Newly-hired tenure-track N. American asst. professors of ecology are 59% women (no further updates)

(Last update: June 3, 2018. No substantial changes since the original post, only very very minor quantitative changes. A recent update adds a link and brief discussion of new data on the gender balance of ecology PhD recipients and postdocs. Those data provide useful context for the data I compiled.)

Like last year, 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. Click the link to the poll for details on how I compiled the data, and why I went with a gender binary even though that’s not ideal.

I’ll present the results first, then the poll results, then some discussion.

Warning: long post ahead. That’s because I’ve tried to discuss the results thoroughly and carefully, and to anticipate and address questions that readers are likely to have. You really should read on, but if you just want the headline results:

• 59% of tenure-track asst. professors of ecology hired in N. America in 2016-17 are women. Combined with last year’s data, it’s 57% women over the last two years. If anything, that’s probably a slight underestimate, for reasons explained below. This is good news!
  
• Ecologists as a group remain unaware of this; many think recent faculty hiring in ecology is <50% women.

Results

I ID’d the gender of 195 tenure-track asst. profs of ecology hired in N. America in 2016-17 (that’s out of ~250 positions I tried to ID; I tried to ID every position in ecology and allied fields). 114 of them (59%) are women. As an aside: those 194 include a very few profs who were actually hired the year before but whom I counted in this year’s data because it was easier than adding them to last year’s dataset and rewriting last year’s post yet again.

We can combine these data with last year’s data on hiring in 2015-16. Last year I identified 193 positions, 54% of which were filled by women (note: it was 51% at the time I originally published last year’s post, but as additional data came in it went up to 54%). That means that over the last two years, 57% (219/387) of newly-hired tenure-track asst. professors of ecology in N. America have been women.

This is a sample, not a census. Further, it’s a sample from a finite population, though one for which the exact population size isn’t known. If we assume that the population size from which the combined data set was sampled is 500 positions, then the finite population 95% confidence interval around our estimate of 57% women is 54-59% women (normal approximation to the binomial distribution, which is fine because we have a big sample size and an estimated proportion sufficiently close to 50%). The bounds of that interval don’t vary much for any plausible assumption about the size of the total population. For instance, the 95% c.i. only expands to 53-60% women if you assume that 700 rather than 500 tenure-track asst. profs of ecology were hired in N. America in the past two years. Conversely, the 95% c.i. shrinks to 55-58% women if you assume a population size of 400 positions. So we can infer with high confidence that recent tenure-track faculty hires in ecology in N. America are >50% women. (note: see below for discussion of possible sampling bias; men are probably slightly over-represented in my sample)

You’re probably wondering if the gender balance of recent tenure-track ecology hires varies among institutions of different types. Across US academia as a whole, women have long been more underrepresented among faculty at large research universities than at small colleges. There’s a trend in that direction among recent ecology hires, but you need to combine data across multiple years for it to come out close to significant in a chi-square test, and its apparent strength varies depends on how you classify institutions. Finer-grained classifications lead to smaller sample sizes and so more sampling noise, but coarser-grained classifications may average away some real variation. In 2016-17, 42/74 (57%) of tenure-track asst. prof ecology hires at R1 universities (or their Canadian rough equivalents) were women, vs. 72/120 (60%) women at non-R1s. Combining with last year’s data, you get 81/157 (52%) women at R1s vs. 138/230 (60%) at non-R1s, a close-to-significant difference in a chi-square test. But if instead you define “research universities” as R1+R2+R3 universities and only look at 2016-17 hires, you find that 65/114 (57%) of hires at research universities were women, vs. 45/74 (61%) women at non-research universities, a difference nowhere close to significant (note that this last calculation omits hires at a few Canadian unis I wasn’t sure how to classify).

