Our comments on Dion, Sumner, and Mitchell (DSM) (Reference Dion, Sumner and Mitchell2018) stem from our experience as journal coeditors—of Politics, Groups and Identities (PGI; Brown) and Comparative Political Studies (CPS; Samuels)—as well as from our work as cochairs of the Working Group on Publications of the American Political Science Association (APSA) Task Force on the Status of Women in the Profession. We have focused our work on attempting to explain Teele and Thelen’s (Reference Teele and Thelen2017) finding of a gender “publication gap,” that women are under-represented in top-ranked political science journals, in many cases well below their raw numbers in the discipline or the relevant subfield.
DSM show that the gender “citation gap” narrows when more women enter a subfield. They also find that a gender citation gap does not exist in papers written by women. These results echo Teele and Thelen’s finding that women are more likely to appear in journals that publish research that engage more women. In short, the gender citation gap and the gender publication gap are flip sides of the same coin.
DSM suggest that the gender citation gap is likely not due to deliberate discrimination but rather due to implicit biases and a lack of senior female scholars. In attempting to explain the gender publication gap, our own findings echo DSM’s conclusion that such “sociological” factors are important.
Senior male scholars edit most journals, and all editors (male and female) should entertain the possibility that implicit biases shape editorial decisions. To address this possibility, we asked the editors of four top journals (plus CPS) to conduct internal audits of their editorial processes: American Political Science Review (APSR), World Politics (WP), International Studies Quarterly (ISQ), and Political Behavior. For all five journals, it was found that conditional on submission, gender does not substantively shape editorial outcomes, whether for internal review (a “desk reject”) or a final decision. Factors that did matter include methodological approach (for example, at CPS, qualitative papers—by men or women—have a lower probability of being sent out for review or being accepted), rank of the most senior author (unsurprisingly, past career success predicts future success), and coauthorship: across journals, coauthored papers had a higher probability of success than solo-authored papers, regardless of the gender combination of the collaboration.
The journals’ reports will be published in a forthcoming symposium in PS: Political Science and Politics, which should nudge other editors to conduct similar internal audits. Yet perhaps obviously, if other editors also find that gender does not matter for editorial outcomes, we have no explanation for Teele and Thelen’s findings. Our results imply that women simply submit fewer articles—and submit fewer articles to particular journals. Why? For both questions, the answer lies with “sociological” factors, as DSM suggest.
Why do women simply submit fewer articles? For example, from 2007 to 2017, 39% of all submissions to the APSR were from a solo male author, while only 13% were from solo female authors. Many sociological factors are at work (Hengel Reference Hengel2017). A key factor in political science is the fact that women remain less likely to engage in collaborative research: another 27% of APSR submissions were from all-male collaborative teams, while only 3% were from all-female teams. That means 66% of all submissions to the APSR were from men, while only 16% were from women—the remaining 18% were from mixed-gender teams. Similar pictures emerge for the other journals included in our forthcoming symposium.
As DSM note, if men publish more, they have greater opportunities to capture future citations. This issue is, of course, broader. If men publish more, they will have greater career success, period. As noted above, collaboration is not only becoming more common; it predicts publishing success. To the extent that women are less likely to be invited onto a team or form collaborations themselves (perhaps because mentors do not think to encourage them to do so), they are less likely to expand their citation networks, less likely to get cited, and less likely to rise through the ranks. The dramatic under-representation of women in coauthorship networks perhaps explains why, as DSM note, the balance of citations does not immediately “adjust” when more women enter a subfield. Women are far more likely to collaborate with men—and when they do, gendered citation patterns persist, perhaps because the senior member of the collaboration is a male scholar, who makes the final call on various aspects of a paper before submission.
To shape the gender “submission gap,” those of us engaged in graduate training should think about ways in which we collaborate, and about whether and how we encourage PhD students to collaborate. Our discipline forces graduate students to complete a “solo-authored” project to obtain the PhD—but this may actually be counterproductive given the growing norm of coauthorship in the process of knowledge production.
