into 31 research topics, three of which had no such late citations at all. For the remaining 28 groups of papers three indicators were defined: ALPHA, the ratio of the number of papers with citations, to the number of all papers of the particular research topic, indicating thus an overall CITATION EFFICACY; BETA, the ratio of the sum of all citations, to the number of the cited papers, indicating CITATION INTENSITY, and GAMMA, expressing the CITATION LONGEVITY for a given research topic as the incidence (number) of cited papers (irrespective of the number of citations) within the 11-year citing period. In addition, three normalized transformations of the indicator BETA were checked. Two-dimensional (without ALPHA) and three dimensional (with ALPHA, GAMMA, and one of the BETA variants) graphical representations together with a pairwise correlation analysis served as preliminary guidance in the latter statistical analyses by (a) Ward's Hierarchical Cluster Analysis and (b) Polar Taxons Analysis. Both of them resulted in good agreement. Thus, the 28 research topics were found to belong to three clusters. Their composition varied slightly for the original BETA and its three normalized values. It was concluded that ALPHA is not a redundant but quite useful indicator, and that one of the normalized BETA-variants appears most suitable for this kind of studies. In the three-dimensional graphs the clusters may be related to scientific merit as judged in a composite way by all the three indicators. This is done with regard to the diagonal joining the indicators' maximal with their minimal values. The citation LONGEVITY (GAMMA) appears to be most important. Cluster stability tests showed fluctuations of few research topics, which was related to their specific features within the given research setting. It emerges on the whole that the research merit of this (peripheral) scientific production is determined neither by the journals status the papers were published in, nor by the authors' institutional status. Rather, it is the very scientific quality of individual papers within a given research topic that is decisive for the citation ‘survival’.


Research Topic Hierarchical Cluster Analysis Beta Variant Cluster Stability Polar Taxon 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Calinski, T. and Harabasz, J. (1974). A dendrite method for cluster analysis. Communications in Statistics 3: 1–27.Google Scholar
  2. 2.
    Cattell, R.B. (1966). Handbook of Multivariate Experimental Psychology, Chicago, Rand McNally.Google Scholar
  3. 3.
    Ferligoj, A., Maričić, S., Pifat, G. and Spaventi, J. (1988). Cluster analysis of citation histories from an institutional setting. Scientia Yugoslavica 14(3–4): 159–169. (Reprinted from Information Research (1988). Neva Tudor-Šilović and Ivan Mihel (eds.), by kind permition of the publisher-Taylor Graham, London, with a documentary addition of the most outstandingly cited paper of G. Alaga.)Google Scholar
  4. 4.
    Gray, H.L. and Schucany, I.P. (1972). The Generalized Jackknife Statistics. M. Dekker, Inc. New York.Google Scholar
  5. 5.
    Jain, A.K. and Dubes, R.C. (1988). Algorithms for Clustering Data. Prentice Hall. New Jersey.Google Scholar
  6. 6.
    Kaiser, H.F. (1961). A note on Guttman's lower bound for the number of common factors. British Journal of Statistical Psychology 14(1): 1.PubMedGoogle Scholar
  7. 7.
    Lindsey, D. (1978). The corrected quality ratio: a composite index of scientific contribution to knowledge. Social Studies of Science 8: 349–54.Google Scholar
  8. 8.
    Momirović, K., Zakrajšek, E. Hošek, A. and Stojanović, M. (1979). Comparative evaluation of some taxonomic algorithms for the determination of morphological types. Collegium Anthropologicum 3(1): 59–65.Google Scholar
  9. 9.
    Narin, F. (1978). Objectivity versus relevance in studies of scientific advance. Scientometrics 1(1), 35–41.Google Scholar
  10. 10.
    Nederhof, A.J. (1988). The validity and reliability of evaluation of scholarly performance, pp. 193–228 in Handbook of Quantitative Studies of Science and Technology (ed. Van Raan, A.F.J.). North Holland. Amsterdam-New York-Oxford-Tokyo.Google Scholar
  11. 11.
    Price, D.J. de S. (1965). Networks of scientific papers. The pattern of bibliographic references indicates the nature of the scientific research front. Science 149(3683): 510–515.Google Scholar
  12. 12.
    Rohlf, F.J. and Fisher, D.R. (1968). Tests for hierarchical structure in random data sets. Systematic Zoology 17: 407–412.Google Scholar
  13. 13.
    Rousseeuw, P.J. and van Zomeren, B.C. (1990). Unmasking multivariate outliers and leverage points. Journal of the American Statistical Association 85: 633–639.Google Scholar
  14. 14.
    SAS/IML (1988). User's guide release6.03 edition. SAS Institute Inc. Cary, North Carolina.Google Scholar
  15. 15.
    SAS/STAT (1988). User's guide release6.03 edition. SAS Institute Inc. Cary, North Carolina.Google Scholar
  16. 16.
    Štalec, J. and Momirović, K. (1983). Some properties of a very simple model for robust regression analysis. pp. 453–461 in Proceedings of the 5th International Symposium ‘Computer at the University’. Cavtat, University Computing Centre (SRCE). Zagreb.Google Scholar
  17. 17.
    Zakrajšek, E., Štalec, J. and Momirović, K. (1974). SS — a programming system for multivariate data analysis (in Croatian), pp. C8: 1–16 in Proceedings of the 1st International Symposium ‘Computer at the University’, Cavtat, University Computing Centre, Zagreb.Google Scholar
  18. 18.
    Ziman, J.M. (1985). Pushing back frontiers — or redrawing maps! pp. 1–12 in The Identification of Progress in Learning (ed. T., Hagerstrand). Cambridge University Press, Cambridge.Google Scholar
  19. 19.
    Zlobec, L. (1975). Comparative evaluation of some taxonomic algorithms. Master of Science Thesis, University of Zagreb (in Croatian).Google Scholar
  20. 20.
    Ward, J. H. (1963). Hierarchical grouping to optimise an objective function. Journal of the American Statistical Association 58, 236–244.Google Scholar

Copyright information

© Kluwer Academic Publishers 1992

Authors and Affiliations

  • Vesna Luzar
    • 1
  • Vesna Dobrić
    • 1
  • Siniša Maričić
    • 2
  • Greta Pifat
    • 3
  • Jagoda Spaventi
    • 4
  1. 1.University Computing CentreZagrebCroatia
  2. 2.Medical Science Studies UnitThe Medical SchoolZagrebCroatia
  3. 3.Rudjer Bošković InstituteZagrebCroatia
  4. 4.Institute of Information SciencesZagrebCroatia

Personalised recommendations