Like last year and the year before that, this year I once again quantified the gender balance of newly-hired tenure-track asst. professors in ecology and allied fields at N. American colleges and universities. I also conducted a poll asking readers what they expected me to find. See here for details on how I compiled the data, and why I went with a gender binary even though that’s not ideal.
This is a long post; grab a coffee and get comfortable! Diversity and equity are important issues on which people have strongly-held views. That’s why I’ve tried to discuss the results thoroughly and carefully, and to anticipate and address questions that readers are likely to have. I urge you to read the whole post rather than just reading the headline results. But if you insist on skimming, I’ve broken the post up into bold-headed subsections, which are listed below. Pick the ones that interest you.
- Headline results: 59% of tenure-track asst. professors of ecology hired in N. America during the 2017-18 job season are women. It’s 57% women over the last three years.
- The 95% confidence interval is 55-59% women over the last three years.
- No, the headline results are not a product of biased sampling.
- The gender balance of recently-hired ecology faculty varies a bit with the research intensiveness of the institution.
- Ecologists as a group remain largely unaware of the headline result, or else can’t quite believe it.
- When placed in the context of other data, particularly on gender balance at other career stages, these results tell a different and more complicated story about gender diversity and equity in N. American academic ecology than you probably realize.
- No, the headline results don’t indicate “reverse discrimination”
- Bottom line: these results are good news. They represent real, systemic progress for N. American academic ecology in one specific but important area. (UPDATE: As I said in the post but which I’ll now say up top here: no, these results do not mean that everything is great for women in ecology now [it’s not], or that women in ecology should all be happy [it’s obviously not my place to tell anyone how to feel], or that people should stop working to improve diversity and equity in ecology [because we should keep working].)
Headline results: 59% of tenure-track asst. professors of ecology hired in N. America during the 2017-18 job season are women. It’s 57% women over the last three years.
I ID’d the gender of 161 newly-hired TT asst. profs in ecology and allied fields (e.g., fish & wildlife, evolutionary ecology, behavioral ecology, etc.) hired at N. American colleges and universities in the 2017-18 job season. That 161 includes a couple of 2016-17 hires I didn’t discover until this year (it’d be a pain to go back and update last year’s post yet again so I just folded those newly-discovered 2016-17 hires into the 2017-18 data). That’s out of 291 positions I ID’d or tried to ID this year. I try to ID every N. American TT asst. prof position that could possibly have been filled by an ecologist.
95 of those 161 new hires (59%) are women. Combining with data from previous years, over the previous three years 57% (314/548) of recently-hired TT N. American asst. profs in ecology and allied fields are women.
The 95% confidence interval is 55-59% women over the last three years.
Gender balance of newly-hired ecology faculty doesn’t vary much from year to year. It was 54% in 2015-16, 59% in 2016-17, and 59% in 2017-18. So if you’re worried that next year’s proportion of women might be a lot lower than this year’s, stop worrying. Seriously, stop. Lots of TT ecologists get hired in N. America every year, and the factors that determine the gender balance of new TT ecology hires change very slowly. For those reasons, the headline result is going to be almost exactly the same every year.
Further, I sampled a large majority of the finite population of TT asst. profs of ecology hired in N. America over the past 3 years. The precise size of the population isn’t known for certain, but no matter what it is, the confidence interval around my estimated proportion of 57% women is extremely narrow. If 750 TT ecologists were hired in N. America over the past 3 years (a plausible number), then the 95% c.i. around my estimate of 57% women is 55-59% women. (Normal approximation to the binomial distribution for a sample from a finite population)
So if you’re trying to remember the headline result, you should not remember it as “about 50% women over the last three years”! 50% is well outside the 95% c.i. Nor should you round off and remember the headline result as “about 55% women over the last three years”. That value is barely inside the 95% c.i. You should remember the headline result as “almost exactly 57% women over the last three years“. (See the poll results below to appreciate why I’m banging on about this…)
No, the headline results are not a product of biased sampling
No, I didn’t oversample (or undersample!) positions filled by women. I sampled quite randomly in my effort to census everyone. And even in the extremly-implausible event that the positions I failed to ID have a very different gender balance than those I ID’d, my sample size is so big relative to the size of the finite population that the estimated proportion of women would only change by about a percentage point. See last year’s post for illustrative calculations if you don’t believe me.
The gender balance of recently-hired ecology faculty varies a bit with the research intensiveness of the institution
There’s a trend for less research-intensive institutions to have a slightly greater proportion of women among their recently-hired ecology faculty. Exactly how strong the trend is depends on how you classify institutions. Combining this year’s data with previous data, over the past three years it’s been 54% women at R1s and 59% women at non-R1s. Over the past two years (the only two for which I compiled full Carnegie classification data), it’s 55% women at research universities (R1-R3 institutions) and 63% women elsewhere.
