SHELLEY CORRELL: It’s great to be here with you today. NCWIT is an organization that’s objectives and missions very much resonate with the objectives of my own research as I think you’ll see very quickly as we dive into that. What I want to talk with you about today is the way that gender stereotypes introduce biases into the workplace that can be disadvantaging to women, so I’ll spend the first maybe 15 or 20 minutes of my talk laying out how these biases come about, and then I want to spend the remainder of the time thinking about ways that we might actually be able to reduce them. I have a lot of research I want to share with you so I want to jump right in, and I want to talk to you about a study that when it came out received a lot of attention in the press, so this might be a study that you’ve heard something about. In this study the context of it is the hiring of musicians into orchestras. In the ’70s and ’80s major orchestras all around the country began making a seemingly minor change to the way they auditioned musicians for hire, and the reason for this is at the time in the ’70s and ’80s women made up only 5% of all the musicians in orchestras around the country, so it was a very male-dominated space. The orchestras were worried that perhaps a bias was creeping into the evaluation process, and was reducing the number of women that were being hired. So what they decided to do was to introduce a screen in front of the musician so that now when the musician auditioned for a position he or she was not visible to the judges, and, therefore, the gender was unknown. This small change had a dramatic effect. The odds that a woman moved on past the first round of auditions increased by 50%, so there’s a 50% increase in the movement past that first round of auditions. Today 25% of orchestra musicians are women so we’re not to parity yet but that’s a big change from 5%. The very sophisticated analysis that was done by the people I cite on the slide shows that this screen really was a major contributing factor to the increase of women. I like this study because it illustrates really two of the main points that I want to try to drive home today. The first is that gender stereotypes bias the evaluations of individuals in ways that we might refer to as male advantaging, and what I mean by male advantaging is that men’s evaluations come out to be more positive then they would be if we did not know men’s genders, and women’s evaluations come out more negatively than they would if we did not know their gender. So this is something what we call a male advantaging bias. So in the case of orchestras switching to this blind audition removed this male advantaging bias. Once you couldn’t see the woman her evaluations went up. This leads to my second point, and that is these biases can be reduced or eliminated, they’re not inevitable. So after explaining how they emerge I’m going to talk about ways to reduce or eliminate them. Now in the case of orchestras the way they reduced them was to put a screen up in front of the person auditioning for hire. Clearly, that’s not a solution for everyone. We could not or would not want to do all of our work life behind a screen, so we’re gonna have to be a bit more creative than that but there are tools and strategies that we can employ. Now before moving on to some of the other studies I want to just say a bit about the use of the word bias. A lot of people don’t like this word, it sounds really negative. It has this kind of ugly connotation like we’re pointing our finger at someone, and accusing them of being sexist, but that’s not at all how I’m using the word bias today, and it’s not how social psychologists use it in general. Instead, when we talk about bias we’re simply talking about an error in decision-making. So if we think about the orchestras again we can probably assume that most of the judges evaluating the musicians were well-intentioned people who truly wanted to hire the best musician out there, but yet gender affected how they saw the musician, and how they rated the musician’s quality. For reasons I’m soon going to explain all of us, men, women, whites, people of color, young people, old people, are all prone to these sorts of biases, but the good news is with effort, and proper procedures in place these biases can be reduced. So this is a no blame message, but a high responsibility one. So why are these biases so ubiquitous, and what can be done about it? If we want to understand how to reduce the biases I think it’s important to first understand a little bit about why stereotypes lead to bias in the first place. So one of the main things that we know is that gender stereotypes function as what we might call cognitive shortcuts in information processing. So to return to our orchestra example, again, imagine that the orchestra is auditioning like 100 musicians, and they only have one or two slots that they’re hiring for. Processing all the information associated with 100 different musicians would be very cognitively demanding. That’s a lot of information coming in, especially, in the time usually allotted for such decisions. So usually we have weeks perhaps to make hiring decisions, not months or years. In these information heavy context, especially, under time pressure, individuals understandably look for shortcuts, some way to help them navigate through that sea of information. Unfortunately, and this is where gender stereotypes come in, gender stereotypes function as a shortcut. What I mean is implicitly or unconsciously we often rely on what we know about categories of people such as men or women when evaluating individual people, individual men and women. So what do we know about men and women? What’s the knowledge that we have about men and women? Here the knowledge that I’m talking about is the knowledge that’s encoded in widely shared stereotypes that are available in our society. Stereotypes are simply that, they’re sort of widely held beliefs. We may not personally endorse them, but widely held beliefs out there in our culture about what men and women are like, and these then are what’s available to lead to the kind of errors in decision-making that I’ll draw out. Psychologists have differentiated between two types of stereotypes. On the one hand we have what are called descriptive stereotypes. Descriptive stereotypes are beliefs about the traits and capabilities that men and women are thought to possess. They’re how we think men and women are. Research on the content of descriptive stereotypes finds lots of things, but one consistent finding is that we still find in modern day American society that most people hold the belief that men are more capable and competent at a whole array of tasks than women. So whether or not there’s any truth value to this or nor these beliefs persist. Now often these differences are small. People think men are a little better at something than women, but they can be more pronounced, and they are more pronounced for the so-called male type tasks, and these are things that I think people in this room care a lot about. Nora’s comments had us thinking about things like spatial ability, mechanical reasoning and the like. For these type tasks the beliefs are actually considerably stronger with people thinking that men are considerably better than women. Women, by contrast, are stereotyped at being better at female type tasks such as those requiring nurturing ability and caretaking, so if we were in a place where we were evaluating people’s abilities in those domains we would likely see a female advantaging bias. However, the thing to know is that nurturing and caretaking themselves