In contrast to previous issues, the following “Notes from the Editors” is the first in which we do not provide summaries of the articles published in this issue. This decision has nothing to do with the articles and letters themselves; rather, we desire to use the space that is provided to highlight and discuss different aspects of the challenges we face as editors of one of the leading outlets in our discipline. This continues our discussion from the previous “Notes from the Editors” on how we conserve and refresh our reviewer pool in times of continuously increasing submission rates. More generally, however, we wish to act in accordance with our commitment to increase the transparency of the review process, which we included in our mission statement when we became the editors of the APSR in 2016. With rising submissions, we hope to continue to use “Notes from the Editors” to briefly update our readership on current trends we witness in regard to submissions and the overall review process.

We aim to work in a transparent way for many reasons, but one important reason is that it helps us in detecting and avoiding bias, whether that bias is against groups of scholars or types of research. We know that bias can be toxic to researchers, institutions, and publications; thus, we try to reduce bias to avoid its implications for our scientific community. Because peer-reviewed publications are frequently the core component of standards for scholarly promotion, it is incumbent on us to reflect on the implications of bias, who and what subfields are most likely to be impaired by bias, and how we can provide our community with a fair review process. We want to highlight three aspects of our editorial decision-making that are part of our efforts to minimize potential bias and to provide for transparent and comprehensible decisions.

First, our manuscript managing system, called Editorial Manager, is not only a system to manage the approximate 1,400 manuscripts we receive annually as well as the subsequent 4,500 reviewers invited to write reports. It is also a powerful reporting tool that allows us to constantly analyze and improve our process. For instance, we can easily identify the manuscripts that have outstanding reviews of 60 or more days, making tracking down reviewers more effective. We track a variety of statistics regarding all stages of our review process—from first receipt to final decision. More importantly, the program facilitates the organization of the reviewer pool such that we receive ample reports that are informative and detailed. When we took over our editorial duties, we requested some updates to the Editorial Manager software as well as the database used by the APSA journals. For instance, we thought it would be more helpful if the indicators for the fields of research in the reviewer pool matched the authors’ areas of expertise. Having better-matched information about reviewers and authors allows the system to provide more accurate reviewer proposals for our editors.

Second, we changed the editorial decision-making procedure when we inherited the editorship. Because of the low acceptance rate at the APSR, most manuscripts are either desk rejected or rejected on the basis of an initial round of reviewer reports. While desk rejections concern the suitability of a manuscript (often with regard to our submission guidelines), report decisions interpret the recommendations of two to five reports. We carefully assess the reports to determine whether the authors will be able to address the concerns of the reviewers. Because biases are more likely to arise when editors interpret reports autonomously, our procedure follows a co-decision-making workflow. Together, a lead editor and a field-specific associate editor decide whether a manuscript is accepted or sent back to the author for revision and resubmission, and in cases of disagreement, there is a consultation with a third editor. Similar to the principle of a collective examination, this procedure offers a major advantage in calibrating individual editor scaling regarding the interpretation of reported reviewer recommendations.

Third, after a round of review, we provide both authors and reviewers some explanation for our decision. This often involves calling attention to particular criticisms given by one or more reviewers that we think present insurmountable barriers to a successful revision and subsequent publication. Inevitably, this effort of transparency may also lead to misunderstanding at times—for example, when authors feel a reviewer's recommendation was more positive than the editors’ interpretation upon assessing the reports. Nevertheless, we think that both authors and reviewers benefit from this transparency in how we are interpreting the reports.

However, even as we strive for more progress, it is nevertheless important to maintain a skeptical outlook of the peer-review process. This is why we require our editors to be transparent and scrutinize our editorial process, especially to be sensitive toward detecting any bias against a specific type of research, methodology, field, nationality, or authorship, among others. A topic that has recently received a lot of attention in academia is potential publication bias against female scholars. Accordingly, it did not come as a surprise that we, as editors of the APSR, were also approached with questions regarding whether publishing decisions at the APSR are biased against women at any point during the editorial process.

We have been attentive to concerns regarding gender in the editorial process since we took over as editors of this journal. Using the statistics from our editorial system, we had examined the role of gender in our editorial process even prior to the recent attention to the issue. We have presented these analyses in numerous ways: we collaborated with other journals to examine gender bias, we presented our findings in our last annual report, we held a panel at the 2017 APSA annual conference about gender in APSR, and there is an article in PS: Political Science and Politics on gender and editorial outcomes in the APSR (König and Ropers Reference König and Ropers2018). However, as we took over editorial duties in September 2016, most of these analyses were based primarily on submissions that were handled by previous editorial teams.

Figure 1 provides some insight into our overall decision-making with respect to gender by presenting a breakdown of APSR submissions with final decisions, disaggregated by gender. The figure only includes submissions handled solely by our team so that we can assess our own decision-making process. Note that the graph does not include submissions currently in the review process, it only includes manuscripts with a final accept or reject decision. The data are grouped by type of authorship; that is, whether authors are writing solo or as part of a team, as well as whether a team consists of single or mixed genders, resulting in five categories. The frequency of submission by gender is far from parity. Of these five types, we receive the largest number of submissions (35%) from solo male authors, followed by male teams (26%), mixed teams (22%), solo female authors (14%), and female teams (3%). Thus, 39% of submissions involve at least one female author and 83% involve at least one male author. If we focus only on manuscripts by individuals and single-sex teams, about 22% have female and 78% have male authors.

FIGURE 1.

Percentage share of final decision outcome by type of authorship for manuscripts for which our team made the first decision

The percentages in the brackets underneath the subplot titles denote the relative manuscript share.

At the desk rejection stage, the major predictor of success is whether the manuscript is solo- or team-authored. Forty-eight and a half percent of male and 46.2% of female solo-authored submissions are desk rejected. Team submissions experience a lower desk rejection rate (33.3% female, 33.7% mixed, and 34.4% male teams). We see a slightly different but similar pattern in terms of finale acceptance rates. In general, team-authored manuscripts are more successful, with higher success rates for single-sex teams (5.0% male, 5.3% female) compared to other authorship types (3% solo male, 2.8% solo female, and 2.7% mixed). We note that we cannot reject the null hypothesis that these five types of authorship will fail to predict acceptance rates in the long run (p = 0.27), whereas we can confidently reject the null hypothesis that there are no differences in desk rejection rates across authorship group types (p < 0.001). Detecting long-run differences in the very low rates of acceptance across different authorship types requires more data than our editorial team has yet produced, even though differences between 3% and 5% acceptance rates are very consequential for the composition of publications.

Admittedly, this figure presents only aggregated descriptive statistics. We do not have a good way of measuring whether the average quality of manuscripts by different authorship groups differs other than the review process itself. We are not the first to note the possibility that women might hold their potential submissions to a higher standard, in which case the similar rates of success for male and female authors and teams could still be consistent with some gender bias existing in the editorial and review process. As the overall submission rates indicate, we receive far more submissions from male than female authors, and very few from female teams. We think that the data do not suggest prima facie bias of the editorial process against female submissions, but we welcome suggestions of ways we might better analyze the possibility that self-selection in the submission process might mask problems in the editorial and review process.