320. Gender Differences in Faculty Retention

Women and men leave academic positions at different rates and for different reasons. In this episode, Aaron Clauset and Katie Spoon join us to discuss their research on the magnitude of and differential causes of gender differences in faculty attrition.

Aaron is a Professor of Computer Science at the University of Colorado, Boulder. He is a nationally recognized expert on network science, data science, and complex systems and he is the recipient of the 2016 Erdos-Renyi Prize in Network Science. Katie is a computational social scientist and a 4th-year PhD candidate, also at the University of Colorado, Boulder.  Aaron and Katie are two of the authors of a paper on “Gender and retention patterns among U.S. faculty,” which has received a great deal of attention and has been discussed in the Chronicle of Higher Education, Times Higher Education, Science Careers, and Nature News.

Show Notes


John: Women and men leave academic positions at different rates and for different reasons. In this episode, we examine the magnitude of and differential causes of gender differences in faculty attrition.


John: Thanks for joining us for Tea for Teaching, an informal discussion of innovative and effective practices in teaching and learning.

Rebecca: This podcast series is hosted by John Kane, an economist…

John: …and Rebecca Mushtare, a graphic designer…

Rebecca: …and features guests doing important research and advocacy work to make higher education more inclusive and supportive of all learners.


John: Our guests today are Aaron Clauset and Katie Spoon. Aaron is a Professor of Computer Science at the University of Colorado, Boulder. He is a nationally recognized expert on network science, data science, and complex systems and he is the recipient of the 2016 Erdos-Renyi Prize in Network Science. Katie is a computational social scientist and a 4th-year PhD candidate, also at the University of Colorado, Boulder. Aaron and Katie are two of the authors of a paper on “Gender and retention patterns among U.S. faculty,” which has received a great deal of attention and has been discussed in the Chronicle of Higher Education, Times Higher Education, Science Careers, and Nature News. Welcome, Aaron and Katie.

Rebecca: Today’s teas are:… Aaron, are you drinking any tea today?

Aaron: I am drinking coffee.

Rebecca: Always a popular tea choice. How about you, Katie?

Katie: I made tea specifically for this. And it’s

Rebecca: Yay!

Katie: …and it’s Trader Joe’s Ginger Tumeric herbal tea?

Rebecca: That sounds yummy.

John: That’s really good. I’ve had that. And I have Prince of Wales today.

Rebecca: And I have some Christmas tea, a little ahead of schedule.

John: We’ve invited you here today to discuss your paper on “Gender and retention patterns among U.S. faculty.” Could you give us a general overview, of the focus of the study?

Katie: Yeah, so we were interested in this topic of gender retention in academia, because it’s really hard to identify faculty who have left academia. So a lot of studies have had to look at a very narrow or specific population of faculty. And so while there’s a lot of studies on this topic, the majority seem to focus on assistant professors, STEM fields, and higher prestige institutions. And so we were really interested in doing a systematic study across career age, institution, and field.

Rebecca: Can you talk a little bit about what data you used for this study.

Katie: So we have a data use contract with Academic Analytics Research Center, which is a company that contracts their data with research teams, and it’s a census for all of the tenure-track and tenured professors at US PhD-granting institutions across, I think, over 100 fields that we grouped into broader domains. And we have data from 2011 to 2020, for all faculty within that period, so we can see if anybody left their job during that decade, then we have that information. And then our study has two parts, we first look at just the rates women and men faculty are leaving their jobs using that employment census. And then we emailed a frame of around 70,000 of those professors from 29 fields, and got about 10,000 responses to a survey looking at the underlying reasons why faculty were leaving.

John: So that’s a remarkably large sample. How did you measure faculty attrition in a sample,

Katie: We calculated attrition risk as just the number of people who left divided by the number of people who could have left in a given year. And then we calculated that across career age. So that’s the year since a faculty member received their PhD. That was kind of our simplest calculation of attrition risk. And then, from that analysis, you could kind of see broadly that women were leaving their jobs at higher rates than men at every career age and the gap seems to grow over time. And so then we did a logistic regression analysis, looking at the odds of leaving a position where we adjusted for other factors.

