>>: All right. We're going to get started. ...

advertisement
>>: All right. We're going to get started. I'm really pleased to welcome you to the third installment of
our gender diversity talk series. We have two great talks today. But before those get started, I'd like to
welcome Peter Lee up here. [indiscernible] next to give us a quick introduction.
Peter Lee: Thanks.
And I did write notes since I wanted to be sure to get all of the details right on our two very, very
distinguished speakers.
Before I do that, let me just set a little bit of context, what the gender diversity lecture series is about. I
think it's been just a terrific series. And we're halfway through now. This is installment 3 of a planned
five lectures this year. And I think in talking to the organizers, I'll name them in a minute, there is
significant inspiration from a company-wide Q&A that Satya held last October that really sparked a
discussion on how to bring more knowledge about the academic research on diversity inclusion to
Microsoft Research. And it makes sense to bring the best research in topics of tremendous strategic
importance to Microsoft to Microsoft Research. And hopefully that research can be translated
eventually into real action.
The series, because of that, was created with the idea of bringing the top [indiscernible] leaders and
researchers in gender diversity to Microsoft Research and help all of us, by doing so, become more
innovative partners in supporting a diversity innovative workplace. And I think it's been really
fantastic. The first two sessions that we had this year I think were really terrific.
So I'd like to spend a little bit of time introducing these two speakers. But before doing that, I wanted
to just thank all of the people, the whole team that's been involved in organizing this, and especially
single out the three people who have really been leaders in putting this together, those being Scott
Saponas, Jaime Teevan and Rane Johnson. So please join me in thanking them for doing this. All
right.
All right. So we have two really quite distinguished speakers. And they've thought deeply about the
challenges and approaches to solving the issues in the gender wage gap in computer science and in
engineering. And I think there's going to be sort of a tag team approach, if I understand the plan.
So to start off today, we will have Dr. Catherine Hill, who is a Vice-President for the American
Association of University of Women, or AAUW. And you can see the logo there. And besides being
the VP for AAUW, Dr. Hill is also a co-PI on an NSF sponsored project on women in engineering and
computing.
It's worth saying just a little bit about AAUW. I spent some time this morning actually on their
website. And they actually have quite a bit of very interesting content there, if you haven't seen it. I
think it's worth spending some time. The nation's leading voice in promoting the equity and education
for women and girls. Interestingly, it was founded way back in 1881, which is quite an amazing thing.
And today has over 170,000 members and over 800 college and university partners, not including, just
to verify this morning in discussion, the University of Washington and Seattle.
They take and examine positions on fundamental educational, social, economic, and political issues.
And in doing so, helps to bring the latest thinking in policy and research on these fundamental issues to
wider accessibility.
Dr. Hill, before coming to AAUW, was a study director at the Institute for Women's Policy Research
and an assistant professor at the University of Virginia. She has a bachelor's and master's degree from
Cornell University and a Ph.D. in policy development from Rutgers University. Today she focuses on
translating scholarly research into reports that are accessible to a wide audience, including to the media.
She will begin today with a talk on the latest data on pay differences by gender and computing
compared with other historically male fields such as engineering. And we'll look at both the long-term
trends as well as projections for the future. So that's part 1 of 3.
Part 2, then, will be our second speaker, who I think many of you know already, Dr. Nancy Amato, who
is a Unocal Professor of Computer Science and Engineering at Texas A&M where she directs the
Parasol Lab. She was also for a time in 2013 and 2014 the interim head of the department.
Dr. Amato's main areas of researcher are in motion planning and robotics, computational biology and
geometry and parallel and distributed computing. She's graduated 17 Ph.D. students with all but two
going on to careers in research in either academia or research labs. She has an undergraduate degree in
Mathematical Sciences and Economics from Stanford, masters degree from UC Berkeley and a Ph.D.
from the University of Illinois. She was an AT&T Bell Labs Ph.D. scholar, received an NSF career
award, an ACM distinguished speaker, a IEEE RAS distinguished lecturer, and it just goes on and on,
including being on the CRA Board of Directors and the co-chair of the CRA-W, as well as the NC with
academic alliance.
