Social and environmental justice issues were the theses of my college applications. As a ethnically mixed woman who graduated with a BA in geology, diversity has always been at the forefront of my mind. Becoming a data analyst at the AGU provided me the opportunity to delve into an aspect of this important issue (especially in 2016, ugh): gender bias in the sciences. I have since returned to the world of groundwater hydrology at the University of Utah, where these sorts of issues will inevitably impact my career. Furthermore, as a young researcher just starting to participate as an author and reviewer, I’ll be living this study…
In my recent publication with Brooks Hanson, we highlight STEM’s gender gap between published first authors (experts) and those reviewing papers as experts in the geosciences. This has been studied before, but these other studies have mostly either assigned gender to first names (This can be problematic…for example, could you simply guess the gender of either author of this paper?), smaller sample sizes, and have not accounted for age. Accounting for age helps reveal some important but otherwise hidden differences.
The American Geophysical Union (AGU), as both a membership and publishing organization, can observe the demographics of people interacting with our publishing system. For context, women are 28% of membership- similar to employed US STEM people. Understanding and measuring our own inclusivity makes us a smarter and more equitable community. I liked this project especially as it illuminates a source of exclusivity that isn’t normally considered in one’s day to day life, but shows a mechanism of how small differences in our decisions impact the careers of entire groups over time.
Measuring the inclusivity of reviewing manuscripts, specifically, is important to the community and our science in a few ways…
For the Individual, reviewing improves writing, and critical thinking skills as well as gives an insider’s look at the frontiers of science in their field. Reviewing also presents networking opportunities: this is important for career development, recognition, and future collaborations. The work is noted for membership society recognition, and can be acknowledged and credited with the advent of ORCID’s.
For Science, generally, inclusivity of views and diverse ways of thought are a primary driver for peer review- because it makes our science more rigorous. Different experiences lead to creative understanding of subject matter. Additionally, understanding and acknowledging our biases makes us, as scientists, better leaders (especially important in this political climate).
Figure: Reviewer distribution by age as compared to other scholarly activities.
What we found:
We were surprised to find that women have higher acceptance rates than men (61% vs. 57%, respectively). Below are some possible explanations. Note that this is not an exhaustive list and the truth is probably somewhere mixed up and in-between.
- Women submit less often than men (2.5 vs. 3.3 papers, respectively, per first author in the study period), and therefore may have more time and resources to dedicate to each manuscript.
- Women expect higher obstacles than their male peers (because of public rhetoric that everything for women in science is harder 1-6), leading to more work on each paper.
- Women are more conservative in submitting papers, needing to be “very certain” rather than a man’s “pretty certain”.
- (Using ‘Keep It Simple, Stupid’) women are smarter. The difference in acceptance rates should be even larger because women are still experiencing bias in the reviewing process, deflating their acceptance rates.
- Okay, yes. Perhaps reverse sexism exists.
However, I think it is more probable that women are just preparing more thoroughly. Women are taught that if “John” and “Jane” are equal, “John” is still seen as better. Wouldn’t they learn that they have to be better in order to be seen as equal?Given that women are less likely to be promoted, given raises and other career-advancing opportunities, we were primed to expect what we found; Women don’t review papers (20% of reviewers) as often as you might expect from their participation rates as first authors, all authors, or members (27, 23, and 28%, respectively). This is not just that reviewers are from older age cohorts, which are male dominated, but a cumulative effect from all age cohorts.The gap mostly stems from both authors and editors not suggesting or inviting women. Things to consider:
- Do you stop to think about whether you’re representing your field equitably when filling out reviewer suggestion boxes?
- Do you pick reviewers that you’ve heard about in a positive light?
- Who of your colleagues do you praise the most?
- Who first comes to mind when you think of “genius”? “Hard worker”? Think about these implicit associations.
While overall women do not decline more often than men, they do within each age cohort. We think that the overall non-difference would prevent any negative feedback loops, but the age-related differences prompted us to think about a couple of things… Why is this happening? We looked at people’s decline responses, but overwhelmingly the response from both genders for review declines was that they were too busy. This isn’t really useful in parsing out gender-related differences, so we have to make an educated guess...
- Are women being asking to be on more committees and expected to do more outreach and non-academic work?
- Are women shouldering more household/family duties?
- Are women falling prey to “impostor syndrome” (where they don’t consider themselves experts, when anyone else would)?
What you can do to help inclusivity in peer review:
As a scientist, it is important to address your own and your community’s awareness of the larger problem. Self assessment and opening up to dialogue about these sorts of issues will help you create the social accountability necessary for change. I am actually creating an organization in my college to create this type of dialogue to apply to our Science and professional community. You might consider doing the same, if groups such as this likely don’t already operate at your institution.
As an author, you should be aware that editors take your suggestions seriously, using about 30% of your suggested reviewers. The list you give them might pre-dispose them (anyone, really) to think of a certain group of people similar to your list.
To Editors and Publishers-
Same thing: awareness, self assessment, social accountability. Action, too. In a constructive way, of course. Be an advocate for those under-represented voices! I hope you frame these findings in a positive light- as a previously unrecognized opportunity for growth.
A more transparent process of choosing reviewers might pave the way for a more equitable distribution of reviewers. Accessing all potential reviewers may decrease the load of those currently doing the work, and meet the growing demand of reviewing work.
Double-blind review would surely take away bias in the reviews, but also takes away the networking and collaborative benefits it provides. Double-blinding the peer review process is treating the symptom, not the source of the problem.
Gender, while not ideally measured as a binary, is easily collected and readily self-reported. It is useful to examine as a large dataset to examine a marginalized group. This study is an example of how career opportunities are afforded to a lesser extent to underrepresented people. Think about how opportunity gaps might magnify into larger differences later in one’s career and how other groups may be impacted by the same social mechanisms that affect women here.
What is your personal experience with gender bias and peer review? Share your thoughts in the comments below.