are often devalued in the workplace, so those are descriptive stereotypes. Prescriptive stereotypes are widely held beliefs about how men and women should be, so they have that prescriptive should element to them. The content of these beliefs are that women are expected to be, they should be nice, women should be warm, they should be concerned about others, and not dominant or assertive where men are expected to be agentic, by that we mean sort of action oriented, get things done, assertive and certainly not modest or weak. As we will see later with some of the studies I want to show you, when people violate these prescriptive stereotypes they tend to be disliked. So what I want to do now is to show you what the consequences of these stereotypes are for men and women in the workplace. How do they play out? One of the findings from my research, and the research of others is that descriptive stereotypes affect the standard that we use to judge the performances of individuals. So it turns out that our standards shift depending upon whether we’re evaluating a man or a woman. Why would this be? Well, if we think about it when a woman performs well at a task, and especially a male type task, so let’s say we find out she has a lot of spatial ability, for example, this runs counter to our stereotypical expectations about women as a group. As a consequence we tend to more critically scrutinize her performance. It wasn’t what we expected so it says we’re trying to look at it and make sense. How did this come about? When a man performs equally well at the task this by contrast is consistent with what stereotypes would let us to expect, therefore, we don’t scrutinize his performance as closely. What this means then is that we end up in many contexts judging men by a more lenient standard than women, and because we’re judging men by a more lenient standard, again, this is out of awareness, the performances of men come to be seen as higher quality. Women’s performances, by contrast, being judged by a harsher standard come to be seen as lower quality, so I think this research again can help us understand what was going on with the orchestra study. Women musicians were being judged when their gender was visible they were being judged by a harsher quality so their musical performance was not being seen as sufficiently high to warrant them a position in the orchestra. This is a graphic that I like to use to illustrate this. Here we have a figure trying to go over a high jump bar, and it says if the bar is a little bit higher for women than it is for men meaning that for any given woman she’s going to have to jump a little higher to get over this bar than a corresponding man would, or in the aggregate if we want to think about it, what this means is that on the other side of the bar we can expect to see more men than we see women. Now in the past, of course, these differences were larger than they are today. Stereotypes 100 years ago were way more stark in differentiating men from women, but even then, of course, exceptional women got over the bar, Marie Curie. We could name other sort of exceptional women that have always cleared the bar, but in my work what I’m really interested in is thinking of ways to bring these standards closer together so that the average woman is rated in the same way that the average man is. I want to now turn to our second research study that shows us how stereotypes affect judgments, and how differential standards play out in that. This study comes from the field of psychology, and what the authors of this study did is they conducted an experiment where they created a resume for a job candidate who was ostensibly applying for a position to become an assistant professor in psychology. This applicant had really good credentials, and he or she had just completed his PhD. So they created this resume, and then what they did is they sent the resume out to psychology faculty all over the country, and they asked them to make judgments about this assistant professor on all kinds of domains. Importantly, they asked people to say how hireable this person would be in their own psychology department for a tenure track position. So these psychology faculty get these resumes, and make these judgments. A randomly chosen half of the participants got the resume on the left with a man’s name on it, Brian Miller, while the other half of the psychology professors out there got the very same resume with the exact same publications, the exact same grants, teaching evaluations. Everything was the same except the resume had a women’s name on it, Karen Miller. So now what the question, of course, is does gender bias these job evaluations? People have the exact same resume. The only thing differs is whether one is presented as a man or one is presented as a woman. The differences were striking. 79% of the people who got the resume with Brian’s name on it said that he would be worthy of hire for a tenure track position in their department while the other randomly selected half of the participants who got the very same resume, but with a women’s name on it only 49% of those deemed her worthy of hire. A second phase of the study really points to this kind of idea that women are being judged by a harsher standard. So in a second phase of the study what they did was analyze the comments that people wrote on the evaluation forms, and what they found is that the participants wrote four times more doubt raising statements on the women’s evaluation forms than on the men’s. So they wrote things like this first quote, I would need to see evidence that she’d gotten these grants, and publications on her own. So notice they’re scrutinizing the information that’s contained on her resume. The second one, it would be impossible to make such a judgment without teaching evaluations. I mean, teaching evaluations are a perfectly valid criteria if you’re evaluating a professor, but that this was being raised only on the women’s evaluation forms I think is quite telling. So to summarize to this point what we see is that gender stereotypes affect the standards that we use to judge men and women and, therefore, can affect the extent to which we come to see them as being sort of higher or lower quality. This is, of course, unfair to women who have to perform at higher levels than men to be seen as equally qualified, but I also want to point out that a lot of the things we’re talking about today are also going to be unproductive for organizations who presumably would prefer to make more accurate judgments about the people they’re hiring, and not being influenced by gender stereotypes. In addition to stereotypes influencing the standard that people use to judge individuals gender stereotypes also affect the criteria that we use to judge individuals. A cool study that shows this is another experiment, but this time what’s happening is people are rating applicants for a police chief position. As you may know, police chief is a very male type job. In this study participants always rated two applicants, and one of the two always had more education, and the other one had more street smarts, or more experience we might say, so that was the key difference between them. They tried to create people who were sort of equally qualified, but one was better on one dimension, and one was better on another. In the first phase of the experiment a set of participants evaluated these two applicants with no names on the file, so there’s no gender that’s salient here at all, and what they found is that participants overwhelming preferred the applicant that had more education, and they justified their choice by