Increasing Girls' Interest in Computer Science | Sapna Cheryan | Talks at Google

So today I'm going to be talking to you about stereotypes as gatekeepers and how to increase the representation of women in computer science. So the past few decades have witnessed an increasing emphasis on diversity. For example, we've seen our standards of beauty change. We've seen a surging population of immigrants in our own neighborhoods. And if we're lucky, we might for the first time see of the first American non white male president. And most relevant to my research, we've seen women enter domains that were previously completely closed off to them. To illustrate this last point, let me show you a graph. This is from the National Science Foundation. It's the percentage of women receiving bachelor's degrees in biology, math and chemistry. And what you can see from this graph is that women have made significant progress over the past few decades. They're now getting almost half of the degrees in the math and chemistry and over half the degrees in biology. Now, this isn't only true at the undergraduate level. We can see this at the faculty level too. This is a screenshot of the faculty in biology at Stanford. And what you can see from this is that women are represented pretty well here. When I counted, they made up about 40% of the faculty in biology at Stanford. Now, of course, I wouldn't be here today if women were complaining. Fully integrated into all domains of society, there are still places where women are significantly underrepresented. Let me show you an example of that. So this is the faculty of computer science at Stanford. This is the computer science department. And what you can see from this is that you have to scroll down almost halfway in the Alphabet to get to the first female name. So those are the two female names on this part of the list. And when I counted, women were making up less than 15 of the faculty at Stanford in computer science. Now, this isn't only true at the faculty level. Women are also underrepresented in computer science at the undergraduate level. So if we go back to this graph that I showed you to start with, and we add computer science, look what happens. So women are still significantly underrepresented in computer science. They make up only 28% of the degrees received in computer science. And what you can also see is that the pattern is different for computer science than it is for these other degrees because the percentage actually seems to be falling as opposed to increasing. And so it seems like women are underrepresented in computer science. And before I talk about why this is the case and what we can do about it, I just want to address a question which is why this is a problem. Why do we want all women in computer science in the first place? And there's two reasons I'm going to talk about. The first thing is that that women are missing out on potentially lucrative and flexible careers by not pursuing computer science degrees. And so I'm sure, as you all know, computer science professions are rated among the best professions in the United States year after year. And women are missing out on these careers that are probably good for them. The second reason that we want women in computer science is because computer science is missing out on female perspectives. And this is important because computer scientists are in many ways the modern architects of society. And to just illustrate this, I'll give you an example from the 1970s and 80s. So some of you might remember the airbag fiasco. This happened when the first airbags were implemented in the cars. And the people who designed the first airbags were men. So it was an all male design team and they designed them for men. So they were for someone who was taller and bigger. And what they found when they put them into cars is that they actually injured, in some cases kill women and children because they weren't designed for them. And what Margolis and Fisher have argued in their book is that if these design teams had had women on them, that perhaps these tragedies could have been avoided. So now we know we want women in computer science. The question is how do we get there? And solving this problem is really a matter of solving two problems. The first thing is that we have to recruit more women into computer science. We have to get women interested in field and then once they're in the field, we have to get them to stay. So it's the problem of recruiting them and then once they're there, also retaining them. Now most of my work is focused on the recruiting side of the question how do we get women interested in computer science. But I've also done some stuff on retention and I'm going to share that with you guys today because one of the studies I've done actually references Google, so I want to make sure to get to that. But first we'll start with the recruiting question. So how do we get more women into computer science? Well, one thing we know from the literature is that girls are less interested in computer science than boys at all levels of schooling. So this is interesting and it might make you ask the question, what is it about women that makes them less interested in computer science? And this is an interesting question, but it's not the question I asked. And the reason it's not. The question I asked is because the answer to this question often involves somehow pathologizing women by saying that women aren't smart enough or they aren't confident enough or something like that. And what I was interested in asking was a different question, which is what is it about the field of computer science that makes it less appealing to women? And what we'll see from the research that I'm going to present is that when we ask this question, we get a different answer and it leads to a different set of solutions. So what do we know about the field of computer science? How are computer scientists professional? And I probably don't even have to tell you guys this, but when people think about computer scientists, they often think about images like these. So it's kind of a geeky field. And to get at this, we asked Stanford undergraduates. We had 188 of them just describe computer science majors to us and I'll read you some of their answers. So one of them said, nerdy techie, stay up late coding and drinking energy drinks. No social life, pale, sometimes socially frustrated. Inquisitive skills, build focus. And then my favorite, they're usually guys, very intense, very intelligent, intuitive and quick. They don't frequently take showers. So what we noticed is that the stereotype of computer science majors is pretty strong. We got over half of the people mentioned that computer scientists were somehow geeky or nerdy or socially awkward or really intensely focused on computers. And you'll notice that we didn't ask them to generate stereotypes, we just asked them to describe computer science majors. Yet they were very willing to give us these really highly stereotypical images of what they thought computer science majors were like. And we did a follow up study to this because we were interested to see whether computer science not only was seen as very geeky, but also was seen as very homogenous, meaning that the people were seen as being very similar to each other. And so we asked students, Stanford students again to tell us how similar they think people are in a series of different majors. And what we found is as we predicted that people rated computer science majors and electrical engineering majors as significantly more similar to each other than they did for these other majors. The math majors, biology majors, econ majors, English majors, both of science majors, history majors, and psychology majors. And so what this led us to conclude was that not only is computer science right Hickey, but it's also seen as very homogenous. Everybody kind of embodies that stereotype. So to just summarize these perceptions that we began looking at you can think about it like this. So this is kind of the image that people have of computer science majors. They're guys, they're geeky. And this is the image of a group of computer science majors. They're all pretty similar to each other. In contrast, people rated other majors as being different. They're kind of less