Investigating the Journal Impact along the Columns and Rows of the Publication-Citation Matrix

Fang Hui

DOI: https://doi.org/10.1007/s11192-020-03715-y

IF: 3.801

2020-01-01

Scientometrics

Abstract:Journal impact factors and diachronic journal impact factors are currently calculated with the data along the rows and columns of the publication-citation matrix of a journal, respectively. The average publication-citation matrix can be obtained by dividing the elements of the publication-citation matrix by the number of papers published by a journal in a given year. Along the rows and columns of the publication-citation matrix, we found that journals in the same subject category can have quite different citation patterns. In particular, some journals have a prolonged impact. To effectively reflect the impact of individual journals with different citation patterns on the scientific community, we propose an integral synchronic journal impact factor that combines the features of the existing journal impact factors and diachronic journal impact factors. This approach utilizes the data along the rows of the publication-citation matrix and the average citations among the papers published in individual years. The length of the citation window can be flexibly set to balance accuracy and timeliness based on citations. Modifications of the proposed indicator considering normalization, the importance of citation sources and a geometric averaging mechanism are provided.