Constructing bibliometric networks: A comparison between full and fractional counting (1607.02452v4)

Abstract: The analysis of bibliometric networks, such as co-authorship, bibliographic coupling, and co-citation networks, has received a considerable amount of attention. Much less attention has been paid to the construction of these networks. We point out that different approaches can be taken to construct a bibliometric network. Normally the full counting approach is used, but we propose an alternative fractional counting approach. The basic idea of the fractional counting approach is that each action, such as co-authoring or citing a publication, should have equal weight, regardless of for instance the number of authors, citations, or references of a publication. We present two empirical analyses in which the full and fractional counting approaches yield very different results. These analyses deal with co-authorship networks of universities and bibliographic coupling networks of journals. Based on theoretical considerations and on the empirical analyses, we conclude that for many purposes the fractional counting approach is preferable over the full counting one.

• The paper shows that fractional counting mitigates distortions from hyperauthorship and highly cited papers.
• It compares full and fractional methods using empirical analyses from university co-authorship and journal bibliographic coupling studies.
• The study proposes that fractional counting yields a more accurate representation of scientific collaboration despite increased computational complexity.

Fractional Versus Full Counting in Bibliometric Networks Construction

The paper "Constructing Bibliometric Networks: A Comparison Between Full and Fractional Counting" by Antonio Perianes-Rodriguez, Ludo Waltman, and Nees Jan van Eck, examines the methodologies used in constructing bibliometric networks. This research paper focuses on comparing the full counting approach—a traditional method widely used in bibliometric analyses—and the fractional counting approach, proposing the latter as a preferable method in many contexts.

Bibliometric networks, such as co-authorship, bibliographic coupling, and co-citation networks, play a crucial role in understanding scientific collaboration and intellectual linkages. However, the process of constructing these networks has received scant attention in the literature. This paper seeks to spotlight the construction step and bring to the fore the differences in results that various methods can yield.

Conceptual Understanding of Full and Fractional Counting

• Full Counting: Assigns a full weight of one to each action (e.g., co-authorship link, citation) regardless of the number of co-authors or citations.
• Fractional Counting: Distributes an equal weight for each action across co-authors or citations, thus normalizing the overall weight to one.

A critical issue addressed in the paper is how full and fractional counting affect the resultant weights in bibliometric networks. For instance, in constructing co-authorship networks, a publication co-authored by five researchers would impart a weight of one to each co-authorship link in full counting, while fractional counting would assign a weight of 1/5. Similarly, in bibliographic coupling, fractional counting treats every citation action equally, preventing distortions caused by highly cited papers.

Empirical Comparisons

Co-authorship Networks of Universities

The empirical analysis utilizing data from the Web of Science database and the VOSviewer software presents a clear differentiation between the two methods. The visualizations of co-authorship networks of 750 universities demonstrated significant differences:

• Full Counting: Results in a network where a few hyperauthorship publications (publications with many authors) dominate and distort the network structure.
• Fractional Counting: Provides a clearer national grouping in the co-authorship network, suggesting that collaborations are more nationally constrained than the full counting method indicates.

Notably, when excluding publications with more than 20 co-authors, the full counting network began to resemble the fractional counting network, emphasizing that hyperauthorship has an outsized influence in the full counting approach.

Bibliographic Coupling Networks of Journals

In bibliographic coupling, the paper shows that full counting can result in misleadingly strong relationships based on a few highly cited papers. For example, the journal Acta Crystallographica Section E appeared strongly related to Scientometrics under full counting due to a small number of highly cited articles. Fractional counting corrected this by normalizing each citation, offering a more balanced measure of relatedness.

Implications and Future Directions

The findings underscore the practical implications of choosing a counting method. The fractional counting method is theoretically more robust for many applications, offering a truer reflection of co-authorship and citation relationships by ensuring each action contributes equally to the network. However, the computational complexity and difficulty in interpretation (e.g., non-integer link weights) present challenges.

Future Work: The paper suggests exploring more sophisticated fractionalization methods, such as incorporating the number of citing authors or researchers. Moreover, integrating the construction and similarity determination of bibliometric networks in combined frameworks could yield further insights and improve methodologies.

Conclusion

For experienced researchers in bibliometrics, this comparative paper elucidates the nuanced but critical impact that counting methods exert on bibliometric network construction. By preference, fractional counting avoids the disproportionate influence of hyperauthorship and highly cited papers, thereby facilitating a more accurate representation of scientific collaboration and intellectual linkages. However, the choice between methods should ultimately align with the analytical goals and the nature of the data at hand. This paper serves as a foundational step towards refining bibliometric methodologies, aiming for accuracy and insightful interpretations in scientific research evaluations.