Gender Inequality and Self-Publication are Common Among Academic Editors

The paper examines the gender composition of academic editors and their self-publication rates, focusing on the interplay between elite status in academia and gender bias. It highlights the influence of editors in shaping the scientific landscape through publication decisions, controlling access to prestige and recognition. Historically, women have been marginalized in academia, facing systemic barriers to advancement, particularly in achieving elite status. While existing research indicates underrepresentation of women on editorial boards, it lacks longitudinal data spanning multiple disciplines and critical comparisons (editors vs. other scientists, editors vs. colleagues, pre- and post-editorship publication patterns, and gender differences in self-publication). This study aims to address these gaps using a large-scale dataset from Elsevier.

The authors review existing literature that documents the underrepresentation of women in academic editorial roles across various disciplines. They note fragmented evidence and limitations in existing studies, such as limited longitudinal data, lack of cross-disciplinary comparisons, and the absence of comparisons between editors and other researchers. Studies focusing on specific disciplines or countries are cited, highlighting the need for a broader, longitudinal perspective encompassing diverse fields.

The study utilized data from Elsevier, the world's largest academic publisher, analyzing over 173,000 editorial pages from 1,167 journals across 15 disciplines. Information on editors' affiliations, disciplines, and tenure was extracted. Gender was inferred using a state-of-the-art classifier (Genderize.io) with a confidence threshold of ≥90%, resulting in a dataset of 81,000 editors and 4,700 editors-in-chief. Publication records were linked using Microsoft Academic Graph (MAG), enabling analysis of 20,000 editors and 1,600 editors-in-chief. The analysis compared editors' characteristics (citation counts, paper counts, h-index, collaborator count, affiliation rank) before editorship began with randomly selected matched scientists to assess their pre-existing elite status. Longitudinal analyses tracked changes in these characteristics over four decades. Gender disparities were analyzed across disciplines, comparing the proportion of female editors with that of female scientists. The study employed randomized baseline models to assess the contribution of productivity and impact to the gender gap. Self-publication rates were calculated for editors over a five-year period after assuming their role, with comparisons made to matched scientists and colleagues. Statistical tests, such as Welch's T-tests and Fisher's exact tests, were used to assess statistical significance. Regression models examined the temporal trends of self-publication rates, controlling for gender and journal fixed effects. The study meticulously addresses potential biases in its datasets.

The study reveals a persistent and significant gender gap in academic editorship. Women are substantially underrepresented among editors (14%) and editors-in-chief (8%), a disparity that remained stable over five decades. This gap exists across all disciplines except Sociology. While career length explains some of the gender gap among editors, it does not account for the disparity among editors-in-chief, suggesting other non-meritocratic factors are at play. Editors, particularly editors-in-chief, exhibit high self-publication rates; 12% of editors publish at least one-fifth of their papers in their own journals, and 6% publish at least one-third. The self-publication rate of editors-in-chief is strongly correlated with that of their editorial board. Compared with their matched scientists, editors publish significantly more in their own journals after assuming their roles. Although men and women are equally represented among those with the highest self-publication rates, men are overrepresented among those with the highest number of self-published papers. Regression analysis shows that men experience a larger increase in self-publication rates after becoming editors compared to women. Extreme cases, where editors publish a vast majority of their work in their own journals, are examined, revealing that this practice can persist for several decades and is predominantly a male phenomenon. Extreme journals, where editors authored a disproportionately large fraction of the published papers, were also identified.

The findings highlight persistent gender inequality in academic editorship despite efforts to promote inclusivity. The significant underrepresentation of women among editors-in-chief, even after controlling for productivity and impact, points to systematic biases and barriers faced by women in attaining elite scientific positions. The high self-publication rates among editors, particularly editors-in-chief, raise concerns about potential conflicts of interest and the need for increased transparency in the publishing process. The study emphasizes the potential for exploitation of power by editors and the need for stricter guidelines and regulations.

This study provides a large-scale analysis of gender inequality and self-publication among academic editors, revealing persistent challenges to gender equity and ethical concerns regarding self-publication practices. The findings underscore the importance of fostering a more inclusive and transparent culture of scientific publishing. Future research could focus on exploring the underlying mechanisms of gender bias in editor selection, examining self-publication practices in other publishers, and developing more robust methods for identifying and mitigating potential conflicts of interest.

The study's reliance on observational data limits causal inference. Although matching and randomized baseline models were employed, the exact mechanisms behind gender inequality and self-publication patterns remain unclear. The data is drawn from Elsevier journals only, limiting the generalizability of findings to other publishers. Gender inference using algorithmic tools introduces potential inaccuracies, although measures were taken to minimize this.