geeky, and they're also less modulus. So they're different from each other. Okay, so given that these are the perceptions of computer science, what I was interested in is whether this image of computer science is one of the reasons that women aren't interested in it. And if that's the case, more importantly, can we change the image of computer science from this to that? And if we do that, will women suddenly express more interest in computer science? And so that was the goal of the set of studies I'm going to present to you today. So what I did for the first study to see whether this image makes a difference is that we had females and males, Introduction to Psychology, fill out a questionnaire for us. And we told them to imagine they were college seniors and that they were about to start their first job and they needed to decide between two companies. And we told them that the two companies were identical in terms of what job they were going to have, what salary they were going to make, how many hours they were going to work, and how many men were in those companies. The only thing that differed between the two companies is what they noticed when they went on their job interviews. So what they noticed was in the environments of the company. And we told them that in the first company they noticed these things. They noticed Star Trek posters, soda cans, video games, comics, junk food and technical magazines. And these were all things that we had pre tested with another group of participants. We'd ask them to tell us what kinds of things they would find in the office of someone who's really geeky. And these are the ones that were most commonly generated. So we took those and then we told them that in the second company they would see things like art posters, water bottles, nature photos, and general interest magazines. And we asked them, based on this information, to indicate what company they think they would choose and how similar they feel to the people in those companies. So let me show you the results. So this is the percentage of the people in the study who chose each company on the Y axis. And then we have the results for the company with all the Star Trek and the video games stuff. And then the company that had the art poster and the nature photos and what you can see Here is that the majority of people who wanted to be in the geeky company said that that would be the company they chose were men. In contrast, more of the people who chose the nine cookie company were women. And if we look at these results a little differently, you can see they're very striking because men pretty much didn't have a strong preference between the two companies. They were pretty willing to say that they could be either in a geeky company or non geeky company. But if you look at the women's results, the purple bars, you see that women were significantly more likely to choose the non geeky company than they were to choose the geeky company. Now, we also asked them how similar they were to people in the company. And these are the results for similarity. So what this means is that the sex was correlated with how similar people felt like they were to the people in the geeky company, such that men said that they felt more similar to the people in the geeky company and perceived similarity was strongly related to interest in the geeky company. So the more that you felt similar to the people in that company, the more interest you expressed in their company. And what's most interesting about this is that that relationship that we saw in the previous graph between, between sex and interest in a geeky company, that relationship actually goes away when you control for similarity in a geeky company. And so what that suggests is that the reason that women aren't interested in the geeky company is because they don't feel like they're similar to the people in that company. So let me just summarize the findings from this study. What we see is that women were less likely to choose a stereotypically geeky company than men. And if we control for similarity using statistically controlling methods that explain the relationship between sex and interest in a geeky company. And what's most interesting about this study to me is that people were willing to draw these conclusions about whether they belonged in these companies and who was even in these companies, not based on the employees in them, because we didn't give them any information about the people in the companies. We just told them what stuff was lying around. But they drew conclusions about the people that were in those companies and how similar they felt to those people and whether they thought they would actually belong there. And so it really speaks to the way that environments can really communicate alignedness. So what does this mean for Google? Well, I think one thing it means is that if you're interested in recruiting a certain type of person or a specific person to A company, you might think about what messages you send that person about who else is like them in their company. So it could be who you choose to interview them or even what objects you have lying around. Those things will probably make a difference in how comfortable that person feels in that company and that environment and could potentially determine what choices they make. So what we want to do for our next study is you'll notice that in this study we didn't actually mention computer science. We just had two hypothetical companies. But I'm most interested in looking at whether we can change the representation of women in computer science. And so we wanted to bring it back into the domain of computer science as a major. And the other thing we wanted to do is rather than giving them a questionnaire about it, we actually wanted people to engage with these environments. So we wanted to put them in these situations where they could see these objects and see if that would make a difference into what they chose. So what we did in this study is we brought males and females Introduction to Psychology, and we actually decorated a room and the Gates Computer Science Building at Stanford. And so what we did is we just went out and purchased all those objects that we had had in our questionnaire in the first study. And I'll show you pictures of what the room looked like. So this is the room, the geeky room. So what we did is we put up the Star Trek poster, we had our science fiction books, and we had our Coke hands, like, stacked up like that. And then in our other room, we had the Necro poster and put it up and put some, like, phone books and some water bottles. So it's the same room. We just simply changed what was in it. And we had half of our participants sit in the room with the geeky stuff in it, and half of our participants sit in the other room. And then we simply just asked them to complete a questionnaire that assessed their attitudes towards computer science. So it was the same questions that we asked in the other study, but in this case it related specifically to how interested they were majoring in computer science. So let me share the results with you. So again, we have how much they considered majoring in computer science on the Y axis. And then we got the men in blue and the women in purple. And these are the results for just the people who came into the geeky room and then filled out that questionnaire. And what you can see is that when they're amongst all these cues, women are much less likely to say that they consider majoring in computer science than men. And what was striking for me in the study is that this is extremely low. So this is a one to seven scale. And the mean for women was almost nearly the lowest they could possibly be. And this also replicates the pattern that we see in the world, which is that women are just significantly less interested in computer science than men. But look what happens when we change the room. So when we change the room, women were suddenly more interested than they were in the geeky room. And even more surprisingly, they were as interested as men were in majoring computer science when there's a presence of these other cues. So for our next study, what we wanted to do is we wanted to see if this was something that we could get more broadly. And specifically, one thing