John: What other factors did you hold constant in the logit regression?

Katie: So we were adjusting for career age, employer prestige, so the institutional prestige, and whether they were trained for their PhD in the US or abroad, and adjusting for those factors, we found women were more likely to leave than men. And that varied a lot across career age, field, and institution.

John: We’ve long known that women are not very well represented in the upper ranks of faculty, but many people would try to attribute that to the fact that we’ve seen a general increase in the number of women entering disciplines. But what you’ve demonstrated is that we’re losing a lot of people along the way, that we don’t see as many women full professors, not just because there weren’t that many women entering that stream a decade or more ago, but also because we’re losing so many people, which I think is a really important result, which I don’t believe had been measured as carefully before. Is that a correct interpretation?

Katie: Yeah, I think definitely both hiring and retention are important for overall gender equity in academia. If you hold the right It’s of attrition constant from our study, then even if you start with a 50-50 cohort of men and women as assistant professors at career age one, then even after a whole career, that falls to about 40% women, and a lot of people leave during that time, like a lot of men too. The majority actually have left over that time, but women are leaving faster, which I think is important to keep in mind.

Aaron: So the gender representation in the U.S. academy is complicated, in part because the academy is really big and diverse. I mean, there’s lots of different fields and lots of different institutions. And this is partly why the literature on this topic is really messy, is because many of the studies look at one institution, they may look at R2 institutions, or fancy R1institutions, or only this field or that field. And so it’s been sort of hard to get a sense of what’s going on. And there’s lots of sort of claims about why it is that you don’t see as many women in the most senior ranks in academia. So the idea that it’s a sort of a demographic inertia thing that in order to be provost at a university, well, you got your PhD probably 20-30 years ago. And so the gender ratio at 20-30 years ago is what sort of sets the maximum gender ratio 20-30 years in the future for those senior people, and of course, that’s a part of the explanation. That’s certainly not wrong. But we were very interested in understanding this attrition aspect of people who are leaving academia. And not just like changing universities, I think that’s a critical piece of the way we measured attrition is that if you went from Princeton to Yale, we don’t count that as an attrition event, you have to leave the entire set of PhD-granting universities in the US in order for us to count you as an attrition event. So these are people who are not just leaving one academic job for another, but like they’re leaving the American academy entirely. And so the results of looking at the data, because this sort of broad census lets us see that the rates, overall, women are leaving at higher rates than men. But it gets really messy when you start looking at individual disciplines. Some disciplines have higher rates than others, some places, it doesn’t seem like there is a gender difference in the rate. And when you drill down to individual fields, it gets even messier because fields have different norms and different sizes and things. So the real power of the study is in part because it has this just enormous view, a quarter million faculty in the data set, which lets us look at differences in these small rates in a really powerful way.

Rebecca: You’re just mentioning that there’s messiness across disciplines and fields. Can you talk about whether or not there was any similarities in STEM and non-STEM fields?

Katie: I think, maybe surprisingly, we found the gender gaps in retention were actually larger in non-STEM fields than in STEM fields. And that was especially true at the full professor level, the gender gaps overall grow with rank. And so that kind of surprised us, I think, because it’s a little counter to what the literature has mostly focused on, which is assistant professors, and especially in STEM. So we actually find that full professors in non-STEM fields are the group with the largest gender gap in retention.

Aaron: This is an important finding, in part because there’s a narrative that sort of underlies a lot of discussion about faculty attrition and gendered attrition, in particular, that once you get tenure, it’s fine. You can’t get fired, tenure is the goal you’re looking for. Everybody wants it. And once you get it, somehow, everything gets better. And so this finding in the paper, that it’s actually tenured women who are leaving at higher rates than anyone else, it kind of inverts that narrative and shows that we’ve been maybe looking in the wrong place for the real effects of this thing.