Qualitatively, this result mirrors a long-standing feature of US academia as a whole. (see here for some not-quite up to date data).
Ecologists as a group remain largely unaware of the headline result, or else can’t quite believe it
As in each of the past two years, I polled ecologists asking them to guess the gender balance of this year’s newly-hired TT N. American ecology asst. profs. The poll respondents are a self-selected non-random sample of mostly-ecologists. But it’s a large and diverse enough sample (193 respondents, comprised of 33% faculty, 33% postdocs, 25% grad students, 9% other) to be worth talking about.
Two years ago, the guesses were way too low: the modal guess was less than 40% women. Last year, the guesses were better, because some respondents had seen the previous year’s data, but still substantially too low: the mean guess was 46% women (median 48%). This year’s guesses improved over last year’s, but were still too low: the mean guess was 50% women and the median was 53% women. Here’s a histogram of all the guesses:
You’ll notice that histogram is bimodal. That’s because many respondents hadn’t read (or didn’t recall) the data from previous posts, while others had. Respondents who hadn’t seen or didn’t recall my previous posts were mostly way too low. Here’s a histogram of their guesses:
Respondents who hadn’t read and recalled my previous posts are surely more representative of ecologists as a whole than respondents who had. We have a big audience for a science community blog (thanks for reading, everyone!). But even our blog’s big audience comprises only a small minority of all ecologists, or even just of N. American ecologists.
Further, even respondents who reported reading and recalling my past posts still mostly missed low. Their guesses averaged 53% women, with guesses that missed low vastly outnumbering those that missed high:
When asked for their reasons for guessing as they did, many respondents who reported reading and recalling my previous posts said they based their guesses on those previous posts. But most of those respondents apparently misremembered those previous posts as reporting a lower percentage of women than they actually did! For instance, one respondent wrote of their guess “It is based on previous posts!”–and guessed 52% women. Another wrote “Hoping it is similar to what I think I remember from last year”–and guessed 43% women. Several other respondents said flat out that they remembered last year’s data as 53-55% women (it was actually 59%).
Respondents who reported other bases for their guesses (e.g., the current faculty gender balance in their own department; recent hiring in their own department; their experiences as women in science) were mostly way off. They mostly guessed that recent ecology faculty hiring is heavily male-skewed.
Humorous interlude: the respondent who guessed 58.38316483% women (yes, really) based on “extrapolation from previous years, plus a super-secret algorithm to derive the EXACT answer out to 9 decimal points of accuracy” made my day. THAT WAS THE MOST ACCURATE GUESS IN THE ENTIRE POLL! 😛 (Seriously, it was!)
I also asked respondents to provide subjective confidence intervals on their guesses. Specifically, I asked respondents to complete the sentences “I’d be surprised if the true value was lower than…” and “I’d be surprised if the true value was greater than…” Most respondents reported quite wide confidence intervals. The average width was 26 percentage points among those who hadn’t read or didn’t recall previous posts, and 23 percentage points among those who had. The correct answer (59% women) will surprise 19% of respondents, all of whom guessed too low.
I confess I find these poll results depressing. I write these posts to try to inform ecologists about progress in one specific area of diversity and equity in ecology. And I’ve mostly failed. Word of these posts hasn’t spread at all beyond this blog’s regular readers as far as I can tell. Those regular readers are a small fraction of all ecologists. And the majority of our regular readers apparently either don’t quite believe the data or don’t recall the data correctly. It’s funny: respondents mostly report being very uncertain as to the true percentage of women among newly-hired ecology faculty. You’d think that near-census data would completely swamp people’s flat, “uninformed” priors. But no–even people who read and recalled previous posts mostly guessed too low and still reported considerable uncertainty in their guesses.
My interpretation of this is that, at a subconscious level, ecologists as a whole are incredibly pessimistic about the gender balance of recent ecology faculty hiring. Ecologists mostly say that they’re very unsure what the gender balance of recently-hired ecology faculty is. But deep down, most ecologists are actually sure that it’s heavily male-skewed, probably often without even realizing just how sure they are. To the point that they not only don’t let the data completely swamp their “priors”, but they actually misremember the data.
I’m absolutely not criticizing anyone for this. I can totally appreciate why ecologists, and indeed people in general, would be subconsciously or even consciously pessimistic about, well, anything these days (I sure am…). But understandable as generalized pessimism is, I think it’s a shame if it prevents us from recognizing areas of progress.
