82

It's no secret that computer science is a heavily male-dominated discipline. It's so male-dominated that some people have given up trying to attract more women. Yet Carnegie Mellon not only has close to 50% women in their undergraduate CS classes, they have lots of female faculty. Why?

I'm also interested in what any other institution that's been successful at achieving something resembling gender parity has done as well.

D.W.
  • 8,221
  • 2
  • 31
  • 57
Allure
  • 127,528
  • 50
  • 325
  • 493
  • Comments are not for extended discussion; this conversation has been moved to chat. Please note that further non-clarifying discussion may be deleted without notice. – eykanal Oct 19 '18 at 14:15
  • "...they have lots of female faculty." I wanted to quantify lots but couldn't find much information about it. Short of going through the web presence of Carnegie Mellon very carefully I guess it's pretty hard to come up with a number there. – NoDataDumpNoContribution Oct 22 '18 at 14:23
  • It's a UK organisation, but the Equality Challenge Unit (ECU) promotes the Athena SWAN charter for equality in higher education and their website includes some example of good practice: https://www.ecu.ac.uk/equality-charters/athena-swan/athena-swan-good-practice-in-heis/ – arboviral Oct 22 '18 at 15:47
  • I remember reading a thread on twitter a couple of months ago by a CMU faculty member where he described in detail what steps they took to do this. I can't remember who that was or locate it... – Bitwise Oct 22 '18 at 19:04
  • 1
    It wasn't always the case that CS was a field dominated by men. Through the 1970's and into the 1980's, computer science was a field that attracted more women than many other STEM fields. https://www.npr.org/sections/money/2014/10/21/357629765/when-women-stopped-coding – Brian Borchers May 06 '22 at 22:28

7 Answers7

79

Carnegie Mellon gives an answer on their website:

The steady climb of women enrolling in these fields at Carnegie Mellon highlights a combination of factors: a strong commitment by leaders at the university, college and department levels; influential pipeline programs for middle- and high-school students; targeted recruitment; closer scrutiny of applications; support and mentorship programs; and attention to diversifying the faculty.

A second, speculative answer: Carnegie Mellon is wealthy. Computer science is where they invest the most. If they want to attract more women to their computer science department, they can use their money to do it. Many of the factors they list cost money. They may be decreasing the diversity of other universities.

eykanal
  • 48,364
  • 18
  • 119
  • 216
Anonymous Physicist
  • 98,828
  • 24
  • 203
  • 351
  • 32
    If they are influencing middle school or earlier, they could equally be increasing the diversity at other universities with the same feed schools. (by supporting the 30% threshold past the age where biases tend to skew the balance) – Sean Houlihane Oct 19 '18 at 11:45
  • 2
    @SeanHoulihane Their impacts could be wider, sure, but I doubt "equal": although I'm not familiar with Carnegie Mellon in particular, "pipeline" programs tend to be pretty strongly branded. Though they may do some general good for society, they are probably also reinforcing explicitly or implicitly "...and your next step is to come to Carnegie Mellon." I don't think it's a bad thing, I'm just skeptical you can estimate the broader impact by looking just at Carnegie. – Bryan Krause Oct 19 '18 at 15:30
  • @BryanKrause, I'm thinking if they improve the pipeline in a single school from an early enough age, that will tend to bootstrap the diversity within the school. Maybe not self-sustaining, but I think there is likely some amplification and it is not at all clear that 2nd choice universities would loose out. – Sean Houlihane Oct 19 '18 at 15:43
  • 3
    Agreed, partly why I said I don't think it's a bad thing. I doubt anyone is really losing out, in any case, I don't think diversity is a zero sum game. – Bryan Krause Oct 19 '18 at 15:45
62

We are a smaller German University of Applied Science and most of our courses have a certain application field in addition to computer science (usually about 70% computer science, 30% application related). The application fields "media" and "medicine" seem to motivate many female students to start studying, even if they later decide to work in the automotive industry or somewhere else.

Furthermore, our university strongly invests in diversity actions, we are trying not to discriminate anyone, and we are trying to promote female students e.g. as student assistants / tutors (which is not hard, since we are usually looking for the best, and most times we are having more female top students than male ones). But for the younger female students it is good to see that others succeeded or are very successful.

Once you crossed the 30% mark, the feeling is that it's not "special" or "strange" if you are female.

One point where we have to improve is the number of female faculty members, but this takes time.

But still, motivating more students to study computer science is a high priority!

