In the first episode of the newly named Diversity Sauce podcast, Kabir and Amy talk to Dr. Maria Klawe, President of Harvey Mudd College, on the secrets to attracting more women and people of color to computer science.
Kabir: We wanted to start with a question I have that we were hoping to get your perspective on. If you look at 25 or 30 years ago, the number of women graduating with computer science/computer engineering degrees as a percentage of the entire population was higher than it is now. And I wanted to open the discussion with why do you think that happened?
Dr Klawe: So, first of all, I think that’s true for computer science degrees. I’m not sure it’s true for computer engineering. But I’ll say why it changed for computer science. Initially, I don’t think people knew what a computer science degree was. If you look back to the 70s or even the 60s, well, what did you know about computer science? Well, first of all, people have to type. We had these decks of cards for coding. And accuracy of typing was important. And so, I think just the fact that at that time, girls learned to type in high school and boys didn’t meant that there was this sort of advantage that young women had in terms of thinking, “Well that’s something I could certainly do.” And then beyond that, it’s not really clear that anyone understood what computer science was going to be. So there wasn’t the same kind of perception of what a computer scientist would be like.
And then, I’d say starting in the 80s, we started to see that, first of all, personal computers were making their way into homes and schools, and if you think about what children did with personal computers, they played games on them primarily. If you think about the kinds of games that were available, it tends to be games that just because of the limited computational power of the processors and computers at that point, essentially the games you could do were things like either shooting things in the sky or playing ping pong or tracking Pacman through a maze. And it turned out that most of those games were much more attractive to boys than to girls. And so very quickly, computers became a boy thing and not a girl thing. And that was perceived by children, by parents, by teachers.
And then the other thing is that if you think about who the dominant kind of person you thought about being a whiz at computer science at that time, it tended to be, you know, there was Dilbert, there was Bill Gates. It tended to be people who were not seen as very social. And also the boys in schools, teenage boys, who really got into computers tended to be people who were not very social. And so very quickly there became this sort of expectations that people who are good at computers don’t like interacting with other human beings. And, just very quickly girls just felt like, “Well, that’s not me.”
Kabir: Yeah, that makes sense. As someone who grew up in the 80s as computers were moving into the home, the games, sort of the boy-centric piece of it seeps in. I think that also reflects how we talk about kids’ media. How you have to see yourself. And you were talking about Dilbert and Bill Gates. If girls didn’t see themselves in that role, they wouldn’t actively want to be computer scientists. That makes sense.
Amy: You saw this then as a problem with enrollment. Like you weren’t just getting the numbers of young women enrolling in your program, right? Because of that bias that already existed.
Dr Klawe: That was happening all across the United States and all across Canada, so the first work that I did about attracting more young women to computer science was actually at the University of British Columbia. That was actually in the mid 90s. The University of British Columbia and Carnegie Mellon University were the first two universities to actually address this problem and to think about it.
Amy: That’s great. So can you talk about some of the tactics that you used to engage enrollment?
Dr Klawe: So it turns out (this is of course true in general but false in specific instances) that young women generally are much more interested in learning something for its usefulness in solving problems in the world. And so, for example, if the way you teach your introductory computer science course is all about the intrinsics of programming language and algorithms and computer logic and those kinds of things, and if you focus just very much on one of the technical issues, it's going to be much less appealing to young women than if you teach that same material but in all of the assignments, you justify it by showing how you would use certain algorithms of techniques or perspectives to solve different problems. So for example, if you're learning recursion which is a well-known element of computer science, you could be solving an abstract problem using recursion or you could be writing a program to show how you would detect the spread of disease. And the second approach is going to be much more engaging and interesting to young women. So part of it how you teach. Part of it is also eliminating-- taking away a very common way that students experience an introductory course. So often when you start taking a course in computer science, there's a very broad range of prior experience among the people in the class. There are a few people, usually young men, who have been programming since some tender age like 12 or 5. And then there are a fair number of people who have never programmed before. And most of the young women will be in the category of never programmed before. And when they're sitting there and there are a few students who seem to know as much as the instructor and use up a lot of the air time, it's very common for young women to say, "I don't belong here. This is for people who know a lot like those guys." And so the second thing you really need to do is make sure that you segregate people according to prior knowledge and so that everybody is having an inclusive, welcoming experience that is not intimidating.
