A comparative study of first and all-author co-citation counting, and two different matrix generation approaches applied for author co-citation analyses

Jesper W. Schneider (), Birger Larsen and Peter Ingwersen
Additional contact information
Jesper W. Schneider: Royal School of Library and Information Science
Birger Larsen: Royal School of Library and Information Science
Peter Ingwersen: Royal School of Library and Information Science

Scientometrics, 2009, vol. 80, issue 1, No 7, 103-130

Abstract: Abstract Aim The present article contributes to the current methodological debate concerning author co-citation analyses. (ACA) The study compares two different units of analyses, i.e. first- versus inclusive all-author co-citation counting, as well as two different matrix generation approaches, i.e. a conventional multivariate and the so-called Drexel approach, in order to investigate their influence upon mapping results. The aim of the present study is therefore to provide more methodological awareness and empirical evidence concerning author co-citation studies. Method The study is based on structured XML documents extracted from the IEEE collection. These data allow the construction of ad-hoc citation indexes, which enables us to carry out the hitherto largest all-author co-citation study. Four ACA are made, combining the different units of analyses with the different matrix generation approaches. The results are evaluated quantitatively by means of multidimensional scaling, factor analysis, Procrustes and Mantel statistics. Results The results show that the inclusion of all cited authors can provide a better fit of data in two-dimensional mappings based on MDS, and that inclusive all-author co-citation counting may lead to stronger groupings in the maps. Further, the two matrix generation approaches produce maps that have some resemblances, but also many differences at the more detailed levels. The Drexel approach produces results that have noticeably lower stress values and are more concentrated into groupings. Finally, the study also demonstrates the importance of sparse matrices and their potential problems in connection with factor analysis. Conclusion We can confirm that inclusive all-ACA produce more coherent groupings of authors, whereas the present study cannot clearly confirm previous findings that first-ACA identifies more specialties, though some vague indication is given. Most crucially, strong evidence is given to the determining effect that matrix generation approaches have on the mapping of author co-citation data and thus the interpretation of such maps. Evidence is provided for the seemingly advantages of the Drexel approach.

Date: 2009
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (8)

Downloads: (external link)
http://link.springer.com/10.1007/s11192-007-2019-y Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:scient:v:80:y:2009:i:1:d:10.1007_s11192-007-2019-y

Ordering information: This journal article can be ordered from
http://www.springer.com/economics/journal/11192

DOI: 10.1007/s11192-007-2019-y

Access Statistics for this article

Scientometrics is currently edited by Wolfgang Glänzel

More articles in Scientometrics from Springer, Akadémiai Kiadó
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().