OBu
  • 13,113
  • 3
  • 34
  • 59
  • 44
    +1 for “Once you crossed the 30% mark, the feeling is that it's not "special" or "strange" if you are female.” —I suspect this is a huge factor. Theoretically, this is supported by work like Thomas Shelling’s “Micromotives and Macrobehavior” – Dawn Oct 19 '18 at 11:51
  • Comments are not for extended discussion; this conversation has been moved to chat. – eykanal Oct 22 '18 at 01:14
42

At the risk of sounding like a broken record:

I'm especially interested in answers from female computer scientists at CMU.

Well, go ask (some of) them directly - don't expect them to stumble onto this page...

The staff directory is available; you'll need to figure out which ones are women by the photos and the names though, and pick out a few people to contact. Or you could start by contacting the Dean or whatever CMU call the head of the School of Computer Science.

The School of Computer Science at CMU also has a dedicated Women @ SCS website, which you should check out, both for immediately-relevant material and for names of women faculty and management personnel which you could contact directly (i.e. put those at the top of your list of people to contact). Thanks to @user3067860 for noticing that site!

PS - As commenters suggest, you can't just start asking specific questions with long answers right away, you need to make sure whoever it is you contacted is willing to converse regarding these issues.

einpoklum
  • 39,047
  • 6
  • 75
  • 192
  • 3
    Maybe the OP should start here: http://www.women.cs.cmu.edu/ – user3067860 Oct 19 '18 at 14:12
  • 20
    i'm sure all the women in the department would LOVE to have some random interrogating them out of the blue about their experiences. – user371366 Oct 19 '18 at 19:08
  • 3
    @dn3s I suspect the intention was to reach out to the heads of the department and ask about approaches they have used, not to contact to women who did join and ask them why. – Captain Man Oct 19 '18 at 19:22
  • 1
    @dn3s: 1. OP would obviously need to sample a few faculty members. 2. With a properly worded introduction, and a question regarding willingness to discuss this issue, recipients would either politely decline or agree, without feeling interrogated by random people. – einpoklum Oct 19 '18 at 20:29
  • I removed some debate about whether this should be an answer, a comment, a vote-to-close, a chat discussion, or something else. At this point (four years and 40 upvotes later) it is staying as an answer. However, this is a singular decision and does not create a precedent that answers like this are acceptable generally -- if another case arises in future, we can discuss on meta at that time. – cag51 May 08 '22 at 05:38
35

The phenomenon you have observed, where some programs have great difficulty attracting female CS majors (and faculty) and others are approximately 50%, is a classic example of a "tipping point" phenomenon. Tipping point phenomena were modeled in economics by Nobel prizewinning theorist Thomas Schelling. They were popularized by Malcolm Gladwell.

The intuition is this: Everyone faces certain benefits and costs from pursuing a CS degree. The benefits include income, intellectual challenge etc. The costs include hard work, eye strain, etc. In addition, minority individuals face the cost of being an "outsider" in the environment. This "outsider cost" is typically modeled as increasing in the segregation of the field. So a field with 90% men would have a higher outsider cost to a woman than a field with 60% men.

The tipping point comes into play when looking at equilibria. One equilibrium is that there are very few CS women. This is because the outsider cost is so high that only women who highly value the benefits (income, challenge, etc.) have a net positive utility from the CS major. But, if you can nudge the system off this equilibrium and get even a few more women, then you can decrease the outside cost enough that a few additional women find CS net positive. And then these women's presence decreases the costs further, and so on. You progressively tip the scales of cost/benefit such that you reach the second equilibrium of 50/50.