saying this person has more education, and I think that’s more important for being a police chief. So education seemed to be the criteria that really mattered to the participants. In the second phase of the study they used the very same resumes, but now what they do is put men and women’s names on them so now one of them is a man, one’s a woman, and importantly the one that has more education is the man, and the one that has more experience is the woman. So now they have a different set of participants evaluate these two applicants. What they find here is that people overwhelmingly preferred Brian, the male applicant, and the reason they gave is that he had more education. So no problem here, education was what mattered in the first condition when we didn’t know people’s gender. Now with a different set of participants we know the gender people are still saying that education is what matters. What’s interesting is what happened in the next condition, and you can probably guess this by now. So now what they did, a different set of participants, same resumes, they just switched the names around. So now Karen has more education, and Brian has more experience. So if education matters Karen should be chosen, but that’s not what happened, or I wouldn’t be standing here today. What happened is that Brian was actually chosen, and when people were asked to justify their choice they said it was because he had more experience. So notice how the criteria shifted to justify what was probably their gut feeling about who was more appropriate for a job as a police chief, the man in this case. So to summarize to this point we see that women, even if highly qualified, can find their past performances discounted due to shifting standards, and their suitability for a job discounted by shifting criteria. Now what we’ve seen across all the examples that I’ve told you, the orchestra musicians, the psychology professors, and the police chief, is that stereotypes affect hiring decisions. So you might be wondering, yeah, but does this matter once we get on the job? Once you’re on the job you’re able to sort of demonstrate your competence to people. Do stereotypes continue to produce this male advantaging bias? The answer here is clearly yes. Research consistently finds that women have less influence in group settings at work. So what I mean by that is in a group setting if women offer a contribution to the group people are less likely to use that information when deciding how to proceed. Women tend to have their contributions judged less positively in the workplace. Very similar kind of experiments that show this. Women are less likely to get credit for their ideas. This third point when I talk to women’s groups this is the one that seems to really resonate with women at work. I don’t know if you’ve ever had this experience, but it’s an experience where you make a suggestion in a group and no one hears you. It’s just kind of like you said nothing, and a few minutes later someone else in the group, perhaps a man, makes the very same suggestion, and they get credit for it. So whenever I talk about this as I said it usually resonates with people, and one of the women in a group I spoke to recently sent me a card with this on it, which sort of humorously depicts the very phenomenon that we’re talking about here. So cartoons are always welcome. It’s not just cartoons research bears this out. So I’ll tell you about a study here. This is a classic social influence kind of study that was done back in 1995. The authors of this study created a four person group with two men and two women, and what the group was charged with doing is to deliberate and come up with a group solution about who should be awarded the custody of a child. So what happened is each of the participants was given a packet of information to read about the case, and they came and deliberated and came up with the answer, who’s gonna get custody of this child. The key feature of the study, though, is that one person in the group was given a unique piece of information that no one else in the group had, and it turns out that this unique piece of information was actually really key to deciding the case. Half the participants were in groups where the person who had the key piece of information was a woman, and the other half were in groups where the person with the key piece of information was a man, so the question is who do you listen to? What they found is that participants were twice as likely to use the key information when it was introduced by a man. This is a very common finding in these studies. Now why would this be? Well, generally, we’re more likely to be influenced by people that we view as being more competent, so if gender stereotypes lead us to sort of unconsciously expect that men are going to be more competent or capable then that means that their ideas are probably going to more easily sound good, right or convincing to us. Just one more current study in this domain that I want to mention to you because I think it’s really important a very similar kind of design, but what the authors of this study did is in addition to one person having more information than the other people in the group they actually had one of the people in the group labeled as the expert, so you think you would listen to the person that is labeled the expert in your group. What they found is that participants were more likely to listen to men experts than to women experts, and this had consequences. What it means is for the groups that had men experts they actually performed better than the groups that had women experts, but it wasn’t because the experts were any better, or knew anything more it was simply that people were more likely to listen to the expert when it was a man. What I like about this example is I think it really shows that the negative effect that stereotypes have not only on women but on workplaces. Presumably a workplace would want other members, employees in the firm, or what have you to be listening to the experts that they, in fact, hire. Okay, so these are some of the kinds of ways that stereotypes can introduce bias. Sometimes, when I talk about this work people want to know are these biases big enough that we should be concerned about them, or are they trivial, I mean, does it matter that someone doesn’t give you credit for your ideas? Are we making mountains out of molehills here? But what I want to argue is that mountains are made out of molehills, and that the effects of gender stereotypes while small in any one instance, actually, accumulate to create increasing disadvantages for women. I want to show you a study that sort of shows how this can happen. This study is a computer simulation study where what the authors did is they created a hypothetical firm that you see represented here, and it’s got that traditional kind of pyramid structure. At the top of the pyramid you see 10 people at the top. These are our leaders. These are the people who’ve moved up to the top. At the bottom in the lower level entry positions you see what we have are 500 employees, 250 men and 250 women, so this firm does not have a pipeline problem, and we’d love to be in this position in computing, right? So 250 men and 250 women in the firm. The way you move up in this firm is if you have better performance evaluations than other people, so if you perform better you move up it’s a meritocracy. It’s the only thing that matters in this computer simulation. In the first condition of this experiment what they do is they give everyone of these hypothetical men and women a performance evaluation. Some have high evaluations and some have