Interestingly, we are having -- I feel a little bit guilty. Dr. Amato has had to come to Seattle three times
in I think about a two-month period. She was here apparently just a few weeks ago. And next month is
the ICRA Conference, for which she is the program chair. ICRA being one of the top robotics
conferences, and it's being held in Seattle. So despite those, she's coming again today. And I'm glad
there's a little bit of sunshine.
In 2014, Dr. Amato received the 2014 CRA Haberman Award, really quite an honor, as well as the
Inaugural NCWIT Harold/Notkin Research and Graduate Mentoring Award. She's a fellow of both the
AAAS and the IEEE.
So after Dr. Hill, Dr. Amato will come and discuss how Texas A&M specifically tackled the wage gap
challenges it faced and how they increased in number of female faculty to nearly a quarter of the
department and what this has meant for the culture of the Computer Science and Engineering
Department. That's part 2.
And then after that Dr. Hill will come back to the podium, to the stage, and transition to a discussion on
practices that can help women pursue historically male fields like computing. And this will be based in
large part on a forth-coming report from the AAUW entitled Solving the Equation, Women in
Engineering and Computing. And as part of this, this will include a special focus on ways that men in
particular can help create a workplace climate for women and people of color.
I think we're really, really fortunate to have two such highly distinguished speakers with us today. I
hope we can all learn a lot and be inspired to take our learning into action.
So with that let's have our first speaker.
[Applause]
Catherine Hill: All right. Do I need talk into this mic?
Peter Lee: You can walk around.
Catherine Hill: Okay. That's nice. I'm going to start off really simple. What is pay gap? It's a ratio.
All right. I'm not going to stay too long on that, because I think you guys can do that math. But the
pay gap is an important issue in this economy. And it's been an issue in America for quite a long time.
In fact, AAUW's first study was back in 1896. And we found a pay gap back then.
In the modern era, on the way the pay gap is measured, is men and women's typical, not average
earnings, and it's based on any of the current population survey or the American community survey.
And it can be looked at in terms of hourly, weekly, and annual. And it's always full-time workers. But
that's 35 hours or more. So there's full-time workers, and then there's very full-time workers. And
sometimes, especially for annual, for example, it includes self-employed and includes annual bonuses.
And I bring up all these weeds, because the numbers, when you see different numbers, people often
think, "Well, it must be wrong," "Oh, it's just not really counting for everything." Well, this is actually - the reason it varies is because there are many ways to measure it. But we have some historical ways
to measure it, which is normally the annual number, weekly numbers often used, and sometimes hourly.
It depends on what question you're asking.
All right. So this is just a description of where the pay gap is in terms of the annual earnings. And I
share this with you, because it starts back in 1960 and moves all the way to basically the present. And
it shows you that there really have been three periods in terms of the annual pay gap. And one reason,
by the way, we continue to use that annual number is that it has this historical background. For the
weekly and hourly earnings, we don't have that same dataset in the same way.
So what you'll see here is that there was a big increase in the pay for women through the periods of
1980 to 2000. And since then it's been pretty flat. And in the beginning it was pretty flat. So the
question in terms of where we're going with the pay gap depends a lot on which trend you think is
going to be most representative of where we're going in the future. Are we going to have a lot of social
change in the future or are we going to have more of a stagnant or flat pattern?
I wanted to point to education, because this is one of the big reasons behind improvements in gains in
the pay gap for women in the 80s and 90#s# and 2000s. So what you see, this is the 70s, and goes up
to a little past 2010. And what we see is that women have made gains, a lot of gains in education. So
this is the percentage. In fact in bachelor's and master's degree, in 1982, roughly '82, we saw more
women were getting degrees than men. And that's still true today. Although it sort of leveled off. It's
not continuing to climb. Like was in the doctor's level, the Ph.D. level, were now slightly above parity,
slightly.
So this tells you a big piece of the story. Well, what we have to question now is when we look forward,
going forward, where are the new gains for women's employment going to be and for women's wages?