In light of the fact that recent ecology hires at less research-intensive institutions are more likely to be women than are recent hires at research universities, it’s likely that my data slightly underestimate the proportion of women among recent ecology hires. That’s because the positions for which I was unable to identify who was hired are mostly at less research-intensive institutions. Here’s a rough back of envelope calculation to show that the bias, if it exists, should be modest. If you assume that 500 tenure-track asst. profs of ecology were hired in N. America in the last two years, of whom I identified 387, and that the unidentified ones were 61% women (roughly as expected if the unidentified ones were mostly at less research-intensive institutions), then the true proportion of women among tenure-track asst. professors of ecology hired in the last two years is estimated to be 58% rather than 57%. Even if you assume, implausibly, that the population of ecology hires was much larger (say, 800), and my sample was more statistically-biased (say, the unidentified positions were 65% women, or going the other direction only 50% women), the corrected estimate only changes by about 4 percentage points in one direction or the other. I can’t see any plausible way the true proportion of women hired into tenure-track positions in ecology in N. America in the last two years could be less than 50%. Taking into account plausible sampling bias and sampling error, I’d be quite surprised if it’s less than 54-55%.

I didn’t break down the results by subfield of ecology. Subfields are too loosely defined for that to be a useful exercise.

Before anyone asks: no, I didn’t also compile data on race/ethnicity. In my view that would be impossible to do based on the sort of publicly-available information I was looking at. For what it’s worth, a substantial majority of recent ecology hires probably identify as white, but I wouldn’t venture to be any more precise than that. This is an important issue, but I don’t have data that speaks to it.

A read-only spreadsheet of all the positions I checked is here. I welcome corrections and additions, and will edit the post as needed in light of them. Please email me with corrections and additions at jefox@ucalgary.ca. I’m the only one with permission to edit the spreadsheet because I want to ensure the accuracy of the information.

Poll results

So what did poll respondents think I was going to find? I was curious about this because last year’s poll respondents were mostly way off. Last year’s respondents mostly thought that recently N. American tenure-track ecology hires are strongly male-biased. And last year’s true value of 54% women was outside the majority of respondents’ subjective confidence intervals around their own guesses.

This year’s guesses were much more accurate, but still way too low. The mean guess this year was 46% women (median 48%), with guesses that missed low far outnumbering those that missed high:

This year I asked respondents for some background information, and about how they came up with their guesses. Respondents who’d sat on an ecology faculty search committee in the last 5 years, and those who read and recalled last year’s results, were a bit more accurate than other respondents, but only by a few percentage points on average. Interestingly, some attributes greatly reduced the variance of guesses. Respondents who read and recalled last year’s results, and respondents who’d sat on an ecology faculty search committee in the last 5 years, rarely provided extreme guesses that were way off. Faculty and postdoc guesses also exhibited less variance than guesses from grad students or people in other employment, though the gap wasn’t huge.

Interestingly, the average guess among respondents who read and recalled last year’s post was only 49% women, i.e. a bit lower than the fraction of women in last year’s sample (even at the time when I first published last year’s post).

To their credit, respondents this year also expressed much more uncertainty about their guesses than last year’s many overconfident respondents did, though in retrospect I think last year’s poll was structured poorly and so nudged some respondents into excessive confidence. This year’s result of 59% women will come as surprise to only 34% of respondents, based on their own statements of what result would surprise them. The upper bound of many respondents’ subjective confidence intervals was 60% women, so this year’s results came close to surprising many respondents.

Many but not all respondents who read and recalled last year’s results reported basing their guess this year on those results. A few respondents said they used last year’s results as a starting point, then dialed down or (more rarely) up based on other considerations. Those adjustments mostly made their guesses less accurate. Many other respondents reported basing their guesses on their own department’s recent hiring, or on anecdotal knowledge of recent ecology faculty hires at other institutions. Their guesses were both more variable and less accurate than the guesses of people who based their guesses on last year’s results–a small illustration of the risks of overgeneralizing from small samples. Some respondents started from the fact that women comprise a majority of US ecology graduate students, and then dialed down based on what they knew, or thought they knew, about retention rates. A few respondents based their guesses on other experiences and information, such as seeing many women at the ESA meeting, reading about sexism in science, and their impression that diversity and equity are taken seriously these days.