that we were worried about is, well, are women really reacting to a stereotype of computer science majors, or are they reacting to something specific? Like maybe women just really don't like Star Trek, or maybe they don't want to be around drunks or something like that. And so what we wanted to do was instead of using objects, we wanted to use another way that people commonly get information about social groups, and that is through the media. So in this study, what we did is we varied how computer science is portrayed by using media representations. So we had females and males on Stanford campus and we went up to them and we had them read one of two articles about computer science majors. And these are articles that we fabricated. We made them up and we made them look like they were USA Today articles that were produced, printed off the web. And so these are what the articles look like. So the one on the left says, study finds computer science continues to be dominated by geeks. This is a geeky article, and it says that computer science is still very geeky. The last line of the article says, computer science equals geek forever. And then the article on the right says, study finds that computer science is no longer dominated by geeks. And this article ended with computer science equals E to no launcher. And so what we had our participants do is we had them read either this one or this one. So half read this one and half read that one. And then we asked them the same questions that we had asked in the previous study when they were seeing the rooms. Specifically, how much have they considered majoring in computer science? Okay, and here are the results. So here we have considered major computer science. Again, on the Y axis we've got the geeky article and the non geeky article. We've got our men and blue and our women in purple. And what you can see is, again this looks a lot like the results I just showed you. After reading the geeky article, they were significantly less likely to say that they had considered majoring in computer science. It's just something that just wasn't interesting to them. But after reading the non geiki article, they were more interested than they were after reading the geeky article and equally interested as men. It looks here like they're more interested, but statistically this actually wasn't significant. So they were equally interested as men in reading the non geiki article and more so than after they read the geeky article. And so let me just tell you about a study that I want to do because I think this will be relevant for you guys. So this is what I'd like to do now based on these studies. So all these studies were done in what we call them the labs. That means we bring people into Stanford campus, we create these environments, and we then measure how those environments affect their attitudes of the behavior. But I actually want to go into real world environments next and study how those environments affect people's attitudes towards them. And so what I want to do is just go to 20 tech companies and actually just take pictures of the environments. And specifically I'm interested in the public spaces. So the spaces that people see when they come visit their friends or they come interview. So places like the lobby and the cafeteria. And I just took a couple of pictures actually I found a picture of Google on quilt, so I included that. And this is the picture of the Adobe lobby. And what I want to do is just remove all the people from the photos because in this study what I'm more interested in is how the environments affect people's choices. So remove all the people and then just show these photos to undergraduates and ask them to rate how much they think they would like to be in those companies based on those photos. And what I'm predicting is that the photos that are fake here in some way will have less interest by women based on my. And what's neat about the study is that you can actually go back through these photos and figure out which ones women were less interested in and then figure out what it is about those environments that make women less interested. So you can code them for the presence of certain cues or presence of certain moods or colors or something like that and figure out what kind of cues make women feel like they belong in those environments more. One more study that I'm looking to do next is things thinking about another type of representation. So what I've been presenting to you so far Is the geeky representation, the Star Trek stuff and the soda cans and the newspaper article that says study fines Computer science continues to be dominated by geeks. And I've shown you the opposite one, the non geeky one, which is more of a neutral like nature artsy one. But there's a third one that I want to test and that is a more feminine one. And so you can imagine that you would bring women into rooms, there'd be like boona bars and yoga. So these are all off the top of my head, I haven't pretended them yet. But yoga class ads and chocolate. And then there could be an article that says study finds that computer science soon to be dominated by women. And I think what would happen here is that women would perhaps express even more interest in being in computer science if it's represented as feminine than they would even maybe as a non geeking condition and probably more as a magikian condition. And I think what would happen with men is we've been seeing that men haven't really been affected by these two. So they've been kind of, you know, they express the same interest in computer science regardless of whether they're reading the geeky article or the non geeky article. And sometimes people ask me what's going on with men, are they just like asleep or they're not paying attention? And I don't think that's the case. I think if we gave them this that we might see an effect here, that men might suddenly distance themselves from computer science. Okay, so to summarize these recruiting studies, I've shown you what we can see is that the way a group is portrayed really impacts who expresses interest in that group. And specifically what we saw is when computer science was portrayed as being really geeky, women expressed less interest in being in that major. And importantly, what we also saw in these studies is that information about these groups is transmitted in seemingly very innocent ways, like the objects in the environment or through media representations. But these things can really affect who expresses interest in the group. And so one thing that might be important is to think about what we put in the environment or to try to change media representations to make them more inclusive and welcoming to more people. And what we saw in these studies is not only maybe an explanation for why women are underrepresented, but also a potential solution. One solution is to change the settings, to change these environments, to change media representations. And this can actually draw more women into the field than are currently there. Okay, so I presented the recruiting side of it and I'm Just going to go move on to the retention side of it now. And there's a lot of previous research by social scientists on the experience of being female in a male dominated domain. And I just want to go over some of the more interesting findings from the field. So one thing we know is there's still discrimination, it's still there. There's lots of studies on this. I'm just going to talk about one. So a bunch of economists sent resumes out to companies and the resumes were identical. The only thing they changed was the name of the person who was on the resume. They changed it from a female name to a male name. So it was like John to Joan. And, and what they found is when they just made that simple name change, those resumes were much less likely to be called for an interview. And this is especially true when the jobs were higher status jobs. So people are still reading into female names and females in general that there's some kind of lower ability or some reason that they don't want them in these domains necessarily. Another way that discrimination can also occur is it can also be more subtle. And there's a group of social psychologists that do a study on how male and female leaders receive feedback