Skip to main content
SpringerLink
Log in

Mapping co-word structures: A comparison of multidimensional scaling and leximappe

  • Published:
Scientometrics Aims and scope Submit manuscript

Abstract

The LEXIMAPPE method and Multidimensional Scaling (MDS) are discussed as methods to visualize (‘map’) characteristics of structures of word-occurrence (‘co-word’) relations. Utilization of MDS is proposed as an alternative mapping method able to circumvent problematic features of LEXIMAPPE maps of the total co-word structure. A comparison of both methods on the same ‘real-life’ co-word matrix demonstrates topological advantages of an extended MDS-mapping.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

References

  1. W. TURNER, G. CHARTON, B. MICHLET,Describing science and technological problem networks using manually and automatically indexed full-text databases: some co-word analysis techniques, Workshop on scientific and Technology Indicators in the Higher Educational Sector, OECD, Report no. DSTI/SPR/85.24/08B, Paris, 1985.

  2. M. CALLON, J.-P. COURTIAL, W. A. TURNER, S. BRAIN, From translations to problematic networks: An introduction to co-word analysis,Social Science Information, 22 (1983) 191.

    Google Scholar 

  3. A. RIP, J.-P. COURTIAL, Co-word maps of science of biotechnology: an example of cognitive scientometrics,Scientometrics, 6 (1984) 381.

    Google Scholar 

  4. A. RIP, Mapping of science: Possibilities and limitations, In:Handbook of Quantitative Studies of Science and Technology, A. F. J. Van RAAN (Ed.), Amsterdam, Elsevier Science Publishers, 1988.

    Google Scholar 

  5. It is argued by M. CALLON, J. LAW, A. RIP (Eds),Mapping the dynamics of Science and Technology, London, MacMillan, 1986, (p. 120) that: “...many peripheral words subsumed directly under a central pole with no intervening structure of mediators — it is often the inclusion maps that are uninformative, while the proximity maps may be richly patterned and can reveal a number of densely connected clusters linked to each other, often by tenuous connections.”

    Google Scholar 

  6. A. RIP, J.-P. COURTIAL p. 385) stress that, “The arrangement in a two-dimensional plane is arbitrary, and chosen for ease of reading the map only.”

    Google Scholar 

  7. Recently a MDS-based spatial location program has been created by the French co-word group [B. MICHELET, L'analyse des associations, Research Report CDST/CNRS (Appendix B), October 1987], although without metric properties. An application of the metric mapping method Correspondence analysis to co-word data is metnioned in J.-P. COURTIAL, Technical issues and developments in methodology, In: M.CALLON, J. LAW, A. RIP, (Eds),Mapping the Dynamics of Science and Technology, London, MacMillan, 1986. However, this application proved to be unsatisfactory, primarily due to the very limited recovery (‘explanation’) of the total of distances between co-words in a low-dimensional space.

  8. MDS-methods have been used in bibliometric studies for various purposes: for example, in displaying relations between nations for publishing and citing patterns H. INHABER, M. ALVO, World science as an input-output system,Scientometrics, 1 (1978) 43., mapping of the cognitive and social structure of scientific specialties via co-citations between core authors or papers [H. G. SMALL, A. co-citation model of a scientific specialty: A longitudinal study of collagen research,Social Studies of Science, 7 (1977) 139], depicting the structure of author co-nominations [P. LENK, Mapping of fields based on nominations,Journal of the American Society for Information Science, 34 (1983) 115], or the structural representation of journal similarities [E. NOMA, An improved method for analyzing square scientometric transation matrices,Scientometrics, 4 (1982) 297].

    Google Scholar 

  9. A perfect reproduction of the proximities with metric MDS yields a minimum stress value equal to zero, whilst the maximum value is approximately equal to 1–2 (c.f. J. C. LINGOES, E. E. ROSKAM, A mathematical and empirical analysis of two multidimensional scaling algorithms,Psychometrika, 3, Part 2, monograph supplement, 1973). The maximum stress value for nonmetric MDS has been subject to various simulation studies, [e.g. J. De LEEUW, I. STOOP, Upper bounds for Kruskal's STRESS,Psychometrika, 49 (1984) 391]. An analytic expression to compute the stress for nonmetric MDS for 10<N<60 and 1<D<5, for randomly generated datasets is presented in I. SPENCE, A simple approximation for random ranking stress values,Multivariate Behavioral Research, 14 (1979) 355. The expression is derived from data gathered via a Monte Carlo experiment performed with the MDS program M-SPACE. Directly comparable results are expected from highly comparable programs such as KYST, MINNISA-I and SMACOF. Spence's ‘random’ nonmetric stress-value is equal to 0.36 for a D=2 dimensional solution with 37 analysis units.

