Identifying single influential publications in a research field: New analysis opportunities of the CRExplorer

TL;DR

This paper introduces advanced methods for the CRExplorer tool, enabling the identification and analysis of influential references over extended periods and their impact sequences across citing years.

Contribution

The study presents new techniques for CRExplorer that analyze long-term influence and impact patterns of cited references in scientific literature.

Findings

Identified influential references across multiple decades.

Analyzed impact sequences of references over time.

Demonstrated methods using Scientometrics publication data.

Abstract

Reference Publication Year Spectroscopy (RPYS) has been developed for identifying the cited references (CRs) with the greatest influence in a given paper set (mostly sets of papers on certain topics or fields). The program CRExplorer (see www.crexplorer.net) was specifically developed by Thor, Marx, Leydesdorff, and Bornmann (2016a, 2016b) for applying RPYS to publication sets downloaded from Scopus or Web of Science. In this study, we present some advanced methods which have been newly developed for CRExplorer. These methods are able to identify and characterize the CRs which have been influential across a longer period (many citing years). The new methods are demonstrated in this study using all the papers published in Scientometrics between 1978 and 2016. The indicators N_TOP50, N_TOP25, and N_TOP10 can be used to identify those CRs which belong to the 50%, 25%, or 10% most frequently cited publications (CRs) over many citing publication years. In the Scientometrics dataset, for example, Lotka's (1926) paper on the distribution of scientific productivity belongs to the top 10% publications (CRs) in 36 citing years. Furthermore, the new version of CRExplorer analyzes the impact sequence of CRs across citing years. CRs can have below average (-), average (0), or above average (+) impact in citing years (whereby average is meant in the sense of expected values). The sequence (e.g. 00++---0--00) is used by the program to identify papers with typical impact distributions. For example, CRs can have early, but not late impact ("hot papers", e.g. +++---) or vice versa ("sleeping beauties", e.g. ---0000---++).