When placed in the context of other data, particularly on gender balance at other career stages, these results tell a different and more complicated story about gender diversity and equity in N. American academic ecology than you probably realize
Here is a bit of context for that headline number of 57% women among recently-hired N. American TT asst. profs in ecology:
- Women have comprised 58-60% of US bachelor’s degree recipients in biological sciences every year since 2002, and have been >50% since at least 1995 (source: NSF). Most though not all people who eventually go on to become ecology profs in N. America have US bachelor’s degrees in a biological sciences field (remember, Canada is a small country compared to the US).
- Women have comprised 55-58% of US master’s degree recipients in biological sciences every year since 2002, and have been at or above 50% every year since 1996 (same source as previous bullet). So, a touch lower than their proportion among recent-ish undergraduate degree recipients.
- Women have comprised about 53% of US PhD recipients in biological sciences every year since 2010, and have been at or above 50% every year since 2008 (they were ~38% back in 1995; same source as previous bullet). So, a bit lower than their proportion among recent-ish master’s degree recipients. Drilling down by subdiscipline, in 2015 women earned 51% of US PhDs in ecology, 53% in wildlife biology, and 54% in environmental sciences (source), and they’ve consistently comprised a slightly majority of US PhD recipients in ecology-related disciplines in recent years.
- Women comprised 46.2% of US ecology postdocs as of 2013. Source: Hampton & Labou. So, a bit lower than their proportion among recent-ish US PhD recipients in ecology and allied fields.
For a more sophisticated statistical analysis of the NSF data summarized above, I urge you to read Shaw & Stanton (2012). Great, important paper that deserves to be much better known.
You may also be wondering about post-asst. professor career stages. It doesn’t do much good to hire women at the asst. professor level if they don’t get tenure. For data on that, see Shaw & Stanton (2012). Conditional on reaching the asst. professor stage, women in US life science fields are actually a bit more likely than men to reach the associate professor stage. That’s been true since the mid-1990s. Promotion to associate professor ordinarily is associated with (or else precedes) the granting of tenure. I’m sure you’re surprised to learn this–I was too! (Note that you have to interpret that conditional probability carefully–it definitely does not imply “reverse discrimination” against men at the tenure stage!)
I emphasize that these data are only a bit of the information one would ideally like to have to fully understand the drivers of gender balance along the career path leading to academic ecology (compare, e.g., this deep dive into the drivers of male skew along the career path in academic economics). But with that caveat in mind, the overall picture here is hard to reconcile with any simple “blanket” story about gender diversity on the path to becoming a N. American academic ecologist. Social forces, individual choices, and other factors operating between birth and college graduation skew US bachelor’s degree recipients in the life sciences somewhat towards women. Then, factors operating at each stage between bachelor’s degree and postdoc reduce the proportion of women by a few percentage points at each stage. Then, factors operating in between the postdoc and newly-hired TT asst. professor stage increase the proportion of women ecologists back to what it was at the bachelor’s or master’s degree stage.
So the famous “leaky pipeline” metaphor is now a bad metaphor for the career path in academic ecology, because the proportion of women goes up at some points along the path, rather than always going down. We need a good memorable metaphor for that. Maybe the “sawtooth”?
One topic I hope we’ll discuss in the comments is what these fluctuations imply for efforts to further improve diversity and equity in N. American academic ecology. What efforts would further improve matters, and what efforts would be misplaced?
No, the headline results don’t indicate “reverse discrimination”
As a further bit of context (and because I know someone will ask about it so I might as well bring it up…), no, there’s no appreciable difference in the qualifications of men and women recently hired into N. American TT asst. professor positions in ecology and allied fields. Take it from the guy who has spent more hours than he cares to count looking at the cv’s, websites, ResearchGate pages, and Google Scholar pages of several hundred recently-hired ecologists. Unless you’ve done that too, you should just trust me on this. 🙂 I really wish this didn’t need saying, but it is not the case that N. American colleges and universities are engaging in “reverse discrimination”: systematically hiring unqualified or less-qualified women ecologists as the expense of qualified or more-qualified men. Nor is it the case that the recently-hired TT women ecologists are better qualified than men on paper. FWIW (not much…), here are a couple of very crude bits of data on two aspects of “qualifications” to illustrate this. First, recently-hired men and women ecologists don’t differ in years of experience. The median PhD year for newly-hired (i.e. 2017-18) N. American TT ecology asst. profs is 2014 for both men and women, and the means differ by only 5 months (mean PhD year 2013.4 for men, 2013.8 for women). Second, TT women ecologists hired over the last two years do have modestly lower h-indices than men on average, but that difference goes away almost completely when you account for the fact that more women get hired at less research-intensive institutions. Within institution types, men and women ecologists hired over the past two years had very similar h-indices on average at the time of their hiring (bachelor’s colleges: men 6.9, women 6.6; master’s universities: men 7.7, women 6.7; research universities: men 10.3, women 9.4). As an aside: for a host of reasons that are hopefully sufficiently obvious that I don’t have to list them, you should not look at those h-index data and conclude that recently-hired women ecologists are even slightly less-qualified than men on average. I interpret these data as indicating that there are systemic forces that make women disproportionately likely to obtain teaching-focused faculty positions in ecology, and that people (men or women) who obtain teaching-focused faculty positions in ecology tend to be less productive researchers on average (at least by one admittedly-crude measure) than those who obtain research-focused faculty positions.
