Overcoming the gender bias in ecology and evolution: is the double-anonymized peer review an effective pathway over time?

Introduction

Females represent half of the global population but are still underrepresented in most work fields, mainly in politics and economics (World Economic Forum, 2019). In the education field, despite recent progress (World Economic Forum, 2019), we still need to reach equity. For instance, females are still strongly underrepresented in STEM (science, technology, engineering, and mathematics) graduate courses, and research positions, where only 33% of researchers are females (Garcia-Holgado et al., 2020; UNESCO, 2021).

Gender bias in the academic environment has multiple dimensions and causes manifested in society, home, workplace, and individuals (Leite & Diele-Viegas, 2020). Girls are discouraged from pursuing a STEM career from an early age due to gender stereotypes, which prevents many from even considering such careers as possible paths (Leite & Diele-Viegas, 2020). Following this initial cultural barrier, females in STEM careers may face additional obstacles in their professional development, such as biases regarding their abilities, harassment, and discrimination (Pell, 1996; Greider et al., 2019; Huang et al., 2020). A central motivation for such barriers is the tendency of male scientists to support, collaborate with, hire, and mentor male scientists (Brashears, 2008; Moss-Racusin et al., 2012; Sheltzer & Smith, 2014), a feature of human behavior known as gender homophily. The principle of homophily states that people with similar characteristics, such as gender, race, ethnicity, age, socioeconomic background, and educational attainment, tend to interact between them more frequently than those with different characteristics (McPherson, Smith-Lovin & Cook, 2001). Therefore, gender homophily can potentially lead to decreased participation and competitiveness of female researchers in the academic environment, particularly in leadership positions (Sheltzer & Smith, 2014). Indeed, female researchers represent only 12% of members of national science academies and 33.3% of STEM researchers, and they also receive fewer grants (UNESCO, 2021). Evidence of gender bias has also been found in authorship patterns in STEM scientific publications (Fox, Ritchey & Paine, 2018; Holman, Stuart-Fox & Hauser, 2018; Salerno et al., 2019).

A gender gap of 27% in total productivity and females representing less than 30% of authors indicate the pervasiveness of gender bias in scientific publications (West et al., 2013; Larivière et al., 2013; Huang et al., 2020). Specifically, in biological science publications, less than 40% of articles have females as first authors, and less than 30% have females as last (i.e., senior) authors (Bendels et al., 2018; Fox, Ritchey & Paine, 2018; Salerno et al., 2019). In the field of ecology and evolution (EcoEvo), despite submitted articles with female authorships being sent to peer review at similar rates to articles with male authorships, females are significantly underrepresented as first, senior, and sole authors (Fox & Paine, 2019). Gender bias is also evident in corresponding authors, as articles with a female corresponding author have lower probabilities of a positive outcome, and female first authors tend to defer corresponding authorship to a coauthor (Fox, Ritchey & Paine, 2018; Edwards, Schroeder & Dugdale, 2019; Fox & Paine, 2019). Additionally, the proportion of female co-authors in EcoEvo is higher when the senior author is female (Fox, Ritchey & Paine, 2018). As a result, only 11% of the top-publishing authors in this field are females (Maas et al., 2021).

Publishing articles is the primary way researchers can receive recognition for their contributions and advance their careers through grants and academic positions (Fox, Ritchey & Paine, 2018). The underrepresentation of females in authorship patterns may therefore reflect early withdrawals of females from the academic community and also highlight the potential for implicit biases that favor males (Sidhu et al., 2009; Martin, 2012; Salerno et al., 2019). Indeed, female scientists have a 19.5% higher risk per year of leaving academia, and their academic careers tend to be shorter than those of their male counterparts (Huang et al., 2020). This results in a significant cumulative advantage for males in publishing over time (Huang et al., 2020). To address this issue, strategies to reduce the early dropout of female scientists from their academic careers and to decrease gender biases in the peer review process are crucial for increasing diversity, inclusion, and representativeness in scientific publications (Huang et al., 2020; Diele-Viegas et al., 2021). Thus, those strategies will ultimately prevent the underuse of valuable skills in the pool of females and other minority scientists, such as Latin Americans (Sheltzer & Smith, 2014; Valenzuela-Toro & Viglino, 2021). One possible solution to avoid explicit bias in article publication is implementing a double-anonymized peer review process.

