Here’s a work scenario many of us know too well: You are in a meeting and your manager brings up a project that needs to be assigned. It’s not particularly challenging work, but it’s time-consuming, unlikely to drive revenue, and probably won’t be recognized or included in your performance evaluation. As your manager describes the project and asks for a volunteer, you and your colleagues become silent and uneasy, everyone hoping that someone else will raise their hand. The wait becomes increasingly uncomfortable. Then, finally, someone speaks up: “Okay, I’ll do it.”
Our research suggests that this reluctant volunteer is more likely to be female than male. Across field and laboratory studies, we found that women volunteer for these “non-promotable” tasks more than men; that women are more frequently asked to take such tasks on; and that when asked, they are more likely to say yes.
This can have serious consequences for women. If they are disproportionately saddled with work that has little visibility or impact, it will take them much longer to advance in their careers. Our work helps explain why these gender differences occur and what managers can do to distribute this work more equitably.
What Are Non-promotable Tasks?
Non-promotable tasks are those that benefit the organization but likely don’t contribute to someone’s performance evaluation and career advancement. These tasks include traditional office “housework,” such as organizing a holiday party, as well as a much wider set of tasks, such as filling in for a colleague, serving on a low-ranking committee, or taking on routine work that doesn’t require much skill or produce much impact.
What is non-promotable varies across fields and careers. For example, in industry, revenue-generating tasks are more promotable than non-revenue-generating tasks; in academia, research-related tasks are more promotable than service-related tasks; and for individuals, a task may be promotable for junior employees but non-promotable for senior-ranking managers.
Studies of industry and academia (by Irene De Pater and colleagues; Sara Mitchell and Vicki Hesli; and Joya Misra and colleagues, as well as many others) have shown systematic gender differences in how work is allocated, with women spending relatively more time than men on non-promotable tasks and less time on promotable ones. These differences matter because they help explain why, despite women’s significant educational and general workplace advances, we continue to find vastly different promotion trajectories for men and women. Women will continue to progress more slowly than men if they hold a portfolio of tasks that are less promotable.
Although what makes something non-promotable varies across occupations, there is typically agreement within an occupation about what tasks are non-promotable versus promotable. For example, in a survey of 48 Carnegie Mellon faculty, we found that 90% agreed that an assistant professor has a higher chance of promotion if they allocate spare time to research rather than to committee work (like being on the faculty senate). Separately, looking at data from a large public U.S. university, we found that when all 3,271 faculty were asked to volunteer for a faculty senate committee, only 3.7% chose to do so — but 7% of women volunteered, compared with 2.6% of men.
There are of course many reasons why women volunteer more than men. It may be that women are better at these tasks or enjoy them more than their male colleagues. To test these explanations we conducted a series of lab experiments at the Pittsburgh Experimental Economics Laboratory (PEEL). A total of 696 University of Pittsburgh undergraduates participated in the studies.
Who Volunteers, and Why?
We designed a simple decision exercise to examine who agrees to do non-promotable tasks. The design mimics the scenario we opened with, that of finding a volunteer for a project at a work meeting.
In the first experiment, we had male and female participants sit in front of a computer in the lab and make decisions in 10 rounds. In each round participants were sorted into new groups of three (they knew they were matched with other participants in the room, but not exactly who) and had to secure one volunteer from the group to click a button on the computer screen. The group was given two minutes to decide, with the round ending as soon as someone volunteered. If no one volunteered, each group member received a payment of $1. If someone volunteered, that person received $1.25, while the two other group members each received $2. So every group member was better off if someone volunteered, but the volunteer benefited less.
Overall, the participants were reluctant to volunteer. While 84% of groups succeeded in finding a volunteer, it typically did not happen until the final seconds of the 2-minute round. Importantly, the rate of volunteering was not the same for men and women. Averaging across the 10 rounds, we found that women were 48% more likely to volunteer than men, and we saw this difference in every one of the 10 rounds.
Because the volunteer task in this experiment was to click a button on a computer screen, we could rule out the possibility that women volunteered more because they were better at the task or enjoyed it more than men.
But gender differences in preferences may nonetheless contribute to the difference in volunteering. In particular, women may volunteer more because they may be more risk-averse or altruistic than men. To examine this, we looked at survey measures of participants’ agreeableness, altruism, nonconformity, and risk aversion. While some of these measures correlated with the decision to volunteer, none of them explained the gender difference.
To directly test the effect of preferences, we also conducted a second experiment, where instead of having both men and women participate together, we had groups of only men and only women.
If the difference in volunteering resulted from women being more risk-averse or altruistic than men, then we should have sees greater rates of volunteering in our all-female sessions than in our all-male sessions. But we found that volunteering rates were identical. Women were no more likely to volunteer than men when everyone was in a same-sex group.
This helped rule out the explanation that preferences caused the gender difference in volunteering in our mixed-sex study. These results instead suggest that the real driver was a shared understanding or expectation that women would volunteer more than men. In a mixed-sex group, men will hold back on volunteering while women in turn will volunteer to ensure that the task is done. But in single-sex groups, this changes — men and women volunteer equally. In these groups men know they have to step forward if they want to find a volunteer, and women expect other women to volunteer, making them less compelled to do so themselves. Interestingly, in women’s groups the volunteering ends up being shared equally across 10 rounds, while in men’s groups it tends to fall on the same men each time.
Who Is Asked to Volunteer, and Why?
To further confirm that people expect women to volunteer more than men, we conducted a third experiment, this time adding a fourth person, a manager, to the group. At the start of each round, the manager had to publicly ask one member of the new three-person group to volunteer. (The manager couldn’t personally volunteer. They saw pictures of the other members of their group and clicked on the picture of the person they wanted to ask.) The manager got $2 if someone in the group volunteered and $1 if no one volunteered. The rules for the group of three remained the same — $1 each if no one volunteered, but if someone volunteered, that person received $1.25 while the other group members each received $2. Managers were free to ask any member of the group to volunteer, but we expected they would be more likely to ask women than men.
This is precisely what we found: Women received 44% more requests to volunteer than men in mixed-sex groups. Intriguingly, the gender of the manager did not make a difference: Both male and female managers were more likely to ask a woman to volunteer than a man. This was apparently a wise decision: Women were also more likely to say yes. A request to volunteer was accepted by men 51% of the time and by women 76% of the time.
Can We Be More Fair?
Our studies demonstrate that although neither men nor women really want to volunteer for thankless tasks, women volunteer more, are asked to volunteer more, and accept requests to volunteer more than men. These differences do not appear to result from gender differences in preferences, but rather from a shared understanding that women will volunteer more than men.
While our results are disconcerting, they also provide a silver lining in suggesting how employees and managers can reduce the inequity in work tasks. The solution is not for women to decline more work requests — which would present problems for organizations and hold repercussions for women — but instead for management to find ways to distribute tasks more equitably. Rather than asking for volunteers or asking women to volunteer because they are likely to say yes, managers could consider rotating assignments across employees, for example. Understanding that women volunteer more simply because men are reluctant to do so should also lead men to volunteer more themselves and should empower women to demand fairer treatment.
Changing this dynamic should be a priority for any organization that wants to advance its most qualified employees. Workers who spend more time on non-promotable tasks are held back from demonstrating their full potential. If this burden falls disproportionately on women, not only is their advancement stymied, but also corporations miss out on capturing valuable talent.