Bottom line: these results are good news. They represent real, systemic progress for N. American academic ecology in one specific but important area.
These data are good news! They represent real, systemic progress. Indeed, I’d go so far as to say that TT ecology faculty hiring in N. America doesn’t have a systemic problem with hiring women anymore, insofar as one can judge that from the available data. At a systemic level, failure to hire women ecology faculty in N. America looks to me like a solved problem.
Which isn’t an argument for no longer caring about diversity and equity of course! Presumably, a big reason why it’s a solved problem is because many individuals and institutions take diversity and equity seriously these days. Nobody should stop taking it seriously!
My interpretation of these data (and I emphasize that it is an interpretation) is that these days in ecology, multiple excellent women and men apply for most N. American TT faculty positions (a state of affairs that of course reflects everything that shapes people’s career choices and outcomes at the pre-faculty stages). So search committees for most ecology positions will have the opportunity to choose between strong men and women candidates. To help them choose well, search committees these days get trained about bias, diversity, and equity, and have to obey HR rules designed to ensure fairness (e.g., rules obliging search committees to ask all candidates the same questions in the same order during interviews). And many search committees these days are keen for their departments to become more diverse on various dimensions, including gender. Which is a good thing for them to want. As I wrote last year, individual faculty don’t exist in a vacuum. Departments, and the colleges and universities comprised of them, are institutional wholes that are greater than the sum of their parts (Brian has a half-joking post on this). Those institutional wholes are best able to teach and inspire the full range of students who come through their doors, and best able to pursue new knowledge, if they’re comprised of diverse, complementary mixtures of people. So gender and other personal attributes are among the many things that search committees consider when they get down to making difficult judgment calls about whom to hire from among (typically) 3-5 well-qualified candidates, each of whom would be an asset to the hiring department in their own unique way (i.e. there’s rarely a single “best” candidate). The net outcome is that, at an aggregate statistical level, the proportion of women among recently-hired TT N. American ecology faculty ends up being modestly higher than the proportion of women among ecology postdocs, without any appreciable differences in average on-paper qualifications between newly-hired men and women.
That interpretation is consistent with the small amount of hard data I’m aware of on actual faculty search committee decision-making. For instance, as of 2004-5, a NRC study found that women were receiving 45% of biology PhDs, comprised 25% of TT applicants in biology, got 30% of TT job interviews and 34% of TT job offers.
Does this mean everything is perfect when it comes to diversity and equity in ecology? Obviously not! So let me emphasize all the many things the data in this post don’t tell you. They don’t tell you whether any particular ecology faculty search was conducted fairly or any particular applicant was treated professionally (e.g., some applicants still get asked “illegal” questions). They don’t tell you anything about diversity, equity, and fairness in any other area of science–grant-getting, paper-publishing, university committee service, faculty salaries, invitations to do peer review, seminar invitations, etc. They don’t tell you anything about the day-to-day lived experience of women ecologists–how often they’re talked over or sexually harassed or etc. (individual experiences and aggregate data are complements, not substitutes…) They don’t tell you anything about ecology faculty hiring in relation to other dimensions of diversity. They don’t tell you anything about what’s going on with faculty hiring in any other field, or anywhere outside North America. And they don’t somehow make up for or erase past hiring inequities or ongoing problems in any other area. Nobody should take from this post the message that everything is now perfect for women in ecology (because it’s not), or that people should stop advocating to make things better for women in ecology (because we shouldn’t)!