Katie: And I think, yeah, more on that point, for us seeing that the effect was larger for full professors kind of points to maybe something else going on besides professional things, or work-life balance in the early career, which is what led us to do the survey and check out what was happening underneath the rates.

Rebecca: There’s often a stereotype that women maybe earlier on in their career leave because of childbirth or raising children and things like that. But that’s definitely counter to what you’re suggesting.

Aaron: Women certainly still leave the Academy for work life-balance reasons. So the observation that the rates sort of run counter to the narrative that has been explored very thoroughly in the literature, Katie was saying it really points to there being more to the story, not that the previous story was necessarily wrong, but like there’s just more going on. And also, things have certainly changed in the past 30 or 40 years. Gender norms change. Men are much more involved in child rearing now than they were 40 years ago, and so our study is at a particular decade of time. So we aren’t able to say things as much about what was going on in the 1980s. But today in the 2010s, when the census data on attrition was collected, we can see that there’s more going on and, as Katie was saying, that’s where the survey really helps us go beyond just looking at the rates of things to try to look at the reasons for these rates.

John: To investigate the causes of attrition, you then did a survey. Could you talk a little bit about the sample that you used, how large it was, how you determined it, and what you found?

Katie: So we took the larger census employment data and we selected 29 fields that we considered broadly representative. We wanted a mix of fields across different sizes, different domains of study, as well as fields that had commonly been studied in the literature, like engineering or computer science versus fields that were less commonly studied, like education. And we then attempted to get emails for all of those people, and then ended up with a frame of about 70,000 people who we then contacted. We got roughly 10,000 responses from both current faculty members, as well as former faculty members who either left academia or retired. And then we also looked at faculty who switched institutions. I think the survey really showed us that while there were no visible gender gaps in certain fields, or at certain ranks, in the first part of our study, where we were looking at the rates that women and men leave, we found that the reasons women and men leave are still very gendered, even in those cases. So for example, I think early career STEM faculty, there was no visible difference in the rates that women and men left. But in the survey, we did find that women were more likely to feel pushed out of their jobs and less likely to feel pulled towards better opportunities. And that, while there was a lot of variation across domains in the retention rates, that there wasn’t much variation in that pattern of feeling pushed or pulled, it was like varied more universally, I guess, across domains.

Aaron: That push-pull dynamic is really key, and it was something that had been studied previously in the literature. And so we wanted to incorporate that into the survey. And I’m just going to read to you so that your listeners can hear what we meant by a push and a pull. So the question we asked for the push was,:”I am unhappy, stressed or otherwise less than satisfied in my current position,” and the pull was “I am drawn to, excited by, or otherwise attracted to a different position.” And so the idea of those pushes and pulls is really about whether the person wants to stay in their current position or not. Are they drawn to something else? Is it an opportunity that they’re pursuing to leave their current position? Or do they feel like they’re being forced out, that they’d rather stay where they are, but they feel like they cannot stay. And so that push and pull really sort of illustrates the idea that it’s up to that person to decide whether they feel like they can stay or want to stay in their current job, or whether they are choosing to move to a different opportunity. And people could choose both. I mean, leaving an academic job is a big decision. Sometimes you might feel unhappy, and you start looking for an opportunity and then you find something that looks more exciting. And so survey respondents could check both. They said, “I could feel pushed and pulled at the same time.” And so our results on the push-pulls were just focused on the ones who said either push or pull.

Rebecca: Given these findings, what are some things that colleges and universities might do to reduce the gender differences in attrition rates and maybe make it more attractive to stay in academia?