low evaluations, but the key thing is they fix it such that the mean performance evaluation is the same for men and women, and the distribution is the same for men and women, so while some men might perform better than some women, or vice versa there is no pattern to the difference, okay? There’s no systematic difference. Not surprisingly given this what they find is when they run the simulation over and over again on average the top positions are 50% women, okay? They call this the no bias condition with equal performance information, and decisions based solely on performance 50% of women at the top. The next condition what they do is they introduce a 1% negative bias onto women’s performance evaluations. In other words, they take that performance evaluation, and they discount it by 1%. That’s a trivial amount of bias I think most people would agree, very, very small amount of bias. The question is does it cumulate over the eight levels of this organization as people are being evaluated? The answer here is yes. With only a 1% bias we go from having 50% women at the top to only having 35% of women at the top. Make the bias a little bigger 5%, but still very small in terms of what we find in a lot of the studies that I’ve reported, and what we see now is less than three and 10 of the positions at top are now held by women, so what this I think shows is that even a small amount of bias if it happens day in and day out as people are being evaluated can cumulate, and negatively affect women’s careers. Now, so far I’ve talked about sort of bias in general, and before turning to what we can do to make it better I want to actually talk about when the processes that I’ve described are worse, so knowing when things are worse helps us I think, sometimes, understand the origins of the bias. So, basically, what we want to try to understand is why is it that under certain situations people rely even more heavily on stereotypes as a shortcut? Well, research has shown that these biases are worse when tasks are male-typed. When there are tasks for which we have pretty strong stereotypic beliefs that men are better than women these processes are more pronounced, so things like technical competence, spatial ability, things like that, those are the tasks that are being considered they’re worse. It’s worse when positions are higher in an organization, so the higher up you move the worse these biases are, and that’s because our stereotypes about leaders, and our stereotypes about managers if you say to people what’s the typical leader like, for example, those stereotypes overlap almost perfectly with our stereotypes about men, so leaders and men stereotypes are very similar, but the stereotypes about leaders, and the stereotypes about women hardly overlap at all, so I’ll show you some examples of that here shortly. My own research has shown that these biases are worse when women are mothers. The stereotypes that we have about mothers turn out to be stronger than our stereotypes we have about women in general. In particular there’s the stereotype that mothers are not as committed to paid work is pretty strong, and has really pretty profound consequences. It’s worse when evaluation criteria are vague. When people don’t know how to evaluate someone they’re just told hire someone, and they don’t know what to do that’s when they more need this sort of cognitive shortcut, and they tend to rely on stereotypes more heavily. Finally, they’re worse when people making decisions are tired, rushed, or otherwise what we might think of as cognitively burdened, when we’re distracted thinking about something else, we’re tired in the kind of ways we almost always are at work, and we’re really looking for a shortcut, stereotypes are readily available. So that’s that the bad news. Let’s spend the last bit of our time talking about what can be done to reduce these things. Here what I want to do in my comments is differentiate between survival skills and organizational solutions. Both are gonna be useful in some way, but I want to differentiate between them. Survival skills are the advice and training that we provide to women to help them navigate a workplace where bias is always possible. If you want to keep yourself busy for days and days on end Google career advice for women. Career advice for men is a much shorter number of hits career advice, so there’s no shortage of things that we tell women in terms of survival skills about how to better navigate their careers, and a lot of these things as I said turn out to be quite useful, so for example, be sure that others notice your own accomplishments. Toot your own horn. One of the problems with stereotypes is people they’re discounting your performance, so self-promote let people know what you’re doing. Be confident, negotiate, ask for things that you want. Speak up at meetings. Don’t wait for people to call on you. Volunteer for leadership positions. Don’t wait for people to come knocking on your door, and because of the negative effects that can children can have don’t draw attention to your children. Don’t put pictures of them up, or if your children are sick say you’re sick. I mean all of these things can be found in handbooks for women. Now as I’ve said these survival skills can often be helpful, but they’re rarely going to be enough, and they can often, also, cause a backlash reaction. That means they can cause an unintended negative effect, so let me show you a study that shows this. This study comes from psychologist Laurie Rudman’s work, and what she does in a series of experiments she has people examine how self-promoting behaviors affect the judgment of individuals. In general what we know is that men are more likely to as the saying goes toot their own horn than women are they’re more likely to self-promote, so given this we often advise women to be more self-promoting arguing that if women self-promote they will be more likely to get credit for their ideas, more likely to be hired and so on. So what Rudman does in her study is she has people evaluate job applicants who are interviewing for a job, and these job applicants as you can see from the bathroom figures here are either male or female, and some of them are either self-promoting the ones on the right with the horns, or they’re more modest, so if you are a participant in her study you would be assigned to one of these four quadrants where you would see one of these applicants who is either male or female, self-promoting or modest. Now the thing that the participants don’t know is these people are actually trained actors that are part of the study and they’re enacting a script, so whether it’s a man or a woman self-promoting they’re saying the exact same thing, so that’s not what’s going on here. So what happens what does she find? Well, what she finds is that self-promoting works for both men and women the more they self-promote the more competent they’re seen in their job interview process, so self-promotion does enhance the competence ratings of both men and women, however, women who self-promote were also seen as being less likable. People didn’t like the woman who self-promoted. They didn’t mind the man who self-promoted, but they didn’t like the woman who self-promoted. Why would this be? Well, when women self-promote, or when they negotiate too hard, or when they give directive orders in a leadership position they violate those prescriptive stereotypes that we talked about earlier those shoulds, okay? These are the stereotypes that say that women are supposed to be nice, warm, concerned for others, not assertive, not dominant and not aggressive, so women are violating