Kabir: I see. That makes a lot of sense. You mentioned some of the things that you've done to make it more welcoming to young women. What has been the reaction then from young men who were taking/interested in computer science? Have you guys looked at that it seen a change?
Dr Klawe: So one of the things that I really think is quite wonderful is that in all the examples that I know of places that have changed the curriculum, their pedagogical approach and so on to make it more interesting to young women, it also turns out that there are a lot of young men whom the prior approach didn't reach either. And so our experience has been that every time we work on this, we realize that we enable a lot of young men who were not feeling engaged and welcomed and making them feel much more interested and included.
Kabir: That makes sense.
Amy: There's also the situation where if you're the only girl in a room or the only girl in your robotics class, there's a lot of pressure there. But you've also talked about when Harvey Mudd tipped to 55% women that that was a unique kind of pressure to women too. Can you talk about that?
Dr Klawe: Yes, so the very first time... when I arrived here, we were a little bit more than 30% female in our student body. I've always been focused on increasing the diversity in all kinds of different areas of STEM. And so I started, and Harvey Mudd is a science and engineering college, so I started to gather with the community, thinking about how we might attract more women. The fourth year I was here, our incoming class was 51.5% female. And it wasn't that we had admitted a higher percentage of women than we had in the past, it’s that we had we had been improving our yield. So when I initially arrived, we were much less successful in attracting female students when they were admitted than male students. We had put a lot of effort into this. And so we ended up for the very first time ever with a class that was more than 50% female. And to the dismay of a number of people but certainly of myself, there was a lot of talking going around that they could have only happened because we lowered the standards for admitting young women. So people were literally saying to the incoming students, "You only got in because you're a girl. They just wanted girls." And of course this meant that we had a lot of first-year women that did not do well in courses that first semester because we know from the work of people like Carol Dweck and Claude Steele that if you tell people that they're not going to do well because of gender, then indeed they will not do well. We're thrilled to know that we actually did graduate 93% of the students who started that year and our 6-year graduation rate, but a lot of young women struggled in their first two years.
So after that we decided for quite a while to stay at about 48% female. Now this is not admission to the computer science program because we don't admit by discipline, we admit just to the school, I suspect we will come in at over 50% this year just looking at the early decision numbers where we have about 10 more women than men admitted in early decision. So we may do it again. I'm relatively confident that we're past where we were in 2010 that we can admit 51% female students and they'll do just fine.
Kabir: That’s really fascinating because this discussion about quality... at times you'll hear the story about how we want a really diverse pool but it's got to be quality, as if somehow creating a diverse pool reduces quality. It's like diversity and quality and have become these choices that you have to make. But I wanted to ask you Dr Klawe, if someone comes up to you and says, "Well why does this matter? Why is your focus so much on the entire student class, or specifically around getting more women in computer science? Why is this so important?"
Dr Klawe: When I answer that question, I look at three different things that are going to be improved by getting more diversity into the tech workforce. So the first one I'll just talk about is supply versus demand. There's lots of studies that show that around the world and specifically in North America the number of jobs looking for computer science skills is roughly twice the number of computer science graduates that we produce. And so it's very unlikely that we'll meet that demand unless we do better in attracting women and people of color who are both very underrepresented in the tech workforce and in computer science programs. Unless we fix that problem, we'll never meet the demand. So that's a big economic impact for countries.
The second reason is these are actually great careers. They are highly creative. They pay very well. They offer lots of flexibility. I was going to say, "If I had a son or daughter"–I have a son and a daughter and I'm very happy to say that our son is doing computer science and math. He's doing information security for Square, for their mobile team. I tried the hardest I could to get my daughter interested but her position is, "Look. Both my parents do this. My older brother does this. I'm going to do something that is more important." So she's doing peace negotiations in Africa, so that's pretty impressive. But in any case, they're great jobs. I just think it would be terrible if we didn't give opportunities to women and to people of color by making sure that they could get the education in a way that appealed to them.
And then the final one is of course the most important which is every single issue that faces the world today, whether it's healthcare or peace negotiations or entertainment or education, they all are going to involve computer science in some aspect of the solution. It won't only be computer science, but we know data science, machine learning, all kinds of things related to software development are important to every single part of society today. And we also know that if we have teams that are working on issues where the teams bring diversity of experience and perspective you get better solutions. So why wouldn't we want diverse teams working on things? The demand, the excitement of the career opportunities, but most importantly, getting the solutions to the problems that face the world.