from people. And so what they did is they trained men and women to lead a mixed sex group and they trained them to do everything identically. So they said the same things, they used the same script, they had the same mannerisms and they led these groups and they videotaped the groups while they were being led by either this male leader or this female leader. And what they found when they coded these videos is that the people who were responding to the female leader, they expressed less favorable non verbals. So they nodded less, they frowned more, smiled less, and the more she talked, the more they did this, the more they kind of gave her disfavorable non verbals. And in contrast, when the man was leading the group and he was using the same script and saying the same things, he just of kind got more nodding, more smiles. And so you can see how this would create a situation where the woman feels like perhaps she doesn't belong there, perhaps she's not doing a good enough job. And this could lead into a feedback cycle where she might feel like that's not her domain. So another way that being female in a male dominated domain plays out is through the expectations that people have for men and women. And there's a study, this is actually one the of favorite studies. This is done by Frank Flynn at the Stanford Business School and He found that women in businesses. So what he did is he went into businesses and he asked people to record their interpersonal interactions. And what he found is that women are asked to do favors more often than men. And so they're basically burdened with either having to say no more often or having to do more work. And what's even more frustrating is not only are they asked to do, but when they do it, they get less credit than a man does the exact same thing. So women are basically giving more and receiving less for doing the same stuff. And finally, in social psychology, there's this theory of stereotype threat. So this is the idea that people have different expectations for the abilities of men and women and that women are stereotyped as having lower abilities in domains like math. And when these abilities are made salient to them, they underperform. And I'll just, I'll just go over the study. So stereotype threat is defined as the fear of confirming a negative stereotype about your group. And when that stereotype about your group is made salient, it can affect your performance in the domain. And this is research by Claude Steele and he's done many studies to show this in many different groups. And I'll just talk about one of them. What they did in this study is they had male and female students with college ethics and they had them come into the lab and they gave them a 30 minute GRE math subject test. And so this was the test. This isn't the test that everybody takes to take the gre. This is the test you take if you're going to go into math as a grad student. So it's a very hard test. And they gave it to undergrads and they divided the test into two 15 minute halves and they told the participants different things about the two halves of the test. So what they told them is they said one half of the test, in this half of the test it has been shown to have gender differences. And then they told them so then they took that half of the test and then they moved on to the second half of the test. And before that half of the test they told them this half of the test has not been shown to have gender differences. And of course they're making this up. Okay, they just divided the test arbitrarily in half to tell the participants that half has had gender differences and half hasn't. And then they take the test and let me show you what happens. Okay, so this is the results for the question the half of the test where they said this test has gender differences. And what you see here is that women underperform, so they don't do as well as men. And this is also the pattern that we see typically on math, on very difficult math tests. But when they said this half of the test, this is the test that has no gender differences in it, that alone was enough to eliminate the performance difference between men and women. And so what they've shown is that what's going on in this case is that women are so concerned with proving the stereotype right, that they're more anxious, they try harder, and they actually, it actually takes away from their performance because they've got all these other thoughts going on. They aren't able to focus on this very hard math test. Okay, so that was just a review of some of what I find more interesting studies in social science about what it's like to be female in the male dominated, dominated domain. And I'm just going to add what some research that I've been doing in this area, and that is the idea that women just don't look like the typical person in that domain. Because everybody expects that if you're an engineer, you're a computer scientist, you're male, and if you're not male, it's salient to everybody. You just don't look like that person, like someone who would be in that domain. And so what I'm wondering is what is happening to women who have these incompatible identities, who have these computer science engineering identities, but also have a feminine inside to them? How do they negotiate their femininity when they're in these domains? And there's some anecdotal evidence that one thing that might happen is that women might distance from being feminine when they're in these domains. And this is a quote from Kathleen Morwitz. She's the director of NYU's current Institute of Mathematical Sciences. And she was asked whether she thinks about her children when she's at work. And she says, no, I'm much more likely to worry about a theme when I'm with my children. And so what she's doing here is kind of pushing away the idea that she's a mother and you know, a mother before she's a mathematician and trying to prove to other people that actually she's a good mathematician. And so one thing I was interested in is will women, when they're in the, in the engineering or math or computer science domain, do they somehow distance from their femininity? And so the way that I got at this is that we had female engineers, computer scientists, At Stanford, come to the engineering quad. We ran them in Packard and Terman, if you guys are familiar with Stanford. And we gave them a questionnaire that reminded them about being engineers. So we asked them to fill out a bunch of questions, questions like, I am a worthy member of the engineering computer science community. And so we did that for half of the women in the study. And so for the other half of the women, they didn't fill out this questionnaire about being an engineer scientists. And so what we were thinking is that in this half, people really made the salient fact that they were engineers and computer scientists. And then what we wanted to see was when women were thinking about being engineers and computer scientists, would they kind of push away from their feminine side? So we did this, we gave them this test, it's called an implicit association test. It was developed by social psychologists and it's a way to measure, uses reaction time to measure how much someone associates to concepts, how closely they associate two concepts. And what we wanted to see is whether people would associate themselves as feminine less strongly when they were thinking about being engineers. So we gave them the self feminine implicit association test. So like I just said, what we thought is that when women were primed with engineering, so that means after they filled out the engineering questionnaire, they would show less of a self feminine association. And indeed this is what we found. So for the women who filled out the engineering questionnaire, they distanced from their femininity more. So they were less likely to say that they were less likely to associate themselves with feminine concepts. And the feminine concepts we used in the study were things like manicure makeup, things that were pretested to be neutral kind of feminine words. So they weren't extremely positive or extremely negative words. What was happening here is that when we were thinking about being engineers, they were kind of pushing away from those words. So to summarize this study, what