Thursday, April 18, 2013

Let me tell you what I know about gender and CS

I was very excited when one of the CS professors in my department, who had admitted he wasn't sure about the causes of gender disparity in CS, accepted my offer to share what I knew. Studying this topic has been my hobby since I stepped into my first core major CS classes and wondered where all the women went.

I've found that professors and others often want to talk about or implement their idea for a partial solution, like a mentoring program or curriculum changes, right away, without taking time to look over the root causes. It was encouraging that this professor was open to learning about the research in the field before taking action.

I based the main arc of my presentation on a book chapter by Whitecraft and Williams that Greg Wilson of Software Carpentry was kind enough to forward to me. It's an evenhanded look at much of the research in this area, including theories that are often out of favor in most places I frequent. It served as a great overview, though I felt it could have focused more on issues involving differences in prior programming experience pre-college and intimidation brought on by "nerdy strutting". (Update: I just discovered a fantastic 2012 report by NCWIT that can also serve as a great overview. It covers cultural issues more comprehensively, with more recent research and more focus on the pre-college years.)

I printed resources from NCWIT for the meeting as well. They have made the process so straightforward with their workbooks on how to increase diversity in a CS department, I can't figure out why they aren't more widely known!

The experience also reminded me how much I don't know. I ran across all the women in computing posts on Mark Guzdial's rather comprehensive blog, and realized there are so many more details to this complex problem! And I really need to read Jane Margolis's books themselves instead of everyone else's excerpts of her books.

Here is the presentation I shared with the professor, plus several slides I added afterwards. It's a bit UofA-centric, but maybe it can be useful to someone. I'm open to discussion, so let me know if you have comments :)

(Here's a link to the presentation if the viewer doesn't show up)

My hope is that this meeting will prove useful for my department. 


  1. Your slide deck is really excellent. A really good piece of persuasion and public policy.

  2. Slide 18? That's why I only took one CS class... 1984.

    I'm going off to drink now.

    1. Enjoy your beverage!

      It's a tricky problem, because you don't want to insult someone who doesn't know that they are being obnoxious because they just don't have the social skills. But that behavior can have really bad effects, as you are well aware. I was thinking that a rule-based approach could be best for everyone, but I'd be interested to see if others have a better solution to the nerdy strutting problem.

  3. None of this made much sense to me until I learned about personality types. It turns out that personality type greatly influences what careers a person would enjoy and personality types are not evenly distributed by gender. The gap is particularly wide for INTJ and INTP, the personality types you usually think of when you think about computer geeks.

    I expect that some of the oft-cited differences between men and women would go away if the researchers controlled for personality type. For instance, the quote on slide 17... if you are female, and fit the obsessive, socially-awkward geek profile, and read stuff like that, you probably conclude that you aren't normal for a woman. Whereas if you knew those are typical traits of a young INTJ of either gender, you won't need to wonder what's wrong with you. Or if you don't fit that profile, you don't need to wonder how to fix the people you're surrounded by, as there's plenty of material available that explains what motivates people depending on the personality type they have.

    Perhaps this information is not terribly useful for defining policies, and clearly it's not the only factor contributing to the male/female ratio in CS. However I've found it extremely helpful for learning how to navigate social environments I don't entirely fit into. Mastering interpersonal communications between differing personality types is at least immediately actionable. However, I do think it's more interesting (and perhaps more effective) to occasionally re-frame the overall debate in terms of encouraging more generalized cognitive diversity rather than gender diversity specifically. I suspect a lot of guys find CS culture alienating for the exact same reasons.

    1. I like how you're framing the debate in personality types, because it brings a new angle to the problem and makes the guys' side more visible. I agree that guys are affected by this same stuff, but it's not as obvious that certain guys are missing from CS. Gender can serve as an easy proxy for diversity in general I think—like Michael Schwern says in this great video, if there aren't any women, who else are we missing? I'm not sure what the research says to support the idea of personality types, but I know my fake internet Meyers-Briggs profile is scarily accurate. I'm undecided if it's in a horoscope sorta way or in a window to my soul sorta way :)

      You also bring up a good point about "wondering what's wrong with you". I'm a bit worried about socially-awkward, geeky, and/or obsessive people feeling like they are being indirectly insulted in these sort of debates, because of talk about the geek factor turning people off from CS. This is a delicate issue, and I've already gotten some negative feedback in this vein from a woman who is proud of her love of video games and Star Trek. My hope is that an emphasis on avoiding arrogant behaviors first of all, and being sure to representing all interests in the CS sphere, both classically geeky and mainstream, will help the situation without alienating anyone.

    2. It might be worth noting that I'm a Star Trek fan too. DS9 anybody?

  4. Possibly the most alienating thing about going through UA's CS program for me was, actually, the difference in maturity level. I saw on your twitter that you wanted to recant that, because you thought it was insulting. But go on a mental journey with me.

    I started as a Junior and only minored in CS, so the kids in the classes I took were generally 2 years younger than me. Wow, it was obvious. I was in a class where the teacher couldn't get a word out because everyone was talking over him. It was like being in kindergarten. The classroom environment was so chaotic and stressful I began ritualistically smoking before it. In my experience, that behavior is not atypical. And whether or not these kids grow up, those early classes are the ones women will take and decide whether they want to pursue the degree. Not sure I would've stuck with it if I hadn't been a Junior needing to complete a minor.

    Furthermore, many people who do start in CS so young seem to be very narrowly focused on it. So much so that they dismiss everything else as being unimportant (namely, activities like 'reading books' or 'learning another language'). I find these types of people to be highly abrasive, absolutist, closed-minded, and immature. Maybe I just got stuck with a bad batch. But frankly, I wouldn't blame anyone of any gender for not wanting to stick that out for four years, much less pursuing a career where there might be more of the same. That shit is exhausting.

    1. I feel like this is a very important issue for diversifying CS, but one that I'm not completely sure how to talk about. I feel that it's easy to come off as insulting, like in the situation from my tweet. But I'm really glad people are talking about it. For instance, I think this video does an amazing job of breaking down typical developer culture and how it could really hurt developers in the future: The video clearly comes from a place of affection, so I hope its message gets across.

      The fact that people are defining the problem is also really important I think. When I started trying to talk to people about this issue, I couldn't really even put it into words, and I think most of the guys I talked to didn't know what I was getting at. I remember telling a prof that I thought part of the reason why there aren't as many women in CS was a bit of a cultural issue and a "geek factor". But he was confused what I meant. "Do you mean students use too much jargon?" "Sort of, but that's not exactly it..."

      I told him about how I had taught my biochemistry classmate (and myself) basic Python during a bioinformatics class. She was so excited by it that she told me if she could go back, she would take several computer science classes (this was a brilliant, top science student). But then one day, we stayed late in the lab and a bunch of guys from the object-oriented design class came in and started talking about their game design projects in a very typical CS student way. She was so alienated by it that she turned to me and said "I don't want to study computer science anymore". My heart broke, and I knew this was a serious problem.

      My hope is that CS profs can make a little more effort to make social rules clear in the classroom, increase professionalism, and counteract single-mindedness by promoting projects from diverse fields in class. If the expectation from professors is clear, it might make a big impact.