those by self-promoting. Now, sometimes, when I talk about this I’ve had women say to me, well, I don’t really care if people like me at work or not I’m just trying to get my job done, but likeability matters, especially, when you’re trying to get a job. What Rudman shows in her paper is that the man who self-promoted was more likely to be hired than his modest counterpart, but for the woman that wasn’t true. The woman who self-promoted was judged to be more competent, but people didn’t want to hire her because they didn’t like her. The modest woman by contrast wasn’t very likely to be hired either because they didn’t think she was very competent, so you see the double bind it’s hard for women to be judged as simultaneously likable and competent. So what advice do we give women in this mess that we find ourselves in? So the survival skills here really involve trying to establish your competence without being threatening, without being seen as aggressive or assertive or dominant, so one piece of advice that I just love that you find over and over again is to be relentlessly positive and pleasant. If you’re relentlessly, I mean, I love that, relentlessly, if you’re relentlessly positive and pleasant than perhaps you’ll be seen as likable, but also still be seen as competent, so notice the tightrope we’re walking here. Express group-oriented motivations when making suggestions in groups. For example, you might say something like it would be best for all of us if we did X, Y and Z, rather than just saying let’s do X, Y and Z, the directive that we might expect of a leader. Use humor to coax subordinates along, and in this last strategy push forward, pull back. I’ll read you a quote to give you a sense of what I mean. This comes from Alice Eagly’s recent book on women leaders. She’s interviewing a Wall Street female executive, and this woman says you have to be strong and assertive without offending people, so you push a little and then back off. You’re always testing the waters to see how far you can go trying not to get angry, trying to think of other ways to say you’re not right without attacking the person. It’s getting more and more difficult the higher I go. So I think as we look at this list, and especially as we listen to that quote it becomes clear that we’re not solving the problem of bias. Instead we’re working around it. If we want to solve the problem of bias we’re going to have to look upstream to more the root of the problem, and we’re gonna have to come up with organizational solutions, so in my last few minutes I’ll talk about what some organizational solutions that I might recommend are, and I have six of these I want to share with you, but they’re all going to sort of share a common underlying principle, and I think whenever you’re trying to enact change it’s really important to think what’s the underlying principle that really could affect the change. Here what the principle is is we need to break the tendency of people to use stereotypes as a cognitive shortcut. So the first thing is actually teaching people about the effects of stereotypes. Now I do this a lot and people would always say to me you’re preaching to the choir, I mean, you go to a group like this this is a group that’s about wanting to help women. Of course, people here are all on the same page, we want to do this, we’re not the ones that need this message, but my pushback here is to say I’m not preaching to the choir, hopefully, I’m arming the choir, and what I mean by that is that we’re giving well-intentioned men and women the tools to be able to avoid bias themselves, and also the tools to be able to think about changes within their organizations. The second solution is to establish clear criteria for evaluations. Tons of research of all kinds now shows that when formal criteria are in place, when people know what they’re supposed to be doing when evaluating job applicants the percentage of women, and the percentage of people of color who are hired goes up. It’s a very clear finding, and you see it in that experimental study with the police chief that I mentioned earlier. In a fourth phase of this experiment what they did is before participants saw the applicants they told the researcher what criteria mattered to them, so they came into the study and they said we’re gonna have you evaluating police chief what criteria do you think is important people overwhelmingly said education was more important. Once they said that they committed themselves to that criteria, which meant that Karen now with more education was actually more likely to be hired, so what people committed to that criteria Karen when she had more education was more likely to be hired. The next closely related suggestion is to evaluate the criteria you’re using in the first place, and to be sure it is the right criteria. A lot of times our criteria came about through historical means. We look around and say whose been successful here in the past, and what are they like, and that’s where we get our criteria. If women weren’t part of the organization in the past we may have criteria that historically men were better at than women. If that criteria is not valid than perhaps we need to change it. Here’s an example that people in the room know very well I think that will illustrate that. Carnegie Mellon increased the percentage of women in its computer science department from 7% to 42% in just five years. Now, since then the number has gone down some, and that speaks to a different issue, and that is how to sustain organizational change, but, nonetheless, the gain is impressive, and it’s worth thinking about what they did to produce that gain and they did many things, but one of the things they did was to change their admissions requirement, so that they no longer required high levels of prior computer experience. As with special reasoning ability this is an ability that could be taught by teaching it in the first semester of school more women were eligible for this particular major. I mean, it didn’t affect the quality of graduates they produced out the other end. The next solution hold decision-makers accountable. This is commonly recommended, but how it works in the case of stereotypes is very interesting. If I have to explain my decision to someone else I have to go and say here’s why I preferred Brian Miller knowing that I’m gonna have to explain the decision to others that I’m gonna be accountable forces me to slow down, and more carefully scrutinize the hiring decisions that I’m doing, so it’s that more careful scrutiny that causes me to not use stereotypes than as a cognitive shortcut. Be transparent track numerical progress. Organizations manage what they measure, and I know lots of people here in the room do this. When you’re constantly posting the numbers about where you are in terms of women, and other forms of diversity what you’re doing is signaling that these are things we need to be thinking about, and in so doing this also helps break the tendency for people to use gender and other forms of stereotypes in making their judgments. Finally, legitimate women leaders. Those higher up in the organization can do a lot to ensure that women leaders are getting the appropriate amount of credit for their expertise, so a study I love that was done way back in 1984 found that female graduate students were rated more positively by undergraduate students after a faculty member vouched for their experience and expertise. This turns out to be important because as many of you may know in higher education female TA’s in science and engineering classes often are rated more negatively by the students they’re