The invariant distribution of references in scientific articles

Abstract

The organization of scientific papers typically follows a standardized pattern, the well-known IMRaD structure (introduction, methods, results, and discussion). Using the full text of 45,000 papers published in the PLoS series of journals as a case study, this paper investigates, from the viewpoint of bibliometrics, how references are distributed along the structure of scientific papers as well as the age of these cited references. Once the sections of articles are realigned to follow the IMRaD sequence, the position of cited references along the text of articles is invariant across all PLoS journals, with the introduction and discussion accounting for most of the references. It also provides evidence that the age of cited references varies by section, with older references being found in the methods and more recent references in the discussion. These results provide insight into the different roles citations have in the scholarly communication process.

The narrative structure as a citation context in data papers: A preliminary analysis of Scientific Data

“…Second, citations are most concentrated in the Background and Summary section, with 36.5% of 949 sentences containing at least one citation, a proportion significantly higher compared to 16.3% across all sections. The latter observation is comparable to Bertin et al ()'s finding based on an examination of research papers across all Public Library of Science (PLoS) journals. This suggests that the Background and Summary section in data papers plays similar “stage setting” roles as the Introduction section in research papers (Swales, , pp.…”

Citations in Scientific Texts: Do Social Relations Matter?

“…(Hu et al, ) found that references cited in the first parts of an article have globally higher citation rates. (Bertin et al, ) found invariant schemes of distribution of citation in the position of citations, but (Boyack et al, ) found important differences across disciplines. The underlying hypothesis in this study and in the current paper is that the position of a given citation is influenced by its relative importance or function.…”

“…The IMRaD (Introduction, Methods, Results and Discussion) prototypical structure has been thoroughly described in the literature, as it is prescribed (and followed) for most science articles (Swales, ). Several recent studies of citation behaviour have focused on the section in which a citation appears, whether to study linguistic variations in the context (Bertin & Atanassova, ), the age of a citation (Bertin et al, ), or its function (Teufel, ). In order to explore this aspect, we manually tagged each section heading in our corpus according to the IMRaD categories.…”

“…These features (distribution, position, frequency, surrounding cues, etc.) are then confronted to external characteristics afferent to the cited reference in order to provide the base for an empirical study (Bertin et al, ) or to train an automatic classifier (Teufel et al, ). Most of the target external characteristics are either subjective (function, opinion or importance, assessed by an annotator or more rarely a posteriori by the author) or loosely related to the citation act (number of citations, age of the reference).…”

Mapping the linguistic context of citations

“…In order to do so, we processed the section titles using a set of regular expressions specifically designed for this task. After analyzing the density of citations along the four main section types, we showed that the distribution of references along the text progression is essentially invariant across the different PLOS journals []. This first result showed that a strong relation exists between the IMRaD structure and the use of citations in scientific writing.…”