Okay. I want to show you just a few other things about the overall number, and then we'll focus in on
computing and we'll look at college students. So overall, this is a weekly earnings. But what you see
that women and men, we see a gap at every level of education -- so but less of a gap at the high school
and less in high school and associate's degree level, and a bigger gap at the professional level. 74
percent.
So it's interesting, education of course was really the crux of how the pay differences narrowed so
much in the past, and yet we see bigger differences among men and women in terms of the pay in
professional fields.
You'll notice, just for all of you with a Ph.D., you'll see the salaries go down a little bit with the Ph.D.
from the professional field. And make your friends who are doctors and lawyers who can pay for your
lunch. So your earnings by age is another important factor. Of course when we start out, young people
earn very similar amounts. There's a smaller gap, 90 percent in ages 16 to 19. And of course during
the prime earning years, during the years of their 30s and 40s and 50s, when you're making the bulk of
the income you're going to bring in, there we see a higher, a larger gap in men and women's earnings.
And these of course is by race and ethnicity. And we see gaps across the board. But the gaps are really
different at different racial groups. So for white and Asian men and women, we see a gap of 78 and 79
percent. In the other fields it's a smaller gap. So the gap itself between Hispanic and Latino men and
women is only 90 percent. Of course you also have to consider the overall earnings. Because of
course we don't want to close the gap, right; we want to raise women's earnings and raise men's
earnings. We don't want to close the gap at all costs.
I wanted just to point out that when we talk a lot about pay, but we need to look a little more broadly
than just pay. We need to look at all the benefits and compensation. I think all of us are probably
aware of the different kinds of compensation that is tied to our paychecks. So pension plans are often
tied to our paychecks. And so over time we see that women -- men are the darker color, the brown or
mustard yellow, and the gray is women. So what you see is that it's more likely for educated men to
have pension plans than for women to have their own pension plans. And that's of course important.
It's one of the reasons why we see women having such -- being so much more likely to be in poverty in
their old age is pensions, and other kinds of savings of course.
All right. Well, this should be sort of familiar. I don't know if you can read these. But this is just the
regular, everybody, civilian employed population. This is computer programmers, computer support
specialists, software developers, computer systems analysts, industrial engineers, electrical,
mechanical, and civil engineers.
So let's go back to, say, software developers. There's an 87 percent gap between men and women in
that field. And there's a similar gap, between computer programmers here is actually a 92 percent gap.
And that's actually really fairly small. And so what we have to say is that within computing, the gaps
tend to be much smaller than they are elsewhere in the economy. This is elsewhere in the economy.
And in computing, we do see differences, but they're not as large.
And if you want to understand the next question, of course -- before I go on to answering that question,
I wanted to just address the question of why all of this is so important and why is it so important now.
And it's because we've seen a real sea change in terms of how people are living, in terms of the
households that people are living in. And back in 1960, very few people were single mothers. It was
very uncommon for women to be the ones who were supporting the family, when they were cosupporting the family.
Today that's changed a great deal. Today 25 percent of mothers are single mothers. And another 15
percent are jointly bringing in income, a large percentage of income in the family. So about 40 percent
really can be looked at as breadwinners, and that wasn't the case in the past. So what women earns
matters a lot more to women today. Especially those that are having kids on their own, or families with
men, or just are single.
All right. So of course one of the questions that I think you probably all have is, okay, so what are you
taking into account? Are these gaps real? What are they based on? Are they based on women's
choices? Or are they based on things that women have no control over? I'll give you a heads up. The
answer is it's both, of course.
But I wanted to talk about a study that we did at AAUW where we looked at men and women at a point
in their careers when they're more similar to one another. So we looked at college graduates one year
after college graduation. We used a dataset that is available on the website for anyone that would like
to use it; it's called the Baccalaureate Beyond Dataset, and it's in the National Center for Education
Statistics website, and it's free. And they have something called the power stat tool. So if you want to
try and reproduce any of our numbers, you should be able to do that. This is not a full apples to apples
comparison. That's a little bit of an overstatement, it's not. But this is a good way people that tend to
be -- most of them are childless, they are not yet married. This is the first time all of them are going to
be working full time. We're not working at astute people that went on to grad school. We're just
looking at people that are working full time and not in school. And they came from similar colleges
and universities and similar grades. And we looked at all of these and found a pay gap of 82 percent
one year after college graduation.