In the double-anonymized peer review process (hereafter: DA peer review), author and reviewer identities are concealed. This peer review has been proposed as an efficient strategy to reduce biases against gender, institutions, country of origin, new ideas, authors’ prestige, and young scientists (Stensrud & Brooks, 2005; Mainguy, Motamedi & Mietchen, 2005; Smit, 2006). Furthermore, DA peer review is seen as a more equitable peer review system (Smit, 2006) that promotes fairness by creating a system concealing identities and forces reviewing focused on content (Wenneras & Wold, 1997; Stensrud & Brooks, 2005; Darling, 2015). However, the implementation of DA is still not widespread among scientific journals (Haffar, Bazerbachi & Murad, 2019). DA peer review has concerns related to the additional journal workload and the possibility of reviewers identifying authors or institutions. In addition, the effectiveness of DA in reducing gender bias in publishing remains controversial. While in the fields of medicine and economics, no significant differences have been found between DA and SA peer review systems, with female authors achieving similar publication rates under both scenarios (Blank, 1991; Cho et al., 1998; Justice et al., 1998; Mahajan et al., 2021), DA has been found to effectively reduce gender bias in computer science (Tomkins, Zhang & Heavlin, 2017). In the EcoEvo field, DA has positively impacted the number of female authors in journals such as behavioral ecology (Budden et al., 2008), but this increase in female authorship may be attributed to the overall female representation (i.e., relative number and proportion of publications) in the field over time (Holman, Stuart-Fox & Hauser, 2018; Salerno et al., 2019), rather than the peer-review process per se (Webb, O’Hara & Freckleton, 2008). For example, the biological conservation journal utilized a SA peer review and did not find evidence of gender bias in publication rates (Primack et al., 2009). Additionally, the proportion of female authors did not differ between the DA and SA journals of ecology and ornithology (Cox & Montgomerie, 2019). While these studies suggest comparable levels of female authorship in both DA and SA journals, the analyses were based on a few journals, which may not accurately reflect the overall publishing landscape in the field of EcoEvo.

Here, using a comprehensive dataset from 18 journals encompassing different thematic areas within ecology and evolution (e.g., animal behavior, applied ecology, behavioral ecology, botany, conservation, ecosystem sustainability, evo-devo, freshwater ecology, and zoology), we investigate the impact of the DA peer-reviewing process on the gender authorship patterns of EcoEvo scientific journals. Specifically, we hypothesize that the frequency of publications with female-leading authorships is higher under a DA peer-reviewing policy when compared to the SA policy (H1). Moreover, we hypothesize that adopting the DA peer-review policy has increased the frequency of female-leading articles on time in journals that changed their peer-review process (H2).

Methods

Results

We obtained articles from the 18 selected EcoEvo journals (Table 1). After excluding the editorials (N = 41), our dataset resulted in 2,231 analyzed articles (Appendix S1). Overall, males led the majority of articles in EcoEvo journals, regardless of the peer-review policy and especially in the senior author position (Fig. 1). Only 24% of senior and 40% of the first authors were women in all analyzed articles.

The gender ratio difference cannot be attributed to the peer-review policy of EcoEvo journals. We did not find a significant difference between DA and SA peer-review policies concerning women publications as first or senior authors (Tables 2 and 3, respectively). Also, we did not find a significant difference between the “pre” and “post” periods (Fig. 2) in journals that switched their peer-review policy from SA to DA regarding both first and senior gender authorship (Tables 4 and 5, respectively).