I got a chuckle out of the respondent who guessed based on “reading coffee grounds”. 🙂 This respondent’s guess was way off, suggesting that a switch to tea leaves may be called for. 🙂

Discussion

This year’s data confirm the good news of last year’s data. They represent real, systemic progress. If you’d collected the same data even a decade or two ago, I doubt you’d have found 57% women among recent asst. professor hires in ecology in N. America!

I don’t think these data are an argument for complacency. Presumably, recent tenure-track faculty hiring in ecology is 57% women at least in part because many individuals and institutions take equity and diversity seriously. We shouldn’t stop taking it seriously.

On their own, these data don’t tell you anything about whether or not ecology faculty hiring decisions are gender-neutral relative to the applicant pools, because I lack data on applicant pools. But a back of the envelope calculation suggests that the applicant pool for tenure-track ecology professorships in N. America is less than or equal to 57% women. Women earn a majority of the US PhDs in biological sciences (e.g., 55% in 2015); that’s been true for over a decade. (Note that Canada may be a bit different; in 2007 [the most recent year for which I could find data], only 47% of new Canadian life science PhD recipients were women. But PhDs from Canada and elsewhere are a small fraction of new ecology faculty hires in N. America and so their demographics don’t make a big difference to my rough, back-of-the-envelope calculation). Drilling down, in 2015, women earned 51% of US PhDs in ecology, 54% in environmental sciences, and 53% in wildlife biology, and women comprise a slight majority of all US ecology PhD recipients since 2000. Collectively, those data show that recent-ish PhD recipients in ecology and allied fields are a bit less than 57% women. And the analysis of NSF’s massive long-term dataset in Shaw & Stanton (2012) indicates that, in US biology as a whole as of 2006, women PhD recipients were a bit less likely than men to go on to a postdoc. Now, Shaw & Stanton didn’t have ecology-specific data, and that the gender gap in transition from grad school to postdoc likely has narrowed since 2006 because the historical trend of slowly-but-steadily narrowing gender gaps presumably has continued. So at a rough guess, I think the overall applicant pool for N. American ecology faculty positions probably is roughly gender-balanced. A bit male-skewed or a bit female-skewed seem quite possible too. UPDATE: we no longer have to guesstimate the gender balance of ecology postdocs based on the data in Shaw & Stanton, thanks to the new paper from Hampton & Labou, which shows that of 2013, 46.2% of ecology postdocs in the US were women.  In light of the new data from Hampton & Labou, I’d now guess that the applicant pool for N. American ecology faculty positions is a bit male-skewed. FWIW (probably not much), the 113 anonymous faculty job seekers who chose to share their demographic information on ecoevojobs.net last year were 48% women. And FWIW (again, probably not much), what I hear anecdotally from people who’ve sat on ecology faculty search committees at research unis recently is that applicant pools tend to be gender balanced or skewed towards men (e.g.). What little data I’m aware of on gender balance of applicant pools vs. hires in biology also suggests that, recently, TT faculty hires in biology skew a bit more female than do the applicant pools. For instance, as of 2004-5, a NRC study summarized here found that women were receiving 45% of biology PhDs, comprised 25% of TT applicants, got 30% of TT job interviews and 34% of TT job offers. Obviously, the gender balance of the applicant pool for any particular search might well deviate from that of the overall pool. But based on the data available, I’d be quite surprised if the applicant pool for N. American tenure-track faculty positions in ecology is as or more skewed towards women than recent hires are. I’d be extremely surprised if the overall applicant pool is a lot more than 57% women, or a lot less.

Assuming that rough speculation is correct, I’m totally fine with that state of affairs. I certainly wouldn’t say it represents “reverse discrimination” if women are a bit overrepresented among recent tenure-track faculty hires in ecology compared to their representation in the applicant pool. I mean, I hope this doesn’t need saying, but the women ecologists being hired into faculty positions are just as qualified as the men and just as good at their jobs! Further, you can’t really rank faculty job applicants on a single “quality” dimension. Especially when it comes down to choosing among the strongest candidates in the applicant pool at the interview stage(s). Typically, the competitive candidates for any given faculty position will vary on various dimensions that can’t be collapsed into a single one, with no one candidate being obviously preferable to the others on every dimension. Personal attributes like gender should be among those dimensions. After all, 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.