And this was really interesting to us, because you know, we had a lot of -- these guys have a lot in
common. Well, we took a careful look at what some of the main issues were, what was driving these
differences. And one issue was college major. Women are more likely than men to major in some of
the fields that did not get higher salaries, the humanities, social science was sort of mixed. But
surprise, surprise, computing and engineering do very well in terms of salary right out of college.
Better than in fact a lot of other professions where you need to get an advanced degree to really benefit
from your education economically.
Okay. So we looked at hours worked was another one we looked at. Women tended to work less hours
by about three hours a week. The college major, differences in college major. Differences in
occupation. So women were more likely to go into -- say a math major might go into teaching as
opposed to a firm. So some of these things are about people's choices.
But what we did is in a regression analysis when we looked at all of these things that we know effect
earnings, and what we found is, yeah, it did explain a big chunk of it. So we have actually 93 percent
that could be accounted for by all the variables that we could find. Which suggested there's a small
amount of difference that was not -- we can't account for. Now, it could be a lot of things. It doesn't
have to be discrimination. But one of the things we were interested in is discrimination because we
take a look at the experimental data. And that's what we're going to be talking about today. We're
going to be talking about a number of different studies where we see that men and women, both, don't
consider women to be quite as good as men in mathematical tasks. So I think this raised some
questions for us and made us want to take another look at this information.
And some ideas for action, I wanted to just throw out now, and maybe we'll talk more about this later.
But some of the things that we've seen in our conversations with both employers and also with media,
and other folks who have reached us on this issue, people have experimented with using initials rather
than names on a resumé, conducting a job audit to monitor pay across companies, different parts of the
company. That's also something that's not always done. This is an idea that may not work for
computing, but it's an interesting idea of setting a price for the job. Like you get paid 80,000 for this
job and that's the price, no negotiation. If you want a different price, you have to be promoted to a
different job. And that's because there are differences in negotiation strategies for men and women.
Women tend not to negotiate as much and also lose out on some opportunities.
And of course the idea is one of the things we don't see is enough women going into the technology
and engineering fields in general. So that's really something that we need to be doing in terms of
women's choices is helping them envision that these fields can be for them.
Another issue we look at is the question of transparency in pay, which is always a little controversial.
But there's different levels of transparency and different ways to talk about it and think about your
company. So those are some of our comments on equal pay.
Now, shall I answer any questions at this point or shall we just go right on to Nancy's section?
Peter Lee: We have time for questions.
Catherine Hill: Questions? Have I convinced you all?
Yes.
>>: What do you think of the idea of starting girls very young in education for technology? What do
you think about that?
Catherine Hill: Absolutely. Well, first of all, stereotypes and biased developed at ages -- very young
ages. And by five or six or seven, you pretty much have biases in place, particularly for this one, that
math is sort of a boy thing. So I actually think that young kids are really important to work with. I
think it's also important that boys and girls play together and are encouraged to play together. They do
when they're very little, but then when you get into the elementary school, you see a lot of segregation.
>>: So would you consider the idea -- do you have any ideas for just getting rid of that stereotype
within the elementary school level? I mean are there -- how are we going to solve this problem of
stereotyping and math and science are for boys?
Catherine Hill: Yes. I think it's something that parents and teachers need to take on. Because if they're
afraid of math, it does effect children, especially the girls. So I've insisted on learning how to program
my own VCR and DVR and all that. But yeah, I think that is important that we do technical things in
our lives, whether it's building furniture we get from Ikea, or whether I'm doing some coding or
whether it's doing the basic programming we need to do just as part of our ordinary lives.
I think that we also took a look and pointed to spatial skills as an area that girls can develop and
improve on. Early, in general, boys tend to outperform girls on spatial skills as early as four or five
years old. So that's the bad news that it is a finding and a difference that we see early on. The good
news is that you can improve spatial skills by playing with blocks and drawing and drawing to scale.