Katie: Yeah, so I think a key finding of our study is that, even where the rates may look the same, looks like there’s gender equity, the reasons could be very different. So I think the first thing would be to also investigate: are the reasons people are leaving gendered, even if doesn’t look like they’re leaving at different rates? And then I think, as far as the particular reasons, we did kind of look into what pushes people were experiencing specifically. And while professional reasons like productivity and funding, as well, so work-life balance, which is things like caring responsibilities and the workload. While both of those categories are definitely still important, especially for faculty at different career stages, we found that climate, the workplace climate, so how someone feels when they go to work around their colleagues, they feel that they don’t belong or fit, if there’s dysfunctional leadership, that those workplace climate reasons were the most gendered, both for former faculty who actually left a position, as well as for current faculty who were considering a hypothetical decision to leave your job. And so I think focusing on the climate reasons, especially as a first step, would be a good way to reduce the gendered impact, specifically, but for retention overall, I think that all the reasons are probably important and listening to faculty, taking people seriously when they say this is why I left, is good.

Aaron: One of the workplace climate findings from the survey that was, I think, most striking is that climate was listed as a very important reason by women at all career ages, whereas the work-life balance issues were most salient early in the career and then it sort of tails off for everyone, men and women. And then professional reasons are sort of high but they also just sort of tail off over time, but the climate reasons are high for young women professors and also senior women professors. So this suggests that this is a place where universities are really not moving the ball forward and trying to improve things because you have both the young faculty and the senior faculty all saying consistently this is a huge issue for their stress. And they’re feeling like they belong, and they feel pushed or pulled out of their academic positions. So figuring out how to move that needle is critical. And there are some efforts. So for instance, Harvard has this CAOCHE survey that many universities do on climate, but it’s typically done at the institutional level, but climate is typically a departmental thing. So if you look at like the computer science department versus the sociology department, they’re very different social sort of contexts. And one department could have a good climate, one is very supportive and inclusive of women, and the other could have a very dysfunctional one. And so these sort of large scale institutional instruments for measuring climate, they may not really be up to the task of allowing an institution to improve in this way. So figuring out how to get down to the level at which climate is actually operationalized, which is the department level, I think, is a key need for institutions of higher education.

John: One of the results that we haven’t yet talked about is, when you were looking at the initial study of attrition rates, when you included a control for the major field, you found that there was no significant gender difference, holding academic field constant, for people at the more junior levels. But at the associate and full professor level, there was a significant gender effect. That result I thought was kind of surprising. I would have guessed that there would have been more women leaving earlier in the career. And I think that’s what many people would suspect because of issues such as work-life balance, and so forth. But I thought that result was really surprising. It certainly has to do with climate, I think, as you’re suggesting, but why would it be more severe at the associate and full professor level than at the more junior level? Or is it just the cumulative effect, perhaps, [LAUGHTER] building up and getting worse over time.

Katie: We added fixed effects for fields to try to get at if you have a woman and a man at the same institution and in the same field, so they’re probably in the same department then or very similar departments. We were trying to get at that average field effect. Our main results are like, if you were to just select someone in the natural sciences, what is their odds of leaving if they’re a woman compared to a man? And women are distributed across fields unevenly. And so basically, adding fixed effects for field lets us adjust for that variation in gender ratio, or other kind of field norms. And so I think the fact that the effect goes away at the assistant professor level, like the effect is not very large to begin with, for assistant professors. And so it’s possible that like, adding that field effect just reduces the gendered effect for assistant professors, whereas the effects are larger for tenured faculty. So really, it’s just saying that even when you adjust for someone’s field and all of these gender ratio, and other field level variables, we do still see something going on. We can’t say for sure why that is, but I think the survey really suggests very strongly, that is pointing to the climate.

