A data-driven comparative analysis on citation intents across scientific disciplines
- 1Department of Computer Engineering, Chung-Ang University, Seoul, Republic of Korea
Abstract
The goal of this paper is to conduct a first comparative analysis on citation intents across multiple scientific disciplines. To this end, we conducted a data-driven investigation into the characteristics of citation intents utilizing a large-scale dataset of 17,717 academic papers and 2,345,674 citation contexts from 21 Web of Science subject categories. We analyzed five key characteristics: i) inter-category distribution of citation intents, ii) average number of citation intents per article, iii) age distribution of citation intents, iv) similarity of the age distributions of citation intent, and v) predictability of the number of citation intents. We found out that there are significant differences in all five characteristics across different research disciplines. Specifically, Philosophy, Sociology, Area, and History prioritize Background and Motivation intents while exhibiting minimal engagement with Uses and Extends. Conversely, Computer Science AI represents a unique case, characterized by a dominant prevalence of Uses and Extends intents alongside the least frequent use of background-oriented citations. Furthermore, the age distributions of citation intents exhibit the lowest similarity in Area, and History. In terms of predictability, Biotechnology and Plant Science exhibit the highest correlation levels in citation intent counts, whereas Sociology, Philosophy, and History yield the lowest correlation coefficients. Our comparative analysis not only uncovers cross-disciplinary variations and establishes a standardized framework for the comparative analysis of citation intents, but also has applications in advancing the normalization of scholarly indicators, facilitating trend detection, and supporting curriculum design for specific scientific disciplines.
Related
- Methodological lineages of concept analysis and mapping of reporting elements in nursing and health-related disciplines: A scoping review protocol
- PaECTER: Patent-level representation learning using citation-informed transformers
- Incentivizing Open Science: Examining the Open Data Citation Advantage in Visual Search Research
- ProLiVis 2.0: Literature-Centric Visualization of Protein--Protein Interaction Networks, with a Citation-Trust Model for Interaction Evidence
- Building better citation practices in public health research: Practical considerations for student researchers, from source discovery to crafting and positioning claims
BibTeX
@article{phan2026datadriven,
title = {{A data-driven comparative analysis on citation intents across scientific disciplines}},
author = {Tuan Anh Phan and Seohyun Nam and Jason J. Jung},
journal = {PLOS One},
year = {2026},
month = sep,
doi = {10.1371/journal.pone.0358327},
url = {https://doi.org/10.1371/journal.pone.0358327},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - A data-driven comparative analysis on citation intents across scientific disciplines
AU - Tuan Anh Phan
AU - Seohyun Nam
AU - Jason J. Jung
PY - 2026
DA - 2026/09/18
JO - PLOS One
DO - 10.1371/journal.pone.0358327
UR - https://doi.org/10.1371/journal.pone.0358327
AB - The goal of this paper is to conduct a first comparative analysis on citation intents across multiple scientific disciplines. To this end, we conducted a data-driven investigation into the characteristics of citation intents utilizing a large-scale dataset of 17,717 academic papers and 2,345,674 citation contexts from 21 Web of Science subject categories. We analyzed five key characteristics: i) inter-category distribution of citation intents, ii) average number of citation intents per article, iii) age distribution of citation intents, iv) similarity of the age distributions of citation intent, and v) predictability of the number of citation intents. We found out that there are significant differences in all five characteristics across different research disciplines. Specifically, Philosophy, Sociology, Area, and History prioritize Background and Motivation intents while exhibiting minimal engagement with Uses and Extends. Conversely, Computer Science AI represents a unique case, characterized by a dominant prevalence of Uses and Extends intents alongside the least frequent use of background-oriented citations. Furthermore, the age distributions of citation intents exhibit the lowest similarity in Area, and History. In terms of predictability, Biotechnology and Plant Science exhibit the highest correlation levels in citation intent counts, whereas Sociology, Philosophy, and History yield the lowest correlation coefficients. Our comparative analysis not only uncovers cross-disciplinary variations and establishes a standardized framework for the comparative analysis of citation intents, but also has applications in advancing the normalization of scholarly indicators, facilitating trend detection, and supporting curriculum design for specific scientific disciplines.
SN - 1932-6203
ER -