Kabir: Right, right. The research shows over and over again, the more diverse means better outcomes. Thanks, that was fantastic. Amy, you wanted to talk about the pipeline, right? The importance of both sides of the pipeline, high school and mentorship, right?
Amy: Yeah, a little bit. And also about how we think about bringing more people of color into STEM careers. One of the things you talk about are the courses available in high school and what kids have access to before they even come to you. Can you talk about what people, even pre-college, can be doing to work on the very earliest part of this pipeline?
Dr Klawe: So there are a number of efforts trying to get more computer science teachers into high school and middle school, trying to get more computer science concepts into everything from pre-K through high school. One of the best ones that's known is code.org and one of the things they do is on a certain day of December each year is they have an hour of code where they get millions of people literally to spend one hour doing code. But on their website, they also have lots and lots of various kinds of computer science curriculum at different levels. And there's also a number of initiatives retraining teachers to teach computer science.
It's one of the problems about having so many job opportunities for CS grads is that very few of them go and become middle school or high school teachers. And so there's a lot of work going on in working with people who are already math teachers or physics teachers or something and helping them learn how to teach computer science and then providing a lot of support for them as they do their first couple of years of teaching. However having said all that, which I think is great and is very important, it's still completely true that we can take a student who has never written a line of code before they get to Harvey Mudd and they will be just as talented a computer science graduate and will be sought after by Google and Microsoft, Facebook, Intel, Apple as much as somebody who started coding when they were 12. So it's an error to say that this isn't something you can start when you first start college. And even more importantly, there are a lot of examples of people who are starting to either at age 39, at age 60. My younger sister is 62 and she's in a web development computer science at the Northern Alberta Institute of Technology in Edmonton and absolutely loving it. She's suffers from fibromyalgia. She's disabled. She walks with a walker and all of the other students in her class are 18-24 and she is doing just fine. The thing I really try to emphasize is this is not something you have to learn while you're young. This is something, and in fact, I had never taken a computer science class or written a line of code before I started my PhD In computer science back in 1978 at the University of Toronto and was hired as a computer science professor (I already had a PhD In mathematics that I had gotten in 1977), and I was hired as a faculty member assistant professor tenure track at the University of Toronto 10 months after I started my PhD. So there’s this mythology out there that say unless you learned it young, you're not going to succeed. That's wrong and that's sad. I do think it's a wonderful thing for young people whether it's kindergarten or middle school or high school for people to have exposure to, it's just that it's never too late.
Kabir: No, that makes a lot of sense I think there is a myth that you need to start computer science at a very young age and that's the only way to succeed. But clearly it's an important skill that if you do learn it when you're younger, that's great, but you're never too old to learn it. Harvey Mudd is a place where you sort of focused on bringing in more women into computer science. I wanted to talk a little but about BRAID. This is something you started with a number of universities that are focused on bringing women into computer science. Can you talk a little bit about what you've done over the past three years and what your goals are?
Dr Klawe: One of my desires for Harvey Mudd is that we can be essentially an innovation lab for undergraduate science and education and that when we discover how to get more women to succeed in computer science and engineering and physics, areas where women are underrepresented all across the country, that we should absolutely be sharing that information. And in fact, there's a really excellent center called The National Center for Women in IT (NCWIT) that has a lot of information very similar to the kinds of things that we’ve done at Harvey Mudd on best practices to attract more women and more people of color to computer science. Those resources have been available for a decade. The kinds of things we’ve done at Harvey Mudd and a number of other places have done, they’re not particularly expensive and they’re not particularly difficult. So one of the questions is, why are so few of the departments making the effort and seeing a transition? And I think part of the reason is that computer science departments just over the last 5 years have been under an intense about of pressure because of the increased interest in majoring in computer science. And also, a number of faculty are being recruited away from computer science departments into industry at much higher salaries. So there’s just a lot of stress going on in CS departments.