you can see is that making salient the computer science engineering identity causes women to kind of distance from their femininity. And this kind of attests to the burden of having incompatible identities, of being female in a domain where you're expected to be male. Now one thing that I'm wondering that kind of goes along with the first half of the research is, well, does computer science have to be male? Does it have to be represented like that? And if it's not, is that going to kind of ease the burden of having to negotiate the two identities? And of course computer science doesn't have to be represented as male, it just is. And in Some countries, computer science is actually considered one of the more feminine disciplines because you're not actually using strong machinery and things like that. And so what we wanted to do was go to a place where computer science wasn't so massive and then see whether women were still doing this distant thing from their femininity. And we didn't go to Eastern Europe or India, which would have been fun. But what we did do is we found a place in America. It's actually in San Diego. So this is the great Hopper celebration of women in computing, where any of you guys were. Okay, a couple of you want. Okay. So you can fill in if you want to, if you want to add some stuff about it. But the Grace Hopper celebrates a movie computer. This is a mission statement. It's to inspire, educate, encourage, and create awareness of opportunities for women in the field of computing and to celebrate the considerable achievement of women in the field. So it was overtly a place where you could be a computer scientist and be a woman and still be seen as a computer scientist. So I just wanted to show you that this is not only the mission statement, it was all throughout the conference. These are some of the ads that were in the program of the conference conference. And what you can see is that. So this is hp, this is IBM. They're both using strong female representation in their ads. So they have a bunch of diverse emails. The Cisco ad had women with different types of heels, which I thought was kind of cool because it shows not only women, but different types of women. And Yahoo used words to make the puende. And we think creating the next incredible app is women's work. And Google is also a major sponsor of this. So I wanted to show you guys the Google Ad look like. Interestingly, Google didn't reference gender in their ad. It's just Google's searching for years of great aspirations. But I don't think that the message was that Google didn't value women, especially because I brought some of the things that Google gave out at the conference. And I'm sure you guys have seen these, but the Google Woman shirt is there, and I got a Google Woman notebook, and I'll show you some of the things that the other companies were giving. So this is a Microsoft shirt. And Microsoft also gave out this makeup bag, which I was kind of surprised about, which actually has a mirror and a lipstick case in, says Microsoft Grace Hopper on it. So, anyway, so the point is that this was definitely a place where women could feel comfortable being both engineers and computer scientists and being women. And I'M just going to show you a couple more pictures. These are pictures that we took. So these are women who are actually at the conference and who participated in the study that we did at the conference, which I'll tell you more about in a second. This is them. So every company that sponsored the conference threw like a huge party on the last night. And so this is at the Microsoft party and then that's them at the HP party. They had like a Shrek theme going on and then this picture. So this is Chris as my collaborator and we just thought we had to do a little reenactment of what happened to him at the conference because he went to go use the bathroom at one point and he walked up to the door and they had put a sign for female over the male bathroom. So they changed, they turned the male bathroom into a female bathroom. And Chris came back and he's like, I think I know what it feels like to be the target of institutional prejudice now. So yeah, he had to go a little further to use the bathroom. And this is just a picture of what we did there. So what we did is we set up a table and we had subjects, we paid them for their time. It was like a 10 minute study. And we just had people come and do the task. And what they did is they did the exact same thing that we did at Stanford. We didn't change anything about it. So all that changed was the context. We took the same questionnaire, we took the same type of students. So we had female engineers and computer scientists, we had 44 of them. We gave them that engineering questionnaire to half of them that timed their engineering identity. And then we gave them that same implicit association test to see how much they associated feminine and self based on reaction time. And what we were wondering is, well, is this going to change things? Are women still going to feel like they need to distance from their femininity when they're in a place where being feminine seems to be okay? And lo and behold, we actually found that when they were primed with their engineering identity, when they filled out that engineering questionnaire, women actually were higher on their self feminine association in this context than they were when they didn't fill out that questionnaire. And so one thing that we want to do now is it could be a possibility that there's something about these women in this conference that would lead them to generate these results. Most of the Stanford students. And so what we're interested in doing is creating these environments in one place in the same population and seeing whether we can actually manipulate the environment and get people to distance or not distance from their femininity based on what's in those environments. So we're going to try to make a little mini Grace Hopper somewhere on the Stanford campus, and we'll see how that goes. Okay, so just to summarize these half of the studies, what we've seen is that distancing from an incompatible identity seems to be a strategy that women use when there's computer science at an easy salient. But importantly, we see that that's not how it has to be. That if we can change the context, we can change whether we need to dissonance our femininity when there are new feelings. Okay, so I just want to conclude now, so I'm going to give you guys time for questions and I just want to tell you what I think are the main takeaways of this talk. First of all, we see that groups can be represented in different ways. So we have these images of social groups, but this isn't necessarily the way they had to be or the way they have to be in the future. We can control how those groups are represented. And the reason this matters is because when we control how those groups are represented, they have real consequences for people in society. And what I've shown are some of the negative consequences of representing. Computer science is extremely masculine. Women tend to avoid the group, and they also are from this position where they have the burden of having to negotiate being female in a domain that's a masculine domain. And what I think is important about my work is that it not only looks for why the reason is that women are enrolled in the field, but it really points to the solutions that we can use to try to get more women into the field. And sometimes I get asked if what I'm suggesting is that we go through companies and we kind of degeek the environment. We take out all the Star Trek stuff, we take out all the soda. And I don't think that the key is actually subtracting. I think what's more important is that we add. So I think what's important is that we add things. We add cues that make people feel like they belong if they're welcome in those environments. And so, you know, sure, take people to see Star wars, but maybe also offer a yoga class or have a barbecue or give something that people can really attach their identities to, something that really makes them feel like they want to be in those environments. So hopefully this work speaks to the importance of constructing environments that are inclusive, good to everyone. Thank you.