  10. J. De LEEUW, J. J. MEULMAN, A special jackknife for multidimensional scaling,Journal of Classification, 3 (1986) 977.

    Google Scholar 

  11. G. YOUNG, A. S. HOUSEHOLDER, Discussion of a set of points of their mutual distances,Psychometrika, 3 (1938) 19.

    Google Scholar 

  12. W. S. TORGERSON, Multidimensional scaling: theory and method,Psychometrika, 17 (1952) 401.

    Google Scholar 

  13. R. N. SHEPARD, The analysis of proximities: Multidimensional scaling with a unknown distance function,Psychometrika, 27 (1962) 125.

    Google Scholar 

  14. J. B. KRUSKAL, Multidimensional scaling by optimizing goodness-of-fit to a non-metric hypothesis,Psychometrika, 29 (1964a) 1.

    Google Scholar 

  15. J. B. KRUSKAL, Non-metric multidimensional scaling: a numerical method,Psychometrika, 29 (1964b) 115.

    Google Scholar 

  16. M. CALLON, J.-P. COURTIAL, W. A. TURNER, “L'Action Concertée Chimie Macromoléculaire: Sociologique d'une Agence de Traduction”. Centre Sociologie de l'Innovation Paris. Group d'Etude sur la Recherche Scientifique, Université Louis Pasteur, Strasbourg, 1979.

    Google Scholar 

  17. H. G. SMALL, E. GARFIELD, The geography of science: disciplinary and national mappings,Journal of Information Science, 11 (1985), 147.

    Google Scholar 

  18. J. LAW, J. WHITTAKER, J.-P. COURTIAL, Mapping Acidification Research: A Test of the Co-Word Method, Paper presented at the European Workshop on The Relation Between Qualitative. Theory and Scientometric Methods in Science and Technology Studies, Amsterdam, The Netherlands, 1987.

  19. The matrix is (partly) presented in A. RIP, J.-P. COURTIAL,Op. cit. note 3. An extended 39*39 matrix was given in A. RIP,Scientometric Studies of Biotechnology, Paper presented to the Conference of the European Association for the Study of Science and Technology, Deutschlandsberg, Austria, 1982. The latter matrix was used as input for our MDS analysis.

  20. I. STOOP, J. De LEEUW,How to use SMACOF-1B, Department of Data Theory, University of Leiden, 1982.

  21. In general, the location of the most frequently occurring words depends on the homogeneity of the data, i.e., if more foci of research interest are found in a co-word matrix, MDS will yield the corresponding words in ‘clusters’ located towards the periphery. A multitude of ‘clusters’ will be positioned in the periphery in a circle-like fashion [e.g., H. P. F. PETERS, D. HARTMANN, A. F. J. Van RAAN, Monitoring advances in chemical technology, In:Informetrics 87/88, L. EGGHE, R. ROUSSEAU (Eds.), Elsevier Science Publishers, Amsterdam, 1988].

    Google Scholar 

Download references

Author information

Authors and Affiliations

  1. Science Studies Unit-LISBON Institute, University of Leiden, Stationsplein 242, 2312 AR, Leiden, (The Netherlands)

    R. J. W. Tijssen & A. F. J. Van Raan

Authors
  1. R. J. W. Tijssen
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. A. F. J. Van Raan
    View author publications

    You can also search for this author in PubMed Google Scholar

Rights and permissions

Reprints and permissions

About this article

Cite this article

Tijssen, R.J.W., Van Raan, A.F.J. Mapping co-word structures: A comparison of multidimensional scaling and leximappe. Scientometrics 15, 283–295 (1989). https://doi.org/10.1007/BF02017203

Download citation

  • Received:

  • Issue Date:

  • DOI: https://doi.org/10.1007/BF02017203

Keywords

Search

Navigation