teaching. Female faculty in those fields often also experience this, so what they found is that if a professor introduced the graduate student, and didn’t just say here’s Susan she’s gonna be our TA, but said here’s Susan she’s gonna be our TA. Susan’s area of expertise is this, she’s written these papers that what they found is this actually led the female TAs to have higher course evaluations having been so endorsed. I’ll wrap up there with just a summary, and a brief conclusion, so what we see here is that gender stereotypes negatively affect women in multiple ways, and that these things can accumulate into larger disadvantages. We’re all prone to these biases, so the goal here is not to blame, but to empower well-intentioned people. Survival skills as we’ve seen can help, but they’re rarely enough and they can actually cause sort of these backlash reactions as well, so to more fully reduce and weed out bias what we need really are organizational solutions that intervene in the tendency the natural tendency to use stereotypes as shortcuts. The goal I think is really to create more organizations that look like this one from the computer simulation. We’d love to solve that pipeline problem at the bottom, and I know lots of people are working on that, but also looking at the way that we can evaluate men and women more fairly without ideas about gender influencing our evaluations. This will obviously be good for women, and here I’m gonna sound like NCWIT a lot, but it’s also gonna be very good for businesses in that it would allow firms to be able to better harness the full range of talent in their firms. Finally, it’s also as we know this is why the National Science Foundation has been so interested in promoting diversity in science and engineering it’s good for the nation as well as our nation is able to more fully harness the talent of all of our citizens. Okay, I’ll stop there. [applause] So happy to take some questions, and I think there will be microphones coming. Yep, here come microphones down the center aisle. AUDIENCE MEMBER: Yes, I was wondering if you could speak to the mix of participants in a lot of the studies you were in terms of gender mix, and underneath that are women less, more, or equally likely to commit the male advantaging even when it’s not in their own self-interest to do so? SHELLEY CORRELL: Yeah, that’s a great question I’m glad you asked it. In the studies that I presented there’s a mixture of different kinds of participants. A lot of the experimental studies are done on undergraduate samples, but you saw the psychology study was not. That was done on actual faculty that were out there, and then the orchestra study was done on people that was actually people who were actually making hiring decisions, so there’s a large diversity in the types of participants from which these conclusions are drawn. I like the question about sort of are men or women more likely to engage in these kinds of processes, and the key answer the most common finding you find across what are now just these hundreds of studies is that men and women both exhibit bias, and they do so to about the same extent. That’s the number one finding is that the bias is about of the same magnitude between men and women, and that were both prone to these biases, and this makes sense if you think about it because what we’re doing is drawing on widely shared cultural stereotypes. They’re in the air we’re all sharing the same air, so I think it makes sense that we would see sort of similar amounts of bias. When you do find differences, and occasionally in some of these studies you find that men are more biased against women than women are. Sometimes, you find women are more biased against women than men are so, occasionally, you find sort of the other possible findings, but that’s far less rare. It’s interesting when I teach this stuff in my class I always ask the students to guess what the answer to that question is before I tell them the answer. The students seem to guess that women are more biased against women than men are. That’s always their guess is that women are harder on women, so that’s what they think. It’s actually the least common of the findings, but it’s interesting that we would have that perception. I think that perception comes about to some extent because we expect more of women, right? We expect women to be more fair in hiring women than men, so perhaps we’re holding them to a higher standard even there it’s a great question.
AUDIENCE MEMBER: Hi, I was wondering if you had any like international comparisons like is the United States more or less bias than countries we might think is offering women more opportunities like northern European countries versus countries that we think of male dominated like Japan or even like Saudi Arabia?
SHELLEY CORRELL: A lot of this research has come out of sociology and social psychology which have disciplines, especially, psychology that it’s origins are in the United States. We have a lot more studies, especially, the experimental studies in the United States, but one of my recently graduated PhD students did a study where she had people evaluating startup business plans for entrepreneurs trying to receive venture capital funding, and she finds a bias against women that is if you describe the project with a women’s name on it it seems less worthy of funding than if it has a man’s name on it. This is something that maybe isn’t surprising to this group, but interestingly and to your point she did the same experiment in the United States and in the UK, and the biases were stronger in the UK than in the United States. That was the prediction going into the study because the stereotypes about men and women are more differentiated in the UK than they are here, so that’s an example. We have fewer experiments that are truly comparative, and we need that kind of stuff, but what we do know, for example, I’ve done a lot of work on the bias against mothers in the workplace, and we do know with actual data that mothers earn less than childless women, and this is controlling for all kinds of things, what kind of job you’re in, how long you’ve been on the job that there’s a pay penalty that mothers experience. That does very cross culturally in interesting kind of ways. It’s, for example, worse in Germany than it is in the United States, and it’s better in France than it is in the United States. Now when you think about that why would that be? Germany has like this great parental leave benefits. France has great daycare for very young children, and the way we understand that finding, and this is getting a little more to the point about family structure, policies that help promote family structure, sometimes are sort of helpful, and sometimes are less helpful. In the case of Germany there’s the very strong norm that women take the parental leave, and it pulls them out of the paid labor force for so long that it starts to negatively affect their wages where, for example, France has early daycare available for children, and that keeps women in the paid labor force, and that decreases the wage penalty that mothers experience. Of course, in the United States we have neither of those things so we’re right in the middle.
AUDIENCE MEMBER: I’m very aware that businesses seem less and less willing to train people over the decades, and in my age group that’s a way that a lot of women came into computer science. Today I’m still in the software industry, and no one even seems to have a concept of training. I feel it acts against women, and I wonder in your attempt to get organizational change do you see resistance to that or are people open to it?