So I was recently talking to Carol Freeze.

I don't know if you've heard of her work at all. She's done some stuff. Okay.

I was originally talking to Carol, please, by the way. I thought you handled that pretty well. Anyway, so she was. She's working at CMU pertly, and she's done a lot of stuff with women in computer science. And so one of the points that she's made is that basically emphasizing femininity, it sort of seems sort of counter.

To some of the stuff you come up with.

Emphasizing femininity in the classroom and trying to sort of think about making coursework better for women, I suppose is causing.

Part of the problem.

So once that's doing, what that's doing is serving to further emphasize the differences between men and women and further emphasize a stereotype that exists that may not be completely true. Like sort of, for example, with the Kiki or Star Trek stuff. I mean, there are plenty of men and women in computer science who might like Star Trek and there are plenty who don't. Right. There are plenty of men who don't and women who do. And so when we say things like.

Oh, we need to get rid of.

Star Trek because, you know, this is alienating women, then you're further engendering the stereotype that men like sardines and women don't.

Right, that's exactly. That's a great point. And actually when I. So I made this a little bit of a simple presentation because I didn't want to like overwhelm people and I didn't want to talk for like three hours. But when I actually write these studies up, I do. I'm very aware of the fact that, you know, we can't homogenize women as all like hating Star Trek and all like, you know, feeling like they want to get manicures and they can't do it or something like that, you know, and they should be valid point. And when I talk about it, I usually say that there's a lot of men who are also probably being turned away from computer science because they don't really feel like they are affected to those domains either. And if we just allow computer science to become more broad, so welcome different types of people, that will get more women and we'll also get more men who aren't that type of person. So, yeah, I think that's a great point. I'm not trying to say that all women are like this or should be like this and that that works actually indicator for this thing. But. But what I'm trying to say is that computer science has this image of being very narrow and that image is having an impact on who is brought into those domains. Yes, I had a question about. I don't know if anyone's done a study about how willing women are to change the environment because like I said, I mentioned grad next who's coming into kind of generic. Start going above it, whatever. Cool. You know, whether it's a game, touch rate or something else and women doing the same thing. And that perhaps this goes along with assisting yourself with stereotypes. If you have like four office mates that are guys, you may not feel as comfortable, you know, bringing in a picture, but you know, flowers or something, you just might think, oh yeah, I'm one of the guys and not do that. Yeah, that's a great idea actually. Like to go in environments and actually take pictures of like women's spaces and see if they differ depending on who they work around. That would be really cool. I haven't, I haven't. Or how likely women are to want to feel they can change it right out. Being perceived as too feminine and not technical enough, not competent because they're so into. Yeah, that's a great idea. Yeah. But I also think it would be cool to see what women are wearing in these environments. I was going to say the exact same thing because designing your environment is one thing, but the main thing you have to design every day is what you're going to wear. So if everyone's wearing a T shirt that was free yesterday and you feel like wearing a skirt, you know. Right, exactly. Yeah. And I'm sure people have done that sort of. Well, no, actually, no, they haven't. No. I mean this idea of like, you know, people oftentimes, especially in social psychology, they'll look at one identity but they haven't done too much work on how you negotiate when you have like two identities. So that's kind of like the up and coming type of work right now. So. But yeah, that's a great idea. Quick question.