Aaron: Our study was not designed to get at the causality of any of the differences in the rates from like a causal inference perspective. And so we can speculate about why senior women might feel like climate is a larger issue in terms of wanting to leave, and then many more of them are leaving, etc., and we can speculate about that. And certainly, if we all have experience in academia, we know women who fall into this category, we probably heard some of their stories. And those stories are real, and probably many more women have those stories than we might imagine. And so we can speculate about that, but the study itself was specifically aimed at looking at the rates for leaving for attrition and then asking in the survey, the reasons that people felt like they could or would or did leave their academic positions. When I was visiting UMass-Amherst last week, and I was talking with some of the folks in their NSF funded ADVANCE grant program. We had a very nice comment session about this specific question, about what it is that the senior women are experiencing. And several of the members of the ADVANCE team there sort of talked about the idea that… it makes a lot of sense… that over the course of a 20- or 30-year career, you might experience sort of the accumulation of these sort of devaluation experiences, and that might slowly sort of erode your belief in the system and your belief in a sense of belonging, even though you are tenured, or you are a full professor, or maybe even distinguished professor… you know, you’re at the top of your career, and yet 20 to 30 years of being devalued by your institution and your male colleagues. I mean, who wouldn’t consider moving on to something else, maybe considering early retirement, for instance. And so it’s not hard to imagine what’s going on with the senior women. But I think that it’s really important that future work really dig into that specific question. Because there is this sort of narrative that once you make it as a full professor, you get to be the boss and everybody else who’s more junior than you. But that’s not what we’re seeing in the data, we’re seeing that the senior women are saying that climate is as much of an issue for them, as it is for the junior professors who have been there for three or four years.

Rebecca: Your point earlier too isn’t lost on me that there’s fewer women in administrative roles. And this is the pipeline for those roles [LAUGHTER] too, that you have to be more senior often to be in those roles. And it’s really concerning about the future of the academy when we’re thinking about how few women are continuing on.

John: So, your study focuses only on PhD granting institutions. One source of attrition, though, that could occur, potentially at least, is faculty may be leaving for comprehensive or four-year institutions and so forth, which wouldn’t be picked up in the data. And that might be a nice thing for someone to explore as a future area of study, to see if some of this might be a shift in institution types. I suspect it’s not, but we do know that there’s a much larger proportion of women in community colleges and a much larger proportion of women in four-year colleges than in PhD granting institution. So it’s possible some of it could be attrition in that form. But that still doesn’t explain why there is such a gender difference in the climate, which seems to have a much larger cumulative effect than immediate effect.

Aaron: This makes me think of our results on prestige. Because nothing makes sense in academia, except in the light of prestige. And one of the things that our research group here has been very active in is in looking at how prestige structures academic careers, and the production of science and so on. So we really wanted to know, like, is prestige lurking on here as well. And we found some really interesting things.

Katie: Yeah, I think the main effect for prestige… it’s gendered slightly… but the main thing is that faculty at lower procedure institutions overall, regardless of gender, are much more likely to leave, then faculty at higher prestige institutions. And again, our analysis is looking at all-cause attrition, so we don’t know if they are leaving to go to a different type of institution or leaving academia altogether, but it’s definitely a huge effect. And then we do find a small gendered effect on top of that, where women are even more likely than men to leave low-prestige institutions, especially full professors.

Aaron: And if we think about how many professors are at those elite institutions, I mean, the Ivy League or the Ivy League plus, it’s a really small number of institutions that receive a huge fraction of our attention in academia. And that’s the prestige game right there. And so if you think about the number of people who are at the elite institutions where the attrition is lowest, it’s not many, most tenure-track faculty are at these less elite institutions. And so they’re experiencing these higher attrition rates. And we don’t have an explanation for why. But we’re very curious to know what is it about these lower-prestige institutions that makes attrition overall higher and more gendered?

Rebecca: Many questions to continue exploring, for sure. [LAUGHTER]

Aaron: Yeah, it’s a mystery.

Rebecca: So why do we care about attrition, anyways? We talk about it with students all the time, we talk about keeping students and retaining students. Why is it important to think about retaining faculty and keeping them in the pipeline in higher ed.?