Does that answer your question?
>>: Thank you.
Catherine Hill: That was a good question.
>>: So switching to the other end of the age spectrum now. You presented an interesting data point
where you stratified earnings as a function of age. And one thing that caught my eye there is during
what you refer to as the prime earning period, is that different jumps, like could you speak to what the
basic primary factors are about that?
Catherine Hill: Yes. One of the primary factors is we do see people starting families and becoming
parents. And we do still see a very big difference between women who have children and women who
don't. That said, women may want to either opt out or partially opt out for part of that time. But they
also want to opt back in. So the question, one of the things I think a lot of companies is thinking about
is how do we help people opt in, back out, back in again. And that's true for women and men in
different stages of their career.
In the past the women's -- the curve of women's labor force participation always used to look very
differently from men's. Always had a big flattening off. So they earned money, flattened out, and then
they earned a little bit more money towards the ends of their careers. Today we're seeing more of the
bell that we see for men. But it's still not there yet. So I think that we do need to keep looking at that.
That said, the majority of women with an infant under the age of one are working. So it's not that
women are leaving the work force in droves.
Another comment?
>>: Yes. Separate from child rearing component, do you think that the likelihood of promotion has
something to do with that as well?
Catherine Hill: Yes, absolutely. And also things are cumulative. So if you don't get a certain training
opportunity because your employer doesn't think you might not be interested, then you're not set up.
So every kind of employment loss or economic los in your career, it does tend to accumulate. And
yeah.
Over there in the back?
>>: So one point you made is that sometimes in some of these numbers you were counting bonuses
and some were not. Can you give a little bit more clarity? It seems like that's a critical benefit to
understand what's going on. Somehow if the base salary is different, it means that there's a prior,
before, you know, you hire the person; you're saying I expect the women to give me less in return. If
the base salary is equal and the difference in [indiscernible], it means it's a post hoc thing.
Catherine Hill: Yeah, well, I think one of the things about this area of research is the datasets are
basically the federal datasets. There isn't another way to get this information. I mean there may be
actually within companies that obviously have this information. But where we're using federal data
sources. And so the annual federal data source that includes the bonuses, because they ask you what
did you earn last year. So it's a really -- now in a weekly earnings, it's a different -- slightly different
question. And as you can imagine, you get slightly different kinds of results, because over the course
of the year you may have pros and cons; and weekly you're not going to have a bonus, right, because
you don't. And the hourly earnings, which are in some ways the best ways, you'd think why not
compare hourly? Well, we simply don't have the kinds of data that we need to do that. And also to
kind of get behind the data and understand how these decisions got made. So it's just a limitation of
some of the data. I don't mean to be [indiscernible]. And I think that's why we use a baccalaureate and
beyond dataset, because it was this great, rich dataset about these people at this particular stage of their
lives.
Peter Lee: Let's do one more question.
>>: Is there any occupation that the pay get [indiscernible] paid zero and the woman get paid more,
like marketing?
Catherine Hill: Yes, there are about 117 occupations for which data exists. There's more occupations
than that, but because you have to have at least 50,000 people in an occupation to be able to say
anything statistically. There are a handful, like three or four. I'm not sure if I can remember them all.
Some of them make sense, like let's see, there's -- counselors was one that is often like the same, the
pay is the same. Often the pay is very similar.
There's also a lot of places where the pay has -- is really close. Computing is one of those areas. So I
think one of the things we can do as we do more research and understand this problem is look at the
places where we see little difference, or in some cases, particularly among young people, women are
earning more. But that's only if you're looking at younger people. And they may know something we
don't.
>>: [Inaudible] I was just curious about the correlation versus causation, the percentage of women in
these different roles.
Catherine Hill: Yeah, you're absolutely right. It's not -- we can't prove this is causal. So they're saying
discrimination, part of discrimination can be being discouraged from taking -- pursuing a field.
>>: Actually I was wondering, I believe I've read things that fields, as the gender changes -- like the
population, more women pursue a field, and then it starts paying less.
