Unveiling the Mechanisms Between Writing Style and Academic Impact: A Multidimensional and Machine Learning Perspective
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Writing style, as a key non-content factor influencing academic dissemination, remains underexplored in terms of its underlying mechanisms with scientific impact. Drawing on an explainable machine learning perspective, this study investigates the effects of multidimensional writing style features on citation behavior. 23,644 paper abstracts from the Web of Science (WoS) in the fields of computer science, management, and biomedicine were analyzed, extracting four categories of features: readability, emotional and subjective expression, lexical richness, and coherence of expression. These were integrated with 17 baseline indicators to construct an analytical framework. Through ablation experiments, feature importance analysis, and multiple regression, we systematically evaluated the independent contributions and directional effects of each style dimension. The results reveal that readability exerts the strongest predictive power among writing style dimensions and plays the most substantial role in citation behavior. Concise and clear expression significantly enhances citations. Additionally, positive emotional and subjective expressions contribute to improved dissemination outcomes, whereas overly complex vocabulary and excessive cohesion may exert negative effects. This study provides empirical evidence for understanding the relationship between writing style and academic impact, offering practical implications for scholarly writing, journal peer review, and research management.
Fields of Study & Subjects
scientometrics and bibliometrics research
Academic Writing and Publishing
Academic Publishing and Open Access
Publication details
- Journal / source
- Exploring Science Academic Conference Series
- Volume
- 14
- Pages
- 41–60
- Publication date
- 27 March 2026
- Language
- EN