Katie: I think there are a lot of practical reasons why it’s good to retain women. Studies have shown that people tend to ask different types of scientific questions based on their identity, such as their gender or race. We also know that having women in leadership roles, there’s been some studies showing that that can influence younger women’s perceptions of being a scientist and things like that. But personally, I think the most important reason for doing this is that scientific talent is not inherently gendered, and so there shouldn’t be differences in how women experience their workplaces. This is what I think.

John: And if people are going to invest in acquiring a PhD, that’s a pretty large investment for them as an individual. And it’s really unfortunate to have that wasted, especially if they’re feeling pushed out. If they were moving on to some other position where their skills could be better used, that would be an improvement in efficiency for them and for society, but if they’re leaving because of a chilly climate, that’s probably not the best use of society’s resources.

Rebecca: It’s a lot more resources to bring on new people over and over again too.

Aaron: Every faculty search is a big lift. And if you hire a bunch of talented women, and then they don’t feel like they belong in the academy, and they leave, then not only do you have to hire new people because you have the faculty leave, but every one of those incredibly talented, highly trained, highly productive women, their contributions to science have now been lost in some sense. And that seems like a moral tragedy for all of us.

Rebecca: I think a lot of our institutions focus a lot on having policies and strategies that are supposed to be equity minded. This is an equity gap that isn’t always attended to.

John: And also there’s a lot of research that suggests that when faculty are happy and excited about their work, that enthusiasm tends to get conveyed to students. And we’re losing that we have a lot of people who are unhappy being in the positions they’re in.

Rebecca: Wait, is that why we’re in higher ed, just teach students and get them excited about our fields. [LAUGHTER] Sometimes we lose track of that. [LAUGHTER]

Aaron: In some ways, that’s like the only reason to be in higher ed.

Rebecca: Exactly.

Aaron: As a tenure-track professor, our job is teach the next generation of students and then also make discoveries about the world and so on. And if we didn’t want to do the teaching part, then we would go be a research scientist somewhere. And there’s plenty of people who can do that. But it’s that combination of the teaching and the research that makes being a tenure-track professor in the American academy really special and sort of the envy of the world. In fact, the American system, to a large extent, is being copied and has been copied by the rest of the world’s higher education systems, because it’s such a powerful synergy. And you’re right. If your faculty are happy, the students will feel that and they’ll be excited about continuing on, and they’ll get infected with the excitement of science and discovery and so on. And in some ways, it’s kind of confusing that institutions are so behind the curve on faculty attrition, and especially gendered attrition, because it’s not like this is a new phenomenon. I mean, our paper is the latest in a long history, you know, 40 plus years of studies looking at gendered attrition at different places in the pipeline. And the fact that it’s still something to study is kind of an indictment of all of higher education.

John: And on that happy note…. [LAUGHTER]

Rebecca: I was going to say, on that powerful note.

John: …we’ll move to our final question, which is: “What’s next? “

Katie: Well, we actually have a follow up study that we’re working on right now. Aaron had kind of mentioned a little bit earlier about this accumulation of feeling devalued across a career. We asked our survey respondents what needs to be different about their positions to reduce their stress, or what needed to be different for them to stay in their positions. And we got over 6000 free-text responses to that question. We then read them all and we’re working on writing up the results. But it’s very powerful reading these stories from people and really narrowing in on like, what exactly is going on in the climate? What is going on for older women? for women of color? for mothers? …and so we’re excited to get that out.

John: Excellent. Well, thank you. It’s been great talking to you. And we’re looking forward to seeing that next study.

Aaron: Thank you.

Rebecca: Yeah, this is important information. I’m glad that we’re able to share it in this space.


John: If you’ve enjoyed this podcast, please subscribe and leave a review on iTunes or your favorite podcast service. To continue the conversation, join us on our Tea for Teaching Facebook page.

Rebecca: You can find show notes, transcripts and other materials on teaforteaching.com. Music by Michael Gary Brewer.

Ganesh: Editing assistance by Ganesh.