darij grinberg
  • 6,811
  • 1
  • 25
  • 43
Dawn
  • 16,786
  • 1
  • 45
  • 75
  • 3
    Agreed. How do you nudge the system, however? – Buffy Oct 19 '18 at 17:29
  • 7
    @Buffy I think that some of those other answers are covering those factors. I am trying to complement those answers, not provide a fully comprehensive alternative. One thing I haven’t seen addressed is scholarship money (unless this is what was meant by “targeted recruitment”). Basically paying women for bearing the “outsider cost.” – Dawn Oct 19 '18 at 17:45
  • I wonder what those tipping points can teach us about fields that have changed from being overwhelmingly male to being overwhelmingly female (teaching and psychology being two that come to mind). – gerrit Oct 19 '18 at 22:35
  • 3
    It is the same idea in reverse.... eventually the men start having a coat from being in a female industry. A few drop out, then the cost increases, then more drop out, etc. I believe it is called “overshooting” but that might not be right. – Dawn Oct 20 '18 at 02:35
  • 1
    As one other answer points out, CM is spending a lot on achieving that ratio, by investing in school recruitment etc. However this is likely to the detriment of diversity at other schools. Nudging the system on the whole would, if possible, require a state-scoped action. I suspect CS is not something women would choose to do on the whole because CS related work tends to be dry,solitary stuff and that is not in women's nature. – Sentinel Oct 21 '18 at 05:00
  • @gerrit not much. The history is well documented. In the 19th century many new opportunities arose for male workers, in new industries and new services. This led to an exodus from teaching jobs. This led to a vacuum and opportunity for early feminists to make the case for females as teachers. They introduced ideas such as that women are more nurturing, and that children are adorable things (rather than little demons, prone to sickness, problematic nuisances etc). Those ideas stuck. – Sentinel Oct 21 '18 at 05:14
  • Remember that before the 1930s, when antibiotics and vaccinations became commonly administered, 1 in 5 children died by the age of 7. To maintain demographic TFR women typically had to have more children than today. The freedoms today are largely down to medical advances. – Sentinel Oct 21 '18 at 05:16
  • 1
    I find your set of comments puzzling. First you claim that solitary work is not in women’s nature. Then you note that other ideas about women’s nature (nurturing) are socially constructed. I would note than “disliking solitary work” is equally sociallt determined. In any case, CS doesn’t have to be a solitary enterprise. In fact, most of my work is coding and I don’t do it in a solitary way (shared office with other female coders in PhD, now have a constant group chat going as a prof). – Dawn Oct 21 '18 at 13:41
  • 1
    Actually, the degree of switching between schools vs. between majors (either overall or at a particular school) is actually an empirical one. There is at least some literature which suggests that the second set of margins is actually quite important. – Dawn Oct 21 '18 at 14:01
  • 4
    Ehm... Do you have any strong empirical evidence that this is a case of a "tipping-point" phenomenon? As far as I can see, this answer is just a bald assertion. I can think of many other possibilities that may explain this phenomenon. In fact it is outright admitted by many universities, this one included, that they "prioritize" female applications or they put "closer scrutiny on applications." Indeed, there is evidence that there is discrimination in favor of female applications given otherwise identical applications (e.g Williams & Ceci (2015)). – Eff Oct 22 '18 at 08:18
  • Here is some striking evidence of the existence of the costs I describe: https://www.sciencedaily.com/releases/2018/01/180123102002.htm – Dawn Oct 22 '18 at 17:05
  • 1
    Here is some evidence in tipping in occupations: https://faculty.smu.edu/millimet/classes/eco7321/papers/pan%202015.pdf – Dawn Oct 22 '18 at 17:06
  • 1
    @Dawn I must admit that I don't find it particularly convincing because these things are massively statistically confounded by other variables. However, it is an interesting thing to note that if this is true, then people who have been publicly arguing that math, sciences, and CS are gender biased fields are themselves (partly) the cause of the disparities. Because the actual evidence shows that the fields are not biased against women. They are, in fact, massively biased in favor of women. So women might perceive these fields as biased, but they're not. – Eff Oct 22 '18 at 17:53
  • 1
    @Dawn One of the major issues with the first paper you cite is that people rank-order fields by how much people think that these fields are welcoming to women. But this is a hugely problematic measure. People are going to rank fields with many women as being welcoming to women, and fields with few women as being unwelcome to women. It essentially becomes tautological, because it will simply show "fields with few women predicts few women choosing this subject." They also don't control for the most well-known factor related to this issue, namely where people are on the thing-people orientation. – Eff Oct 22 '18 at 18:08
  • 6
    @Eff The evidence says that women perceive this as a cost, which is all that is required for the tipping point phenomenon to occur. In this case, perception is reality, you don't have to show that the perception is accurate or impervious to bias. The fact that the number of women influences perception is not relevant. Similarly, it is not necessary to control for these other factors like people-thing orientation. – Dawn Oct 22 '18 at 18:13
  • 1
    @Dawn You're completely correct that perceived bias doesn't have to be related to the reality of bias. Given a hypothesis of "tipping point," perception of bias may be all that is needed. However, you're not correct that it's not necessary to control for other factors, because this is a massively confounded problem. Women may simply enter the fields that they're interested in, and then these choices will end up correlating with people's perceptions about bias. Therefore, the fact that perceived bias correlates with major choice does not provide very persuasive evidence. – Eff Oct 22 '18 at 18:22
  • @Dawn In other words, what I'm saying is that you would completely expect perceptions of bias to correlate with major choice--even if there is absolutely no causal relationship between perceptions of bias and major choice. Because people are going to perceive fields which women don't major in as more biased towards women. – Eff Oct 22 '18 at 18:25
  • 1
    Let us continue this discussion in chat. – Dawn Oct 22 '18 at 20:40
6

If you want to attract some contingent of students to your institution, then provide an attractive work environment for them. It is really as simple (complex) as that. It doesn't have to be just the male/female breakdown here, but it applies there of course.