Discussion

Using a comprehensive dataset of 2,231 articles from 18 journals encompassing different thematic areas of ecology and evolution, we showed that despite the increasing efforts to avoid gender imbalance in science, it is still pervasive in the EcoEvo field. We found that females were not likely to publish more in DA peer-reviewed journals (H1) and adopting the DA peer-review policy did not increase the representativeness of female researchers as first or senior authors over time in EcoEvo journals (H2).

Adopting a DA peer-review process is usually considered a possible solution to reduce the gender gap in academic publications (Budden et al., 2008). However, our results showed that this policy alone is not reducing the gender gap in EcoEvo journals. In fact, our analyses showed that a SA policy produced a slight effect on reducing gender imbalance.

Although the nine journals included here as DA or switched-review claim to follow a DA peer-review process, their actual implementation is not guaranteed. For example, the DA is optional in The American Naturalist and Animal Behaviour, meaning that the authors can opt between hiding or showing their identities to the reviewers. In the Journal of Evolutionary Biology, the reviewers’ names can be shared with authors if they sign their reviews. Mammal Review indicates in their authors’ guideline section that papers with conflicts of interest may not be suitable for DA, while Behavioral Ecology informs that the author’s identity is “as much as possible” kept from reviewers. These uncertainties highlight the difficulties in actually implementing the DA peer review, which might be affecting our results.

Preparing and reviewing a DA paper demands both authors’ and reviewers’ specific care to ensure their anonymity (Cox & Montgomerie, 2019). Reviewers were able to identify the authors in 40% of the performed DA reviews in medicine journals (Cho et al., 1998; Justice et al., 1998), and 10–26% in computer science journals (Le Goues et al., 2018).While adopting a DA policy could reduce nepotism, and geographic, institutional, and gender biases (Budden et al., 2008; Cox & Montgomerie, 2019; Fox & Paine, 2019), such improvements are probably linked to a well-implemented DA that guarantees a masked review (Cho et al., 1998; Darling, 2015).

The lack of gender equality in the authorship of papers published in DA journals could also indicate that conscious and unconscious gender biases are still pervasive in the peer-review process, which may be related to the gender bias in editorial boards of EcoEvo journals (Liévano-Latorre et al., 2020). Female editors represent less than 30% of editorial boards in academic journals (Liu et al., 2023), and this trend is similar in ecology and conservation journals (Cho et al., 1998; Liévano-Latorre et al., 2020; Sperotto et al., 2021). As male scientists support other male scientists (Moss-Racusin et al., 2012), editors could bias the peer review process by rejecting submitted articles with female authorships (Brodie et al., 2021). On the other hand, female editors tend to accept more women-authored articles (de Barros et al., 2021). In this sense, double-anonymization could reflect the structural bias and inequalities in research institutions and universities, where tenured positions are biased towards men (Sheltzer & Smith, 2014; Holman, Stuart-Fox & Hauser, 2018; Diele-Viegas et al., 2021). Homophily and gender equity suggest that the inclusion of more women in leadership roles could improve the gender balance (Amrein et al., 2011; Mauleón et al., 2013; Cho et al., 2014; de Barros et al., 2021).

Journals should move towards a better implementation of the DA peer review aiming to fight intrinsic biases of the peer-review process, but the adoption of the DA review system alone could be insufficient in fostering gender equality in scientific publications. In this sense, before adopting the DA peer-review policy, the first step to promoting gender equality in EcoEvo publications is promoting gender-equitableness in these editorial boards (Liévano-Latorre et al., 2020). In addition, supporting affirmative actions focused on the entry and permanence of minority groups in science, diversifying decision-makers, and dismantling mechanisms that allow sexism in academic environments are deep actions that would reduce gender bias in science (Diele-Viegas et al., 2021). Furthermore, more transparency in the peer-reviewing process is needed. One alternative is using an open peer-review system, which could promote reviews more accurate and courteous, reducing biases and hostile or abusive comments (Walsh et al., 2000; Ford, 2013). Another alternative could be a triple-anonymized peer-review process, where editors, reviewers, and authors are anonymized (Brodie et al., 2021; Conklin & Singh, 2022). However, both alternatives would demand adaptations and changes for journals and researchers (Walsh et al., 2000; Brodie et al., 2021).