By the same token, knowledge of all those other things doesn’t tell you anything about gender diversity in N. American ecology faculty hiring. I suspect a big reason that many ecologists mistakenly think that recent ecology faculty hiring remains skewed towards men is because they overgeneralize from unrelated experiences and data. A key broad insight of feminism, and of ecology, is that “everything is connected”. That what happens in one area of life has consequences in other seemingly-unrelated areas of life. So overcoming systemic inequities requires joined-up, systemic thinking. But joined-up, systemic thinking isn’t the same as overgeneralization. It doesn’t mean assuming that Everything Is Terrible. William Gibson (sort of) said that “The future has arrived–it’s just not very evenly distributed.” He wasn’t talking about gender diversity and equity in ecology, but he could’ve been. When it comes to gender diversity at a systemic level in N. American ecology faculty hiring, the future has arrived.
After reading a draft of this post, Meghan made a very good point to me that I’m sharing with her permission. Why does many women’s lived experience in ecology often contrast with the good news of the headline results? Well, one reason among others is that many women faculty are serving on search committees and working hard to make that good news happen. Which can be exhausting and not always rewarding work, for instance if you have to fight against biases against women applicants. I think that point generalizes. The hard work of driving progress can be so exhausting and unrewarding that your day-to-day lived experience hides from you the very systemic progress you’re helping to drive.
If current trends hold, how long would it take for N. American ecology faculty to reach gender parity? A ballpark answer is “about 20 years, or a bit less”.* See the epic footnote if you care how I calculated that answer.
In conclusion, I hope this post is good news to you too, and I hope you’ll remember it and share it widely. As I said, it’s good news only about one very small and specific bit of the world–but a small and specific bit that I’m guessing many of you care a lot about. I would like to think that this good news could be universally known and celebrated among ecologists without diminishing–indeed, enhancing!–ecologists’ collective motivation to continue to work for positive change.
As always, looking forward to your comments. A couple of reminders, to help ensure a productive thread. First, you can comment anonymously if you wish, just fill in a made-up name and email address if WordPress asks you for them (but remember that the moderator–that’s me–can see the IP address from which you’re commenting). Second, we welcome disagreement with the post, and with other commenters, but please treat each other with respect. Personal attacks, offensive comments, and trolling are out. Such comments will be blocked/deleted. A second attempt to make such comments will get your IP address blocked. And you’re on thin ice if your comment seems to indicate that you didn’t read the post. We’re proud to have amazing, thoughtful, professional commenters–let’s all keep up that proud tradition.
*I can’t answer exactly because I don’t know various bits of information you’d need to give an exact answer. But we can ballpark it pretty well by making some simple assumptions. Let’s assume that the total number of ecology faculty remains constant. That’s a reasonable approximation, since it only changes very slowly (aside: slow shrinkage of the total number of faculty would slow progress towards gender parity, while slow growth would accelerate progress) Let’s assume that every faculty career lasts 35 years. That is close to the average length, plus random variation around the average is irrelevant for our purposes, and changing the number slightly would only change the answer slightly. Let’s assume that ecology faculty career length doesn’t vary with gender. That’s a reasonable assumption. For instance, there’s no hint of a gender gap among recently-hired ecologists in how long ago they got their PhDs (see above). And I’m not aware of any data suggesting that women faculty tend to retire at a different age than men (happy to be pointed to data on this). Let’s assume that current faculty are uniformly distributed from 1-35 years into their careers. That’s not actually true, but other assumptions wouldn’t dramatically change the answer. Let’s assume there’s no gender variation in career stage. That’s a conservative assumption; in fact, the average woman faculty member is earlier in her career than the average man, primarily due to the male skew of faculty hiring decades ago. Every time someone’s career ends, they’re replaced with a new asst. prof; that’s what holds total faculty numbers constant. 57% of those new asst. profs of ecology are women. Assuming a constant 57% seems reasonable, given that proportions of women at previous career stages seems to be constant or only changing very slowly. In this hypothetical scenario, in 35 years N. American ecology faculty will be 57% women, no matter what their starting proportion. That’s because 1/35 of faculty are replaced each year and so it takes 35 years (the length of a career) to completely replace the faculty. Gender parity would be reached somewhat before that, but how long before would depend on the starting proportion of women. For instance, if ecology faculty are currently 40% women (that’s just a rough guess based on data here), then if I’ve done my sums right it would take 20 years of hiring 57% women/year to reach gender parity. In reality, it would be a bit faster than that because in the real world ecologists who retire in the near future are mostly going to be men. If you want to reach gender parity substantially sooner, you have to almost stop hiring men. For instance, starting from a faculty comprised of 40% women and retaining our other simplifying assumptions, you’d need to hire 75% women every year to reach gender parity in 10 years. I know no one would actually recommend hiring almost-exclusively women. I’m just running the numbers to illustrate how long it takes faculty gender balance to change. Faculty careers last a long time! (Aside: I feel like everything in this footnote is boringly obvious–is it?)