So I guess one of the high level objectives or main takeaways is that hey, maybe women were missing out on their different fields. Right. Technical fields being one of them. But curious if you guys have done. Or if you've read about her parallel studies in other male dominated fields such as business or law or whatever else. And how polarizing are those when you're sort of presented with an idea that reinforces the male dominant?

Yeah, that's a good question. I mean, I haven't done any. The only Thing I've done is comparing computer science to these other fields and showing that the stereotyping computer science is just a lot more strong, which is why I chose to focus on computer science. But a lot of, I mean a lot of the work on retention has been done in math and I can't think of anything. And the business studies will go and look at how women are stereotyped as being bad in negotiations and how that influences their negotiations, their actual negotiations, when that stereotype is made salient to them. But I don't think anybody's gone into like what are the stereotypes of lawyers and how does that affect who's in it? And one thing I think is interesting though is that there are like when you think of a doctor you usually think of a male. That's kind of common. But now women are making up more than 50% of people in med school. So somehow those fields have done something different than computer science has done because they've come from being such a male dominated field to changing to being more osint. So it'll be a really good thing to would compare those fields more directly.

First off, I think that the presentation was great. You had some really good statistics. I have two small comments and a small suggestion relating to them. So I think you touched on a little bit how this happens at all levels of education that females may be a little distant from computer science. So one thing that for me is not obvious and I think you'll probably agree and it was kind of mentioned before that it's not obvious to me that Star Trek is inherently more male than female than feminine and that water feels more than females than males.

So.

I think that perhaps looking at, and I'm sure there's studies that do this, like why females at a young age are exposed to more of these things and don't like Star Trek would be valuable. And also on that point of maybe touching a little bit earlier, one of the statistics I was looking at was where you've been working some of the philosophy 101 students or sorry, sorry, psychology.

Oh yeah.

And I think that even that if we could bring it a little step, another step earlier in the student's life would be a big step to the presentation. Just because I think maybe a lot of those students have already made that decision away from computer science and already said I'm going to pursue psychology. So if you get it a little bit earlier when it's more of a fresh mind and see why are you not liking Star Trek and that stuff.

Yeah, that's a really important point and the reason that I use college students is because it's easy for us to go into high schools is actually a pretty big deal because when you run subjects you have to get human subjects approval. And getting human subjects approval to do anything on Myers is crazy, crazy process which could take all your years in grad school even yet. And so I chose to focus on college students and I think it's actually pretty amazing that we are getting these advice at Stanford because, you know, all these women obviously probably did pretty well in these fields, you know, but I think it would be really important to go and see what kind of images that they're seeing and. Yeah, exactly why. And I guess I'll have. I just have a one question about your first. I don't actually know why women don't like Star Trek more than men either. But you know, I do know that computer science, it used to be. So the first computer programmers were women. I don't know if you guys know that, but they were actually women, they were hired by the army. And when men started realizing that it was actually not, it was more than just typing, it was actually something that could be high status. They came into the field and they, the base took it over and salaries went up, the usual story. And what happened. Actually, you know what, I'm going to show you guys a graph because I think it also might address the point that you guys question you guys might have had. So this is the pattern of women in computer science, the percentage women getting bachelors. And one thing that I was interested in, and this is just kind of, you know, I'm not a cultural anthropologist or anything, so this is just cursory, look at this. But I was interested in what happened here. Why did women suddenly start tanking here? And I started thinking about the stereotype because that's what my research concerns. And I just did some searches and I found that some of the movies that we consider like the geekiest things were actually released around this time. And this is also the time that corresponds to the whole rise of Bill Gates and Steve Jobs and the idea of this, you know, really singular geek, you know, male geek who does things, does things alone. And you know, these movies kind of correspond to that idea. And you can see here that this also comes, you know, at the same time that women are decreasing. And so I mean it'd be cool to study this in more depth and see if that's really one of the reasons that this happened. But yeah. Did you have another thing? I think there's a Lot to thank you.

So I appreciate you tackling this problem. I think this is a really tough problem A lot of people dealt with with respect to something like this, that representative computer sciences over time. There are these stats out of UCLA that you probably see where we look at men versus women. They're interested. And around the mid and early 80s there was a surge in computer science, both men and women. And then Most recently around 2003, those numbers have come down dramatically. But for women they've come down even more. And so I'm wondering if what you feel is different about that time versus this time.