Assuming that my rough speculations above aren’t way off base, the results presented in this post fit with my admittedly-scattered reading of the literature on gender bias in academic science. My impression is that, in contexts in which scientists and their work are formally evaluated by other scientists, the evaluations mostly are gender-neutral (there are important exceptions). I’m thinking of pre-publication peer review, grant and fellowship application reviews (e.g.), and faculty hiring. Yes, everyone has subtle unconscious biases–but those can be overcome by people making a conscious effort to be fair, and by following formal procedures designed to ensure fairness. For instance, asking all faculty job candidates the same questions in the same order in interviews, and using score sheets in the initial stage of ranking faculty job applicants. When I think about gender bias and equity in academic ecology at a systemic level, I tend to worry more about academic ecologists and their work being evaluated casually by people who aren’t necessarily trying to be fair (e.g., student evaluations of faculty teaching). I worry about people choosing reviewers, visiting speakers, symposium panelists, etc. by just picking the first people who come to mind. And I worry most about the systemic forces that constrain and shape individual career choices and that can differentially affect women (e.g., parental leave policies, division of parenting effort).

Hypothetically, if current trends of increasing representation of women among ecology undergrads, grad students, postdocs, and new faculty continue, 20-40 years from now academic ecology might become predominantly women. There’s precedent for that, I think. My possibly-faulty recollection of long-term historical data on gender balance of different professions is that professions that move from very skewed to gender-balanced rarely stop there. Over the course of several decades, they more often cross over and become skewed in the other direction (sorry, can’t find the link to the paper in which I read this). Computer programming is one example, but there are others if memory serves. And hypothetically, if academic ecology ever did become very skewed towards women, I’d consider that a problem. But just because “crossover” has often happened in the past is no guarantee it will continue to happen in the future, because the future background context will be different. For instance, people have different attitudes about gender than they did decades ago, and today there are many more women in the workforce in a greater range of professions than there were decades ago. Plus, academic ecology is still a long way from becoming mostly women, if it ever happens at all. So I don’t see any practical reason to worry now about that hypothetical, distant future. I’m fine with people speculating about the distant future just because they’re curious and like speculating, though.

That recent tenure-track faculty hiring in ecology is 52% women at R1 universities vs. 61% women at other sorts of institutions isn’t surprising. I suspect it is a real difference and not just sampling error or sampling bias, because it parallels a longstanding pattern for US higher education as a whole, across most every field: women are a higher proportion of faculty at less research-intensive colleges and universities. And that gap hasn’t been closing as far as I know, though of course I lack the long-term data to say whether it’s closing in ecology specifically. Rather, the slow, steady long-term increase in women’s representation among faculty has taken the form of parallel increases in representation at more and less research-intensive institutions. Curious to hear folks’ comments on this–is it a problem, and if so what should/could be done about it?

I tentatively suggest that these results weaken the argument for blinding the initial stage of faculty job searches. It seems to me that, if current hiring procedures are working well, at least at the aggregate level and at least as measured by this one metric, then we shouldn’t fix what isn’t broken. On the other hand, there are other arguments for blinding based on other considerations. But I don’t have strong feelings on this and I’m curious to hear what others think.

I’m glad that poll respondents are now much more aware of the current gender balance of recent tenure-track faculty hiring in ecology than last year’s respondents were. But as a group they still mistakenly believe that recent hiring is skewed towards men. And the people who read last year’s post are a small fraction of all ecologists; others remain even more mistaken in their beliefs about the gender balance of recent ecology faculty hires. Between last year’s and this year’s poll results, I bet that the typical ecologist mistakenly thinks that recent tenure-track faculty hiring in ecology is 40ish percent women, rather than 57%. It seems like a shame to me that people’s impressions of systemic progress lag behind actual systemic progress. But it’s understandable, and I certainly wouldn’t criticize the poll respondents who guessed incorrectly.