SHELLEY CORRELL: Yeah, you’re right. Our businesses in this country, and our universities as well seem to want the people they’re interviewing to already have all the skills that they want them to have, so I do see that sort of resistance to training people. What organizations to me seem most receptive to is, and I do a lot of this training myself is going in, and sort of doing trainings these sort of “survival” skill trainings for women teach women how to negotiate, teach women how to do this or that as if women are in some ways the problem, but training that might be more skills based training like can we get people up to speed in the way that say the Carnegie Mellon was willing to do with computer science that’s a much harder sell. I mean, I think there is sort of a strong sense that people should be sort of ready to go with the criteria that we have in mind. Interestingly, I do work with some of the tech firms in the Silicon Valley and I won’t name the firm, but one of the people was sort of telling me about that one thing is that the men that came in, the entry level men just were much more interested in sort of the kind of hacking culture that is computer science, and I said, well, does this even matter in terms of how people end up doing in their job? How much does this sort of like sort of almost nerdy programming matter in terms of people’s trajectory in your firm? They said not really at all. After two or three years nobody is doing this anyway, and I thought, well, what a shame that women are coming in, and sort of feeling really kind of chilled out on the place when the kind of thing that’s being sort of selected on really doesn’t even seem to have long-term consequences for the firm, so I think it’s a real problem, and I feel like a lot of what I do these days is try to think of new ways to frame things to open doors to those kind of conversations because I think it’s really important.
AUDIENCE MEMBER: So within your research or other research, obviously, an emerging issue now is gender identity not just biological gender, but a person’s gender identity. Where does that play into these gender biases when we have biological males who are psychological females, and all those other more complex issues of gender that HR departments are struggling with, and, obviously, in academia is a major problem?
SHELLEY CORRELL: That’s an area I have students that are just starting to work on this. I always feel like the academic research has to catch up to sort of the new things that we’re dealing with in the actual workplace, and we don’t have a lot of research on this. There is a book that was written by the sociologist Kristen Schilt at the University of Chicago that looked at the experiences of trans men in the workplace, and how sort of stereotypes about gender influence them as they transition. It’s called Just One of the Guys? That title comes about because once they had transitioned they’re pretty much accepted as men by their co-workers, so that’s one of the very few studies that we have about that, and is certainly an area where we need more research, so I wish I had more to tell you, but there’s just not that much right now.
AUDIENCE MMEBER: Thanks so much it was great, enjoyed your … I’m sorry I’m supposed to stand.
SHELLEY CORRELL: It’s that and the light, I can hardly see you.
AUDIENCE MEMBER: Your research because it resonates so much. I’m sure that it was included in AAUW’s Why So Few? report, and is very useful in the workplace. One of the things that was included in that report was the Harvard Implicit bias site, and I didn’t know whether you recommended that, whether you’ve used it, whether enough people because it is anonymous have been willing to talk about it after the fact and so on?
SHELLEY CORRELL: Yeah, that’s been a very productive area of research. I understand I think somebody maybe last year came and spoke about implicit bias in this particular group, but implicit bias for those of you who don’t know is a way of detecting how quickly we make cultural associations between categories of people, and the traits that we think that we have, so can you push a button more quickly when you see the word science and then a man’s name then when you see the word science, and you see a woman’s name or something like that, and it just shows how quickly we make cultural associations that make “sense” to us in our society, and how much more difficult it is when those don’t. I find this body of research to be very convincing for showing people the kinds of biases that we all hold. If you’ve ever been in one of these participations they kind of go through a demonstration with you, and you sort of see how even verbally, and even though you’re a person whose sort of actively working on breaking down barriers you yourself are having a hard time associating say women with science as quickly as you could associate men with science, or something like that, so I think it’s very useful for getting people to see the way that gender biases continue to matter, and that we all possess them, so I think it’s very useful in that way. I’ve not personally done any of the research myself, but I will say another reason it’s useful is that there are some things for which implicit measures of bias are more predictive of our behavior than explicit measures of bias, so if you were to ask somebody questions about do you think women should be computer scientists somebody that would say no to that is exhibiting a rather high level of a very conscious bias, but these sort of implicit biases, actually, in some ways can be more predictive. They actually, especially, predict the negative reactions to women who self-promote and things like that the violations of those prescriptive stereotypes are heavily predicted by implicit bias.
AUDIENCE MEMBER: Thank you, I found that really fascinating when you were talking about the study … Sorry, I’m right here. I found it really fascinating when you were talking about the study that used the two different resumes, right? So there was the male resume, and there was the female resume, and then we kind of saw the results of that, and I was wondering if you were aware of any research that does that on different levels of difference, for instance, with names that are associated with particular cultural groups, for instance, Latino names, Latina names and how that might also make a difference?