I came up with a theory that based on studying several departments that had managed to attract many more women and looking at the fact that in almost all cases, the person leading the effort was a person in a position of leadership, either department chair or former department chair or dean, something like that. And so the idea of the BRAID Project was to recruit a small number of department chairs who wanted to take on a commitment to really working on the things we’ve seen be successful at other departments and that in exchange we would attempt to raise some funds to support their initiatives. So this started almost three years ago in the summer that I was giving a preliminary lecture at a conference called Snowbird for department chairs. There were about 250 department chairs in the audience and I was talking about how six departments had made a lot of progress. University of British Columbia was one of them. Carnegie Mellon, Harvey Mudd, Cal Poly Slo, MIT, and the University of Washington was the group. And the thing I was talking about is that they all did similar things and they weren’t particularly difficult. I could see that the department chairs were really listening and were really engaged. So on the spur of the moment, I said that for the first ten department chairs that signed up/sent me an email that said they wanted to do this, I would go out and fundraise to get them some resources so that they could, for instance, take their students to the Grace Hopper Celebration of Women in Computing or the Tapia Conference.
By the time I finished my talk, 15 department chairs had signed up and so I went out to fundraise from Google, Microsoft, Facebook, and then eventually Intel and that started the project. Interestingly, one of the people I talked to at Microsoft, Brad Smith who is president of Microsoft said, “You know, it would be really helpful if you could get a research team to study what’s happening in these departments right from the beginning.” And so I reached out to Jane Margolis who has done a lot of research in this area and she suggested another professor at UCLA, Dr Linda Sax and Linda said yes. And so we have had right from the beginning of this project a research team from UCLA studying what’s going on in the departments. And they now have a granted from the National Science Foundation for about 2 million dollars to do a longitudinal study of what is happening in the classroom departments at BRAID and how it is affecting the students. And it turns out that there has never been a study like this done. And it’s incredibly exciting to see the program evolve. We are now—our funding was originally for three years. Funding from the four companies was in place for three years. We’re now reaching out and trying to extend it for another three years since obviously, it typically takes the students four or more years to graduate from college. We think we need at least six years to see whether we’re really having impact. I think we’re likely to see that some departments are more successful than other departments just because it’s a very diverse set of universities. It’s from quite small places to huge places like Arizona State University. But generally, we’re very encouraged by the progress that has been made.
Amy: That’s really exciting, especially as we think about providing outlets for people in the professional world to hire with inclusion and diversity in mind. It seems like a place that they could start is with all these BRAID institutions. So that’s very promising. Are you sort of bridging that between hiring companies – how do organizations approach Harvey Mudd about job possibilities for your graduates?
Dr Klawe: Well, we are currently pretty much at the top of the list for most tech companies right now partly because we have great students, partly because we have a very diverse student body now. We’re not only roughly 50/50 male/female, we have a significant number of both African American and Hispanic students. That makes us very attractive to any company who wants to increase the diversity of their tech workforce. But the other reason they’re very interested in our students is we have a big emphasis on communication skills, collaboration, and creativity, people skills in addition to highly rigorous and broad curriculum. So I think our attention to diversity as well as attention to what some would call the softer skills and I think I would just call the 21st century skills to be successful in a tech career, that’s made us very attractive.
I’ve also spent a fair amount of time talking with HR leaders and other executives in tech companies. And I was just on the phone this morning with Ellyn Shook who was the chief HR officer and leadership person at Accenture. Based on taking some of the things that we had done at Mudd and then translating them into the equivalent kinds of things at Accenture, they came in with hiring over 40% female in their tech workforce last year. And they're also looking at a number of things like pay equity, paths that they can aim at women that will put them in significant technical leadership positions. All of those same kinds of things that we've been doing at Harvey Mudd. And it's just very exciting to see a company take a look at what worked here, really taking a close look at their promotion retention practices and look and what they need to change to make those careers equally attractive to women as they are men. I have no doubt that all kinds of companies could make their work environments more attractive to women and be much more supportive of promoting women into leadership positions. Unfortunately, so far, it's a relatively small set of companies that are truly being thorough and rigorous at looking at all elements of everything from recruiting, where you recruit, how you interview, how you negotiate, whether you provide mentoring or not. All of those kinds of things. But I'm optimistic that as a few companies start to make that progress, others will follow.
Amy: I mean especially if you can expect the outcomes will be really good for the business bottom line as is shown time and time again. So, fingers crossed.
Kabir: So, Dr Klawe, we wanted to thank you for coming on. We really appreciate your time. We have so much more we could discuss, but we'll have to bring you back to talk about the wonderful things you're doing, and of course, the continuous great outcomes that you have at Harvey Mudd.
Dr Klawe: Well, thank you so much. It's been a pleasure talking with both of you.