Right.

Because if it was just the nerdy images around that time, then we should have seen that big gap happening then we didn't actually see it until 2003. And so I'm just wondering, do you conjecture what is the difference between now those.

Well, I mean I think one big thing that happened is the dot com bust, you know, and I think that probably has a lot to do with these statistics. And so I don't.

That affected men and women.

Right, Right.

But the disparity between men and women actually grew.

Yeah.

And so the economics of it alone doesn't necessarily explain gender disparity.

Right. Unless men are somehow. Or women are less concerned with it or something. But yeah, I mean, I don't know. I've seen those statistics. I mean Natural Science foundation doesn't go out that far. So I don't. But I feel that it's become even a bigger problem and a lot of Christianized farms are really motivated to solve the problem. But I haven't. I guess maybe I've been. I haven't. It would be nice to have to read some sociology or read something about like what's been going on recently. And I haven't seen anything like that. And so it's hard for me to comment. My research is all very, you know, at the individual level, but I think it would be great to do some kind of cultural analysis of what's been happening and seeing if there's anything that's. That I don't know if you guys have any ideas of anything that you've seen in the last few years. Yeah.

So just a comment on the. Just a comment on the 2003 drop. So if you look at that, you're.

Born in 1984, when all this stuff is.

It takes until about 2003 for you to get it.

That's true. Right. So. Thanks so much for your presentation. I have two questions.

One is just like methodology in the process. So in terms of the pre testing.

And everything like that.

What steps do you take when you.

Form, like a questionnaire? And also, you know, the final questionnaire.

That you present to your participants, like, what sort of back, like the process right before that, the preliminary process, like.

What are the steps you take in.

Terms of pre testing and what are you looking at to see which one. Which types of questions are the best questions to ask and which types of sort of labels and cues are the best ones? So, yeah, that was my first question.

And then I had another question. Okay, yeah, I can answer that one first. That's a great question. I mean, what we do is a lot of times we develop the questionnaire and then based on previous literature or based on our earlier findings, based on our intuition, and then it goes through usually many revisions based on pilot data. So I think most people will pilot their questionnaires on a small sample of people and just look to see if, first of all, people can answer the question if they make sense to people, if they're worded. Okay. And then also, you know, sometimes you just want to. You just want to make sure that the statistics are looking okay. So you'll do some initial statistical analyses and sometimes you check your questionnaire because it didn't make any sense or your idea was totally off base, you know, and then you'll just try another version. And with pre testing, things like, for example, the newspaper study. So the two articles that I presented, you know, that had some specific things we wrote. Yeah. So we, we wanted to make sure that the two articles were both equally believable. So we didn't want it to be biased by one, like, people saying, oh, yeah, computer science is still really geeky. There's no way. It's not really not geeky. So they just didn't believe the article. Then it wouldn't have worked. So that was really important. So they believe it. With the objects in the room, we pretested to make sure people noticed the objects. So, you know, pull them out of the room afterwards. And we were just like, write down what you saw, what you saw in the room. So it kind of depends on the stuff, but that kind of gives you an example of things that we're looking at. Okay, that helps.

And in terms of just you talking at the end about, like, solutions. And I know that I was just wondering what kind of conversations you think are possible between, like a company like Google and some of the research that you're doing, you know, that like our diversity recruiting team is doing, you know, a lot to work in you know, K through 12 as well as in college, university diversity, ethnic diversity, gender diversity, all that stuff. But do they actually have conversations with folks like you or people who do research, sort of bridge that gap? Because I know that they are working really hard, but I'm just wondering if they're getting some of their ideas from some of this research.

Yeah, that's a good question. I mean, I guess I'm not the best person to answer that because I think anybody. Well, maybe Hazel would be a good person to answer that. So, professor, in our department at Stanford, so I don't know, do you have any ideas about whether the companies are moving to us or. No, they don't. They don't. Yeah. I mean, we, yeah, it's really important at Stanford to get our research out there because, you know, we're not just interested in getting published. You know, we actually care about what's going on in the world and we actually want to make positive changes. And also it helps our research to be informed. That was high on the real world because, you know, sometimes academics are kind of out of punch and you know, you want, you want to. That's why it's so great that I'm here, like from my perspective, because I feel like it's just going to help my research to hear what kind of questions you guys have just to talk to you afterwards. But we, I mean, we're totally open. You know, we love interfacing with people who aren't Stanford people and people who have ideas about and engage with these issues every day. So call me, call me or email me if you ever. And then actually so, you know, I am looking to actually come into companies to do studies. So that's another way that interfacing can happen. Because, you know, one thing that I think would be cool to do is to actually get a list of what good companies, companies that have high percentage of emails, what they look like, and then make those results kind of widely, you know, widely distributed to companies that don't have females and you know, maybe they can make it just changes and then we can come back in and see if that's helping or something like that. That would be great. Any other questions? Okay, well, I just want to thank all you guys for coming and thanks again to the Luminit Google team and my collaborators. If you have any questions, I'm here, so feel free to come afterwards. Thank you.