SHELLEY CORRELL: It really does, it really does make a difference. There was a study that was done out of the University of Chicago economics department back in I think 2002 that looked at the extent to which having a common African American name affected whether or not somebody would be hired into low wage work, so that was the kind of job they were looking at. Again, it was the same kind of design where you had the very same resume, but it would have say Tyrone on it instead of Bob, or something like that, and it found a very strong bias against African Americans in the hiring process so the results looked very similar to this here, so we do have several studies out there. Not as many as we do with gender, but an increasing number that at least look at the experiences of African Americans. Recently, there was another one of these kind of studies that was done to look at the extent to which gay men were discriminated against. This was done by a PhD student in sociology at Harvard, and basically he sent resumes out to jobs in I think five different cities across the United States. What he would do is that on some of the resumes you learn that the person had had a very important position in a group when he was an undergrad, a very important position in some sort of a gay and lesbian group, and the other person was involved in something else, and found a really pretty substantial bias against gay men, although, it varied regionally as we might expect, so the bias against gay men was much smaller in states that tend to be more progressive on these issues, and worse in say the Deep South, so there have been studies that look at other forms of difference as well. What I think we have lacking here is something that’s truly more intersectional that is that really brings together in the same study a concentrated focus on the combined effect of different kinds of identities. One example that I’ll give, though, that is quite striking is on the motherhood stuff, and the author of this study was interested in the way that stereotypes about black mothers and white mothers might play out differently when people are making evaluations, and what she did in the study it was timed to coincide right with Mother’s Day, so what she did is she had people, a random sample of the United States population read about a couple that had children, and the mother in the couple was either white or she was black, and she was either a stay-at-home mother, or she was employed in the paid labor force. The question participants were asked is how much money should be spent on her Mother’s Day present? All these couples had the same income level, so this is a great example I think of something that’s a little more intersectional what they found is that white stay-at-home mothers were given larger Mother’s Day presents than the ones that were in the paid labor force, so if you were a white mom you want a big Mother’s Day present be a stay-at-home mom, but for black mothers they got bigger Mother’s Day presents when they were in the paid labor force than when they were staying at home, so we sort of see the way that cultural conceptions about parenting are not only gender they’re also simultaneously raised, but these are the kind of studies I would love to see more of.
AUDIENCE MEMBER: Thank you, I was curious to understand what role as you looked at this introversion and extroversion plays in that bias if any, meaning if you’re an extroverted female is there any more or less bias versus an introverted female or male?
SHELLEY CORRELL: Yeah, it’s like I guess the prediction would be just in general to the extent that extroversion starts to look anything like being assertive or what have you that that would lead to a violation of prescriptive stereotypes, but when people are extroverted they’re not always extroverted in just an assertive way they might be bubbly and outgoing, or something like that as well. In these particular studies it ends up not affecting the result. People are randomized into conditions, so there’s equal numbers of introverts and extroverts across the condition so it doesn’t matter in that particular way. In the one study I showed you those were actually scripts they weren’t acting, so that was kind of muted out, but in the other studies it’s common to collect these individual personality measures, and to see if they affect our results, and with the sort of these sanctioning of people it doesn’t seem to matter so much.
WOMAN: We have time for one more.
AUDEINCE MEMBER: Do you see generational differences in the sciences?
SHELLEY CORRELL: That’s a great question. The content of our stereotypes have changed over time. For example, there used to be a pretty strong stereotype that men are more intelligent than women that’s no longer the case. There’s no difference in how people assess just the intelligence of men and women, so the content of that stereotype has shifted somewhat, so we can see that with other kinds of things. The stereotype differences aren’t as great as they used to be on most dimensions, although, those prescriptive stereotypes that women should be nice and concerned about others haven’t budged an inch, so we’re really kind of stuck on those for sure, but amongst the sort of does gender, I mean, does age of say the participants matter, yeah, you do see it. I think things are getting a bit better in this regard that we do see sort of smaller biases in general against younger people. It’s not a tremendous difference, but we do see some of that. If that sticks and doesn’t wash out as they themselves get older that would be encouraging news.
WOMAN: I think we have time for one more, so is there a hand over here?
SHELLEY CORRELL: There’s a hand over here I see.
WOMAN: Over here, oh, over there.
SHELLEY CORRELL: We may not have time she’s so far.
WOMAN: I can project [mumbles]
SHELLEY CORRELL: And we still like you.
AUDIENCE MEMBER: So as you were talking about generational differences because I’m looking at the three male leaders in my office, and we only have three leaders and they’re all male, all of their mentors were women. Lucy and I met with my boss and he had to leave because he had to get the kids from daycare because his wife had to work, so as you see that shift and actually every man in my office when their kids are sick they stay home from work because their wives have jobs is the response we get real quick. My wife has a job I have to go home. Do you see that changing, you know, there was the Eagly study that was done out of Harvard showing that there is this shift in the role men are taking at home in terms of housework, cleaning, et cetera, do you see this shift in how those men as they get into positions of power how they drive that down?
SHELLEY CORRELL: Yeah, I mean, we have seen a shift in men doing more housework, and in particular doing more child care that’s where we’ve seen sort of the biggest increase there, so that’s another sort of encouraging trend, and in some ways if you think about women who are partnered with men, and they have children to the extent that men are doing some of that it also means that women aren’t having to do as much which is good for avoiding some of the bias that mothers face, so I do think this is a trend that we’re seeing, but at the very top, I mean, here it depends on what kind of job we’re talking about. At the very top with the sort of the more extreme careers people who work over 50 hours a week, for example, we don’t see as much help from men in that sector, so that’s an area where we still I think really need some improvement, so women who are married to men who work more than 50 hours a week, for example, are very likely to cut down on their hours, or opt out of paid labor the other way around it doesn’t happen. Men’s hours are not really responsive to how many hours their wives are working. There’s still a lot of improvement that I think needs to be made, but it is encouraging. Somebody who teaches college students at an elite university I’m very encouraged at least by what they say at this point in time about how they want their lives to be in terms of being fully involved with their children, and having a partner whose fully involved in the paid labor force.
WOMAN: Thank you, please join me in thanking. [applause]