The underrepresentation of women in leadership positions also emerges as a potential explanation for the minority of females as leading authors in scientific articles, especially as the senior author as evidenced here. Even considering that female participation in STEM and specifically in biological sciences has been increasing in the last years, females are still underrepresented in senior positions (National Science Foundation, 2018; European Commission, 2019). Female scientists receive less support in academia and harsher reviews in scientific publications, which, together with recurrent sexual and moral gender-based harassment, prevents the retention and advancement of females in STEM careers (Leaper & Starr, 2019; Greider et al., 2019; Diele-Viegas et al., 2021). Lack of support for female scientists is also evident in the funding and grant allocation, as women have a lower probability of obtaining a grant than men, which also contributes to lower rates of female permanence in academia (Moss-Racusin et al., 2012; Wijnen, Massen & Kret, 2021).

White male cis-gender researchers from Global North still dominate the ecology area (Nuñez et al., 2021) and STEM as a whole, contributing to the lack of diversity in the academic environment (Maas et al., 2021). Meanwhile, underrepresented researchers (e.g., women, people of color, and LGBTQIA+ researchers) have different barriers to keeping and advancing in a scientific career. For instance, female researchers from Latin American countries deal with the intersection of sexism, colonialism, and racism (Bernal et al., 2019; Valenzuela-Toro & Viglino, 2021). Female Latin American scientists develop their careers in countries that invest less in STEM and present a culture that highlights male pride, besides hampering literature access and present language barriers (Bernal et al., 2019; Valenzuela-Toro & Viglino, 2021).

Besides structural challenges female researchers face in the academic environment, the COVID-19 outbreak added extra barriers to their maintenance in STEM fields. The pandemic has negatively affected the productivity, networking, community building, and well-being of women in STEM, especially mothers (Myers et al., 2020; Langin, 2021). In addition, it led to disrupted collaborations and pauses in the career progressions of female scientists, as they face challenges of remote work conflicting with caregiving responsibilities (Hipólito et al., 2020; Staniscuaski et al., 2020; Myers et al., 2020). Consequently, female scientists have been more isolated, losing contacts and publication chances, affecting their job stability and funding (Gabster et al., 2020; Myers et al., 2020).

All academic levels could promote a fair and equal academic environment by fostering diversity, inclusion, and affirmative measures, such as scholarships and research funding (Diele-Viegas et al., 2021; Maas et al., 2021), especially for mother researchers (Heidt, 2023). Furthermore, reforms in the education system, mentoring, and academic publishing are needed to reach equality in science (Holman, Stuart-Fox & Hauser, 2018). For instance, creating new evaluation metrics and implementing inclusive policies, such as encouraging gender equality in the editorial boards, could reduce the gender gap in STEM fields (Liévano-Latorre et al., 2020; Diele-Viegas et al., 2021; Sperotto et al., 2021).

Conclusions

Our results highlight that adopting the DA peer-review system alone could be insufficient in fostering gender equality in EcoEvo academic publications. Thus, we suggest the application of means to increase female representation on editorial boards and promotion of more transparency in the peer-review process, adopting strategies to improve the DA system, or exploring other alternatives, such as open peer-review or triple-anonymized review. Ecologists and evolutionists understand how diversity (expressed in diverse facets such as functional, genetic, phylogenetic, and taxonomic) is important to ecosystems’ resilience in facing environmental changes. The question remaining is: why is it so difficult to promote and keep this “diversity” in addition to equity and inclusion in the academic environment? We thus argue that all scientists, mentors, and research centers must be engaged in promoting solutions to gender bias by fostering diversity, inclusion, and affirmative measures.

Supplemental Information