Columbia University, where I work, is a fantastic place to do science. But like at most other institutions, gender bias is a pervasive issue. I’ve repeatedly called out various seminar series at Columbia for having an absurdly skewed gender ratio (favoring men, of course).

lol the @Columbia @ZuckermanBrain Neurotheory seminar speaker lineup has fewer women than people named Michael: https://t.co/anJdThKkKo

— Anita Devineni (@BrainsExplained) October 19, 2017

This semester, they have an equal number of women as Dmitri(y)shttps://t.co/RtPiTLljiv https://t.co/LyQzxbtiUX

— Anita Devineni (@BrainsExplained) January 16, 2018

So when I recently had the opportunity to participate in selecting next year’s speakers for the Columbia Neuroscience seminar series, I was excited to get an inside look into the selection process and try to advocate for a more diverse set of speakers. By diverse I mean in gender as well as race/ethnicity, seniority, scientific approach, etc., but today I’m just going to discuss gender. (I’m well aware that race in particular is an even bigger issue than gender in most areas of science.)

As an aside: Hopefully it’s obvious why gender balance in seminar speakers is important, but here are just a few reasons. 1) A skewed ratio is a clear message that men are more highly valued in the field than women. 2) Being invited to give an extramural talk, especially at a prestigious institution like Columbia, enhances a researcher’s profile and may help them forge valuable collaborations. 3) Female speakers serve as successful examples and potential role models for female trainees.

Being part of the selection process was very interesting. First of all, I was able to see a list of all the speakers that were nominated by members of the Columbia neuroscience community as well as the people who nominated them. As a data junkie, I decided to analyze the gender ratios of both the nominators and the nominees. After I describe this analysis I’ll then discuss more about how the speakers were actually selected by the committee.

Overview of the nomination data and analysis

Email requests for speaker nominations were sent out multiple times to students, postdocs, and faculty in several departments related to neuroscience. Ultimately 154 nominations were submitted (although 1 entry failed to include the nominee’s name).

Often the same person submitted multiple nominations, and some nominees were nominated multiple times. This analysis focuses on each of the 154 nominations as a separate data point, so an individual nominator or nominee may be represented multiple times.

Nominators were asked to list their position (student, postdoc, faculty, etc). Based on their name, picture, and/or personal acquaintance, I was able to determine the apparent gender of 145 of the 154 nominators and of all 153 nominees. Nominators whose gender was not determined were excluded from the analysis below, though their nominees were still included.

Men submitted many more nominations

Let’s start with a breakdown of the nominators (people submitting a nomination).

25% of nominations were made by graduate students, 32% by postdocs, 40% by faculty, and 3% by people holding other positions (e.g. staff scientist). Without more information I’m not sure if any of these groups were overrepresented.

Breaking down the data by gender reveals a very clear effect: 71% of nominations were submitted by men and 29% were submitted by women. This is a huge gender gap that I didn’t expect to see.

I thought maybe this gender gap was due to the large number of faculty submissions, since the faculty at Columbia (like most institutions) skew heavily male. For faculty submissions, 77% of nominations were submitted by men and 23% by women. So this is a slight overrepresentation of male nominators compared to the overall average of 71%.

I believe the gender balance of postdocs and graduate students in the department is relatively even, so I would expect that equal numbers of men and women from these groups would submit nominations. Compared to submissions from faculty, submissions from postdocs showed a far less skewed ratio: 58% of nominations were submitted by men and 42% by women. But things flip the other way when we look at nominations from graduate students: 79% of nominations were submitted by men and 21% by women.

So the skewed ratio of nominators cannot simply be attributed to the overrepresentation of men holding faculty positions. For some reason male students were far more likely to submit nominations than female students, and postdocs showed the same trend as well.

Main conclusion: men submitted nearly three-quarters of the nominations for seminar speakers.

 

Men were much more likely to be nominated, especially by other men

Now let’s break down the nominees (people who were nominated as speakers).

73% of nominated speakers were men, whereas 27% were women. This is a huge and disappointing gender gap, though it’s about what I expected to see.

Are male nominators more likely to nominate male speakers? The answer is a resounding yes.

Of nominations submitted by men, 78% were for male speakers and 22% for female speakers.

Of nominations submitted by women, 55% were for male speakers and 45% for female speakers.

As I suspected, women were much more likely to nominate female speakers. Of course the data can’t say whether this is intentional (women are actively trying to rectify the gender balance) or subconscious (women are more likely to respect female scientists). Either way, it seems that encouraging more women to submit speaker nominations might help narrow the gender gap in nominees.

I also analyzed whether any specific group – students, postdocs, or faculty—was more likely to nominate female speakers. The numbers were pretty even: female speakers got 28% of nominations from faculty, 27% of nominations from postdocs, and 26% of nominations from graduate students.

Main conclusion: female speakers received only about a quarter of the nominations, and they were twice as likely to be nominated by women than by men.

 

A proactive approach can narrow the gender gap

So with a list of nominated speakers that was 73% men, who ended up getting selected (and how)?

I’m probably not supposed to reveal too much about the selection process. But one thing I’ll say is that at the start we were told that the main criterion was to select exceptional scientists, and diversity was not mentioned until a female postdoc brought it up. I was the only one to point out that three-quarters of the nominees were men and we would have to proactively advocate for women speakers if we really wanted gender balance.

The selection committee broke up into groups covering different areas. Within my group there was definitely a focus on selecting similar numbers of men and women, although some participants seemed to care much less about this issue than others.

The final list that I saw included 20 men and 14 women, representing a ratio of 59% men to 41% women. The committee members seemed to think this was a job well done.

Sure, this ratio was a lot better than the ratio of nominees that we started with, but I believe we could have done better. There were plenty of excellent women nominees who hadn’t been selected, so there was really no reason to settle for a skewed ratio at all. I tried to say as much, but by the time the gender ratio had been tallied many people had already left and certainly no one was willing to go back to the drawing board.

 

Final thoughts

Overall, this was an enlightening inside look into how gender bias in seminar speaker selection seemed to occur at every step of the process:

Men submitted most of the nominations, and most of the nominees were men.

Anecdotally, women on the selection committee generally seemed much more concerned about gender balance than the men (with some exceptions, of course).

Even with a proactive approach (led mainly by women) the committee ultimately decided to invite significantly more men.

I hope the data and stories I’ve presented here help provide some insight into why many seminar series and conferences end up with far more men than women. I know there are tons of resources out there to address this issue, including advice for achieving an equal ratio, lists of women scientists, and websites that keep track of gender ratios at conferences. As scientists we are used to solving problems, and this is a problem that shouldn’t really be so hard to solve if we set our minds to it.

 

Do you have experience on selection committees or further suggestions about how to address gender bias in speaker selection? Leave a comment below!