Sunday, June 19, 2022

The "service work" imbalance

Imagine you're sitting at a computer with a button you can click. You're in a room full of other people who are also participating in this experiment. You've been randomly placed in a group with two other people, though you don't know which two.


If, at the end of a two minute timer, none of you have clicked your button, all three of you earn $1. But if someone clicks the button, the round immediately ends, that person earns $1.25, and the other two group members earn $2.


There are 10 rounds. Nobody knows who volunteered in which round.


What would you do?



When researchers ran this experimental game of chicken, they found that in mixed-gender cohorts, women volunteered 3.4 out of 10 times on average, while men volunteered only 2.3 out of 10 times. In contrast, when the group of people in the room was single-gender, everyone volunteered about 2.7 times out of 10.1


The researchers concluded that in mixed-gender groups, women are more willing to volunteer to do tasks that incur a cost to themselves, and men are more likely to hope someone else (perhaps a woman) will do it. This wasn't because the women were better at the task or enjoyed it more, nor was it because women were nicer and more altruistic: in single-gender groups, women and men volunteered at the same rates! So what's going on?


This experiment was conducted by the authors of The No Club: Putting a Stop to Women’s Dead-End Work. These four successful women started a "No Club" because they kept finding themselves agreeing to take on work tasks that someone needed to do but that did not benefit their careers. Together, they slowly got better at saying "no" to these non-promotable tasks (NPTs). This made their workloads more manageable and gave them more time for promotable (i.e., career-benefitting) work. But then they noticed that the NPTs they turned down were just being taken on by other female colleagues instead.


When they began to conduct studies and interviews on the question, they found that women were doing more than their fair share of NPTs across industries ranging from academia to law firms to the TSA to bartending. Women were likelier to volunteer or to be asked to do these tasks, and once asked, women were likelier than men to say yes.2


One study of a consulting firm's records on their employees' billable hours found that women spent about 200 hours more per year than men on non-promotable work! For women in more junior roles, this time ate into their time spent on promotable work; for senior women, it was on top of their promotable work. The senior women were simply doing more work in total than senior men—perhaps they realized they had to in order to get where they are!


One of The No Club’s authors, economics professor Linda Babcock, gives a personal example, comparing her schedule on a non-teaching day to that of a male colleague. Hers looked like:


8:30-10:30 IRB meeting

10:30-12 curriculum review meeting

12-1:30 student presentation

1:30-2 interview with reporter

2-3 executive education meeting

3-4 research

4-5 prepare talk for womens' group

5-6 faculty meeting


His looked like:


8:30-12 research

12-1:30 student presentation

1:30-5 research

5-6 faculty meeting


He had time for 6 more hours of research. This is an extreme example, but it illustrates the point.  



In the fall semester of last year, I remember apologetically telling a fellow graduate student I couldn't co-run LGBTQ math socials with her. I was doing too much of this sort of thing already: Math Includes, Real Representations, Girls' Angle, DRP, WiSTEM, and being a GSC rep. When I listed them all, I realized it might be wise to cut down on these things in the spring semester. In retrospect, I mostly failed to do this.


Last month, I was asked for the first time to review a paper for a TCS3 conference. I agreed, of course: reviewing papers is a service to the community, and for me it's also a new skill to develop. Coincidentally, less than a week later, I was asked again by someone else to review a different paper. Now I was torn. I felt bad saying no, but I was busy and I knew it would take away from my research time. Eventually, I clicked on the paper to glean some hint of how time-consuming it would be to review. I promptly saw that my PhD advisor was one of the authors, upon which I breathed a sigh of relief and informed the requester that I had to say no due to a conflict of interest.


The replication crisis has taught me to take all social science studies with a grain of salt. But I can certainly say this stuff comes up in my own life. There's a name for NPTs in academia, and it’s "service work," and I do a lot of it. And I wouldn't be surprised in the least if this is correlated with my gender.


So what do we do about this?


Naturally, we should change the current norms and expectations so that we no longer expect women to take on service work at higher rates than men. Individual women can volunteer less and be more willing to say "no" when asked to take on an excess of service work. And everyone—especially people in positions of authority—can consciously strive to be gender-balanced in who they ask to do service work, to fight back against subconscious biases toward asking women.


There is however a complication: sometimes you do specifically need women4 for a particular role. For example, on a faculty committee focusing on diversity and inclusivity, you obviously want representation of perspectives from underrepresented groups. For a math mentorship program for women, you want female mentors to serve as role models. And so on.


I think the trick is to pay attention to when a woman is actually needed for the role, and when a man would do. Don't do what a university referenced in The No Club did—they bragged that their faculty committees were all 50% women, despite only a quarter of faculty being women! Although this practice was well-intentioned, it meant female faculty had to spend a lot more of their time on committee-related service work, while men were comparatively off the hook. They would have done well to think carefully about which committees needed gender parity most.


Finally, departments can make service work promotable, or at least less non-promotable, by counting mentorship and service as an important factor when choosing candidates to hire or promote. This was on my mind last year when I was asked to serve on the grad student interview committee for new faculty candidates (another NPT, I now realize, ha!). We grad students were there to judge whether the candidates would be good teachers and PhD advisors, but I made sure to ask each candidate a question about what they were doing to support diversity in math. I'm not sure if my initiative-taking actually affected anything—optimistically, one hopes that these questions about service are included elsewhere in the interview process. But I'm not privy to the rest of the process, so who knows.



Reader, how about you? What is the gender balance of service work in your department or office? And what can you do about it?




Thanks to Rebecca Giblon, Holden Lee, and Dennis Yi for feedback on a draft of this post.



Multiplying this by three, you can see that there were some rounds in which no one volunteered. 

The authors acknowledged as a limitation of their research that their studies focused only on men and women; more research is needed to understand how those outside the gender binary fit into all this. 

Theoretical computer science. 

Or other underrepresented groups—my focus here is on gender, but I expect this part is more broadly applicable.