Catherine Hill: Yeah. That's a very good point. And it's hard to measure, because you look at
veterinarians. There are many, many more women veterinarians. Twenty##, thirty years ago, that
wasn't true. They earn a lot less than they did overall. The issue is the field also has changed in terms
of the number of small animals versus large animals and that kind of thing. But I think that is
something that they found is when you have more women sort of segregated into a field, and the fields
that women were doing a hundred years ago are a lot like the fields that most women are doing today;
teaching, some kind of retail shop work, and some kind of sort of care giving. Nurses, other kinds of
medical care giving. And that's a profession, too. So everything changes, but it's not changed as much
as we'd like.
Peter Lee: All right.
[Applause]
Nancy Amato: So thanks for having me here. And Peter, I have to say that I don't mind coming to
Seattle that much, because I'm actually from here, and my parents are very glad that I have been
coming so frequently. Although this trip is pretty painful, because I left after my class yesterday, and
I'm going to take the red-eye back tonight so I can get back in time for my class tomorrow morning.
So what I want to talk about is some kind of more specifics about how we've dealt with some of the
these issues at Texas A&M. And I'll first start off talking about some institution-wide salary analysis
that we've been doing and how we're using that to basically make changes. Then I'll talk a little bit
about how we've increased the number of female faculty in our department from about 8 percent to
around a quarter now. And then I'm going to give a plug for some of our CRA-W programs that we've
been using to, I think, have helped us with the latter part there.
So first off, basically I'm going to start off and tell you a little bit about my own experience. But
basically state universities usually have fairly typical uniform starting salaries as assistant and a
standard raise on promotions.
Annual raises are basically determined pretty much entirely by the department head. And at Texas
A&M, like many state universities, our salaries are publicly available. You can find them online. So
that's also some of the transparency issues which companies have; right. You know, you don't really
know how much that person next to you is making necessarily.
But here's just my experience. I started at Texas A&M as an assistant professor in 1995. And I've been
there ever since. I was basically promoted early to associate, so in 2000. And a little bit early also to
full in 2003. So you would imagine that I was doing pretty well. So when I got promoted to associate
professor, there was a guy -- I don't know who it was. I have a suspicion. But somebody would go
around our department and they would make a photocopy of all the salaries for our department from -they'd go to the library and get that and they'd put it in all our mailboxes.
So when I saw that, right after I was so proud I had been promoted and all that, and I thought I had a
pretty good raise. Well, I looked, and my raise of 5 percent, which I thought was good, was less than
some people that were still assistant professors, you know, and ones that I knew weren't doing that
great. And they would have like 10, 11 percent raise. So I was like shocked.
So I talked to my department head. And I remember he still came to my office and I remember him
writing on my whiteboard, and he said -- those of you who don't know, my husband's also a faculty
member in computer science and engineering. He came a few years after me. So he was still an
assistant professor at the time. But he told me, "Nancy, you and Lawrence together make almost as
much as me." And I'm like, "What?" How does that -- what's the relevance there?
And then he also told me that these two assistant professors that I noticed had received much higher
raises than I did, on my promotion year -- remember, that they had families, they had wives and
children that they needed to provide for. So I was just honestly flabbergasted. And I'm not that shy, but
I think it was pretty clear that I was pretty upset. So I'm like what can I do?
So then about six weeks later, I get a note in my mailbox that there's been an error and it's been
corrected. And I then had a 10 percent raise. Which I later on found out everyone is supposed to get a
10 percent raise when you get a promotion from assistant to associate or associate to full. So I really
don't know what's going on there. But that was one of my experiences.
And then when it came time for me to go for full, I did want to go a little bit early, but I wanted to go
up for -- you basically decide you're going to put forward -- you have often no choice for the assistant
to associate. I did choose to go early at the time. That's another long story I won't share with you. But
for the associate to full, I wanted to go a little bit early, because I wanted to go on sabbatical and I
didn't want to deal with that process while I was on sabbatical, and I thought I was doing pretty well.