The second important factor shaping why women submit to some journals more than others is the social perception of the journal. Together with Mala Htun and Brittany Ortiz (University of New Mexico), we surveyed APSA members in 2017 about their perceptions of different journals: whether they thought they would ever submit to that journal and whether they believed their work had a chance of appearing in that journal. Obviously, the most important factors shaping a journal’s profile are the types of papers it publishes, substantively and methodologically. Yet even controlling for those and other factors, we found statistically significant “gender perception gaps” for several journals.
All else equal, women say that they are simply less likely to submit to some journals—including Political Analysis. It is possible—as per DSM’s findings regarding gender citation gaps—that scholars perceive journals as more or less “open” to work by women depending on two key factors: the relative number of women (1) who are likely readers of the journal and who thus might serve as reviewers, and (2) who are on the journal’s editorial board. Not surprisingly, women were more likely than men to submit to PGI, for example, which has two female colead editors and eight women (of ten) on the editorial team. A journal’s social “image” matters for shaping the pool of submissions, which obviously affects the proportion of women who appear in the table of contents, and thus how frequently women are cited.
DSM confirm that “social” factors shape citation patterns, undermining women’s relative success among male and even mixed-gender coauthor teams. We agree that changing the status of women in the profession requires not just recruiting more female scholars but raising awareness that implicit biases are real, pervasive, and substantively important for the health of the discipline. Implicit biases undermine the marketplace of ideas, as they suggest that novel ideas and good research are simply ignored or do not get the attention they would get because a man did not offer them. They also shape career success—from graduate school through the job market to opportunities for promotion. The fact that many women succeed is not evidence that biases do not exist.
What can editors do to change norms and alter perceptions? For one, they can conduct an editorial audit and communicate the results with their boards, associated APSA section members, and readers—and take action if biases are found. Even if no biases are discovered, editors can take several actions to spark awareness of potential implicit biases, among authors and reviewers. First, they can limit self-citations, which can become gratuitous—and as DSM note, men are far more likely to self-cite gratuitously. At CPS, for example, we ask authors to limit self-citations to four. Second, journals can encourage authors to address the gender citation gap directly. ISQ’s online submission guidelines, for example, urge authors to pay attention to “overlooked authors and literatures,” while at CPS we include a similar statement in the email sent to authors of accepted papers. ISQ’s editors sometimes go further, suggesting citations for authors to include.
Third and perhaps most importantly for the issue of perception, journals should consider adopting a triple-blind review process. Triple blind means that editors do not know the author’s name, at least at the initial stage of the editorial process. ISQ, CPS, and WP all use triple-blind reviewing. There is no way, even for a journal with a relatively narrow focus, that editors can know the author of every submission, and removing an author’s name, rank, and institutional affiliation from the internal review eliminates the possibility—however remote—of implicit bias on the editors’ part. It also sends an important “social signal” to submitting authors that editors treat all papers equally.
The same logic holds for maintaining anonymity after internal review. We are aware that Political Analysis has recently adopted single-blind reviewing. In our opinion, this is a mistake, partly because we think editors should not know authors’ names, at least at the initial stage of the review process. To be sure, an author’s name is a Google search away. But just because reviewers can discover an author’s name is not reason to assume that they will do so. Even in the Google era, single-blind review sends a signal that the game is played by and for insiders. In our view, single-blind review sends the wrong social signal to junior scholars, scholars from “unlikely” institutions, and members of under-represented groups—particularly female authors, and particularly in research areas with few female scholars. It is true that in many areas of our discipline reviewers may be familiar with an author and his or her work. Yet given evidence of pervasive and substantively important implicit biases, editors should work to reduce such biases and send whatever signals they can that they are actively working to do so. Editors can shape the social perception of their journal—and should be aware that gender does affect the journal’s social perception. Journals that are seen as unreceptive to women’s scholarship are missing out on potentially highly impactful papers.
Individual scholars are likely unaware of the ways they participate in or contribute to the perpetuation of implicit biases. This does not mean they do not exist. Along with DSM, we hope that our findings contribute to a broader understanding of the extent of such biases, and help promote discussion about ways to eliminate them.
