Visualization of JOV abstracts

Abstract

Since the abstract can be found at the beginning of most scientific articles and is an essential part of the article, several attempts have been made to explore the rhetorical moves of abstracts in various research fields. These studies dealt only with accepted articles since they can be easily accessed. Although the findings of such works have some pedagogical implications for academic writing courses for young researchers who are relatively new to their fields, they do not contribute enough to the transparency of the peer review processes conducted in research fields. Increasing transparency requires considering rejected articles since they help to clarify the decision criteria in the peer review. Based on 591 abstracts of accepted or rejected articles submitted to Journal of Visualization (JOV), the present study aimed at exploring the differences between the accepted and rejected abstracts. The results show that there are significant differences in the structures of the abstracts. Since we also successfully develop a classification model for the decision using a machine-learning technique, the findings of this study have some implications for developing a semi-automatic reviewing system that can reduce the reviewer’s burden and increase the review quality.