Finally, to prevent misinterpretations of these data, here’s a list of all the things these data don’t tell you:

• What’s happening at any other career stage, such as tenure. If you’re interested in that, you need to read Shaw and Stanton (2012). In particular, did you know that the odds that an asst. professor of biology progresses to the associate professor stage (which is typically associated with receipt of tenure) are higher for women asst. professors than for men asst. professors in the US and have been since the early 1990s? (I know, it surprised me too when I first learned that!) You have to be careful with how you interpret that conditional probability, of course; you definitely should not leap to the conclusion that there’s “reverse discrimination” against men at the tenure stage. But on its face, it provides some grounds for optimism that progress in diversity and equity in faculty hiring at the assistant professor level will carry through to higher ranks. More broadly, Shaw & Stanton quantify how much of the change over time in gender balance at each academic career stage is ultimately attributable to changes in the gender balance of undergraduate degree recipients, vs. changes in people’s probabilities of progressing to the next career stage given that they’ve reached the previous one.
• Anything about any particular search or department, or the treatment of any particular applicant.
• The experiences of individual ecologists in their day-to-day professional or personal lives. Nor do these data substitute for or devalue discussions of individual experiences. For instance, that recent asst. prof hiring in ecology is 57% women does not mean that women no longer have horrendously sexist things said to them or about them, and doesn’t somehow make those experiences any less horrible. It doesn’t mean nobody ever gets asked “illegal” questions in faculty job interviews, or make it ok to ask such questions. Etc. Conversely, I don’t think individual experiences substitute for aggregate data like these. For instance, that some people still say horrendously sexist things in 2017 should not lead you to infer that recently-hired asst. profs of ecology must be mostly men–because they’re not. I think both aggregate data and discussions of individual experiences have important and complementary roles to play in discussions about diversity and equity.
• Anything about recent faculty hiring in other fields, or about recent ecology faculty hiring outside N. America.
• Anything about your own personal chances of obtaining a faculty position in ecology.

Protips for anyone wanting to emulate this exercise

If you’re inspired to do a similar exercise in your own field, here are two tips. First, you need a big sample size to estimate a proportion with any precision. Even something like, say, 50 positions is not a big sample, unless that’s close to a census of all the recent hires in your field. If I’d stopped after the first 50 or so positions I ID’d this year, I’d have incorrectly reported that 2016-17 asst. prof. hires in ecology were 65% women. Second, don’t just rely on crowdsourcing the data from social media. The subset of positions that I was informed of in response to my blog posts and tweets skewed heavily towards positions filled by women, to a greater extent than my full sample. That likely reflects the fact that this blog’s audience, like a large chunk of science Twitter, includes a lot of people who care deeply about gender and equity issues. Possibly, if you did the same exercise for a different field, asking around on social media would give you a sample biased towards positions filled by men; I dunno. None of which is a criticism of people on social media! It’s just a reminder that people on social media are a small and statistically-biased sample of whatever population you’re trying to crowdsource data on. If you want to emulate this exercise, I recommend seeking data via several different means, including but not limited to asking around on social media. Honestly, if you really want to do it right, you’re probably going to have to do what I did: suck it up and do a lot of googling and a lot of browsing of department websites.

Conclusion: please comment

As always, I look forward to your comments. In order to encourage open discussion and to ensure that my eventual replies are carefully considered, I’m going to stay out of the comments for the first few hours. I’ll also remind all commenters to treat one another with professionalism and respect. Disagreement with one another, and with the post, is fine. Indeed, I’m sure that at least some of you disagree with at least some of my comments on the results, and I hope to hear that disagreement. I certainly don’t think my own views about these results are the only legitimate or defensible ones! But personal attacks are right out, obviously. Also, please keep in mind that overheated rhetoric runs a risk of being seen as an attack. And remember that it’s best to ask for clarification first before leaping to the conclusion that I or another commenter believes something that seems to you to be seriously wrong or otherwise bad. We’ve always had a great commentariat, let’s all keep it up.