Citation Profiles and Research Dynamics

Abstract

What can we learn about research dynamics from ‘citation profiles’, a free product offered by Google Scholar? In this contribution we inspect citation profiles for a number of scientometrics scholars. The selected profiles range from highly prestigious scholars to moderate and lower stature colleagues, and from active and highly productive to inactive or former colleagues. Author citation profiles are compared to a theoretically expected pattern, based on scientometrics theories about publication delay and the immediacy effect in actively growing research fronts. The found profiles show a general increase of author citation levels from the millennium onwards, pointing to raised interest and activity in the field. Contrary to expectations based on the immediacy effect, we find prolonged high citation scores of no-longer-productive authors. Comparison of the author citation levels shows a lasting dominance of the founding fathers of the field. Explaining these patterns, requires both focused and explorative elements in the selective attention of citing authors in a community of researchers. The elements combining in citation dynamics are threefold: (1) fast consumption of novel contributions targeted to audiences in research fronts, explaining immediacy; (2) ritual referencing to basic ideas and methods, as paradigm focal points, explaining prolonged high citation; and (3) explorative searching, explaining (renewed) citation of earlier or remote work.

Appendix to Citation Profiles and Research Dynamics: Immediacy and Citation Functions, an Example

In this appendix, we inspect an example article for the relation between immediacy and the three formulated theoretical functions of citation: (1) connecting to a research front; (2) ritual citation to basic paradigm concepts and/or its prestigious authors; and (3) explorative citations to earlier or remote work to gain novel research ideas. Immediacy is defined as the percentage of citations to the most recent earlier work (0–4 years old) that researchers build upon in their current research.

Example: High immediacy and front connection.

As an example we analyse an article selected at GS from http://gut.bmj.com/ on 27 May 2019. With high immediacy, the research is expected to be highly focused and containing mainly ‘bricks’ as references. The article is published in Endoscopy news, open access: Ebigbo A, et al. Gut 2018;0:1–3. https://doi.org/10.1136/gutjnl-2018-317573. Below we discuss its references in context. Here below an extract of the example article.

Immediacy

With eight out of nine of the cited references aging within three years since publication, the immediacy, 89%, is very high. Only one cited reference is not immediate, dating from 2010.

Citation functions

The first three references, the first being the only ‘older’ one, are cited together legitimizing the research effort: “The incidence of BE (Barret’s oesophagus) and EAC (early oesophageal adenocarcinoma) in the West is rising significantly, and because of its close association with the metabolic syndrome this trend is expected to continue. 1–3”. The first, and only non-immediate cited reference, from 2010, receiving 135 citations by the end of 2018 (GS), reviews the evidence of rising incidence from clinical trials, meta-analyses, and large cohort and case-control studies, and points to the importance of early detection of oesophageal cancer, outlines strategies for prevention and describes features of oesophageal cancer to assist generalists in diagnosis. This reference is thus not to basic theoretical concepts, but underpins, together with references 2 and 3, the societal relevance of the study by Ebigbo et al., relating to the grants received from the Alexander von Humboldt Foundation and the Deutsche Forschungsgemeinschaft. Pointing to the study’s societal relevance, the prospect of improving medical practice, is a ritual citation function, as it does not directly bear to the content of the study, but affirms and reassures it’s wider goals.

References 4–7, given together, point to earlier use of the technique the group works upon to improve by computer aided learning next to handcrafted learning: “Reports of CAD (computer-aided diagnosis) in BE analysis have used mainly handcrafted features based on texture and colour”. Reference 8 then specifies databases used to improving aided learning technique used by the authors: “… to train and test a CAD system on the basis of a deep convolutional neural net (CNN) with a residual net (ResNet) architecture.” Finally, reference 9 points to their recent earlier study now further worked upon: “In this manuscript, we extend on our prior study on CNN in BE analysis.” Their research goal, improving computer aided diagnosis, i.c. “endosopic assessment of BE”, and its wider relevance, “enhancing patient management”, are directed at medical practice, not medical theory. Thus, this article with high immediacy has six out of nine references that are built upon to improve cancer diagnosis techniques, thus ‘bricks’, and three out of nine legitimizing the effort by pointing to scientific evidence of the rising incidence of BE and ACE, a ritual function.

Discussion and Conclusion

In this example article, immediacy of 89% goes together with 67% bricks and 33% ritual citations. The ritual function here relates both to theoretical elements of the paradigm (evidence of rising incidence), and its societal relevance (improve medical practice). As the ritual function is performed also by two most recent articles, it follows that 100% immediacy would not exclude ritual citations. Whether, even at the fastest growing research fronts, paying ritual dues to scientific or societal relevance, though less efficient, is an inevitable requirement, is a question for further examination.

Reference

Ebigbo, A., et al. (2018). Computer-aided diagnosis using deep learning in the evaluation of early oesophagealo adenocarcinoma. Gut 2018;0:1–3. https://doi.org/10.1136/gutjnl-2018-317573