First, consider the answer here of user Dawn who suggests a tipping point. But how do you get to the tipping point. Others have suggested money, but I'm pretty sure that is insufficient, though it may be necessary.

First, how many women do you have on the faculty? Is it enough? Why not. Is it harder for them to get tenure? Why is it harder to get tenure? Women (in the US, at least) often have societal demands put on them that they can't avoid and that makes a rapid advance to tenure impossible. In a related case, I had a brilliant female colleague who had to delay her career to take care of an ailing mother. Men almost never have that problem as someone else (a woman) will do that instead. Women also need to have accommodation made for childbirth as no one else will take over that burden. Men don't have to delay family or choose between career and family. Single parents are usually (not always) women. Do you pay young women enough that, especially if they are single parents, they can afford proper childcare without sacrificing their research? Rigid rules around tenure and salary work against you here. Tailor the system so that it is consistent with your overall goals and flexible enough to accommodate special situations.

So, work to build up (and tenure) your female (or other "minority") faculty.

Next is the question of how welcoming the environment is for students. Look not only to universities with a better balance, but look also to small women's colleges (and Historically Black colleges). Why do students want to go there? The faculty has a lot of women, of course, but not all. However, the environment has a lot of people who "look like me." (Tipping point again). But, beyond that, those places also seem to me to be better at mentoring young students. In many of them the faculty and students are on first name basis with one another. The professor becomes a role model, not just a teacher.

So, make sure mentoring happens. This can work for everyone, not just women.

Next, is the actual work schedule sane or insane? Do you require long hours in the lab that may be possible for men (who have external support for the daily tasks of living) and not so much for women (who often are left with those tasks. Does your environment lead to burn out for some? That is unhealthy for everyone and leads to extreme stress. Some questions on this site suggest that it is more common than it should be. Some of the techniques for reducing stress, by the way, require additional time and effort. That may be harder for those with outside responsibilities. On a related note, is your work schedule so intense and competitive that backstabbing among students (or even with professors) happens? Do you have a way to make that stop? A number of questions on this site indicate that it happens more than it should. Even people "stealing" the research product of others.

So, assure that the work schedule is sane and not overly stress inducing.

Next, how are women treated day to day by both faculty and other students. Is there some low level of sexual harassment that is tolerated? Are you sure? There are faculty who prey on students, of course. Are you sure that you have a way to both recognize that and to force it to stop? Do you have formal standards of conduct that apply to both faculty and students. Note that mentoring (of both males and females) can help here. A quiet word from a mentor can change behavior when needed.

So, make sure that no one needs to deal with the "ick factor".

Next, who's questions and opinions are valued? Is it only extroverts who get a chance to say anything or to represent a working group? Does your faculty have sufficient training to assure that every student's ideas are valued? Fairly simple, but required, teaching seminars can improve the classroom/lab environment. They will probably be resented by some, but you have to ask, what is important? If you only value high pressure ego driven research progress, you will attract only people who are comfortable in that environment.

So, consider your values and find ways to operationalize them, not just talk about them.

And note that you don't have to sacrifice research quality or output to do these things. Take CMU, for example.

For what it is worth, I never had a situation in which I thought a female colleague was less able than her peers, including myself. I have a pretty big ego, of course.

And yes, I realize that I answered a slightly different question.

Buffy
  • 363,966
  • 84
  • 956
  • 1,406
5

To your question of other institutions getting gender parity: Until visa and exchange rates started making things harder, Australian universities had a lot of Iranian postgraduate students. I noticed that the proportion of women in this cohort was very high, and when I asked some of those (female) students about this they told me that computer science is a popular subject of study for women in Iran, and that they felt very accepted in the field. Articles such as this one from Forbes seem to bear this out. I hesitate to draw any conclusions from this, but it's an interesting example of a place where computer science in not assumed to be a male-oriented field.

Peter Mortensen
  • 836
  • 7
  • 7
beldaz
  • 555
  • 2
  • 8
3

At my low-diversity STEM Ph.D. program (arguably top 10), the biggest change we made during recruitment week (a week where accepted student applicants visit the school) was that we hosted a women-only social event. Current female Ph.D. students in the program chatted with female recruits socially. No men were present. Shockingly that year we went from maybe 20% of female recruits accepting offers (in the past) to 100%. I think the event showed that we as a program acknowledge that being female in our field comes with additional challenges, and having a safe space to talk about those challenges is important. I think it also created a sense of social cohesion with the female recruits. They liked each other, they liked the possibility that they could all go to school together. They also got to ask really important questions about the program that they may have felt uncomfortable asking if there were men present. Literally, nothing else changed, this one small change to recruitment week had such a massive effect.

WetlabStudent
  • 10,674
  • 5
  • 42
  • 61