The Normalization of Co-authorship Networks in the Bibliometric Evaluation: The Government Stimulation Programs of China and Korea

TL;DR

This paper evaluates Sino-Korean research collaboration using bibliometric analysis, highlighting the importance of normalization techniques like fractional counting over traditional integer counting for accurate performance assessment.

Contribution

It introduces a comparative analysis of integer and fractional counting methods for co-authorship networks, improving evaluation accuracy of international research collaborations.

Findings

Fractional counting reveals paradoxical effects on publication and citation metrics.

Integer counting overestimates collaboration impact, making it less suitable for evaluation.

Both methods help identify key countries and institutions, but serve different research questions.

Abstract

Using co-authored publications between China and Korea in Web of Science (WoS) during the one-year period of 2014, we evaluate the government stimulation program for collaboration between China and Korea. In particular, we apply dual approaches, full integer vs. fractional counting, to collaborative publications in order to better examine both the patterns and contents of Sino-Korean collaboration networks in terms of individual countries and institutions. We first conduct a semi-automatic network analysis of Sino-Korean publications based on the full-integer counting method, and then compare our categorization with contextual rankings using the fractional technique; routines for fractional counting of WoS data are made available at http://www.leydesdorff.net/software/fraction . Increasing international collaboration leads paradoxically to lower numbers of publications and citations using fractional counting for performance measurement. However, integer counting is not an appropriate measure for the evaluation of the stimulation of collaborations. Both integer and fractional analytics can be used to identify important countries and institutions, but with other research questions.