So I went and I talk to our associate head. He was kind of in charge of the promotion in tenure
committee, so he was the right guy to go talk to. And he'd been sitting in and doing the annual
evaluations of me every year with the department head. So you would imagine that he would be
familiar with my record. And what he told me is he told me, "Nancy, you know, people don't get
promoted full professor for service." And I was like just shocked, because I thought I was doing pretty
well in terms of research.
So then I thought, oh, boy, maybe I'm not doing that well. So I went to actually one of my colleagues,
the only other female professor in the department. And she was on the promotion and tenure
committee. And I asked her, "Hey, Jennifer, can you look at the letters I got when I was promoted to
associate? Maybe there was something wrong with them? Maybe they weren't good.? So she checked
and she came and she told me, "No, they were really good. Don't worry about it." So I didn't worry
about it, I went, and it worked. But basically I got that negative feedback that I wasn't somehow good
enough or ready to go for full.
And then -- and I think that this happened frequently. Basically because he knew I did a lot of service,
so he now thought if you're doing that much service, you can't possibly be a good researcher too, or
something. But he was a value to me every year.
After and then the other thing is that I kind of got a more global picture of how things work. A year
ago I was our interim department head, and basically the department heads are pretty much solely
responsible for deciding what the raises will be for all the faculty and the staff in the department. The
department head proposes the raises. You typically have some raise pool which you are given some
parameters as to how you can allocate it. And then the Dean has to approve it. But basically they
pretty much just agree to what you do.
So now I want to tell you about something that Texas A&M has been doing, which I experienced as a
department head. And I think it's a really great thing. So we have an advanced grant from NSF. And if
you're not familiar with NSF advanced program, it's a program which is basically trying to help
institutions make change with regards to gender equity in their institutions through maybe leadership,
promotion, pay, many different aspects. And Texas A&M has a grant. And part of that program is that
they do an annual university-wide analysis of the pay for all the tenure and tenure track faculty
members. And what they were trying to do, kind of the initial goal, was to see if there's any different
between a male and female pay rates. And correcting for things like years in, rank, et cetera.
And they also used this as an example to see if there were any differences based on race and ethnicity,
and also natural origin, were you native born in the U.S. or not. So they were kind of developing some
kind of statistical measures just to see if there were any issues. And then they also were flagging
individuals. And both of those things have been very effective. And it's been so successful that the
university's committed to continuing this after our advanced grant expires this year.
So how does it work. The researchers were -- actually they're both from Texas A&M. One is a social
scientist researcher, and the other one is a researcher basically working with the Dean of faculties
office. But they wanted to come up with a predictive model for salaries for the tenure-track faculties
that based on -- oops, that's one of my edits -- based on ten years of data. And they were trying to use
demographic variables such as gender, title, age, race and ethnicity, the years in service. It does not
account for productivity. But it's just things that are measurable. And they come up with these models,
and they develop a separate model for every college, and they allowed for different indicators within
the departments. And they got input from the deans and department heads into this model. I will talk a
little bit more about that later.
So here's some results of what happened just kind of the statistic-wise from our last one, in the College
of Engineering. So the computer science and engineering is in the College of Engineering, as well as
ECE is in the College of Engineering. So all the computer science faculty are in engineering.
And what you can see here, this is -- wait. So this is for College of Engineering, Science, and
Agriculture and Life Science. They have things like Biochemistry and Biophysics in there. But the
female salaries as a percentage of the male salaries. And in Engineering, you see that the females,
women, are making 102 percent of the male salaries, so they're doing a little bit better on average than
the men. Now there are many fewer women; right. And but it's very similar basically. Just looking at
these numbers, it seems like Engineering's doing pretty well. And it is in fact.
In Science, which includes Math, Biology, Chemistry and Statistics, the women aren't quite the same as
the men. And in some of those departments there are quite a few women. But still men are a majority.
Now, in Ag and Life Sciences, they broke it up into the ones that are considered in the STEM fields or
in the non-STEM fields. And interestingly in the non-STEM fields are doing worse than in the STEM
fields. But in the STEM fields we don't have many women generally.
Here's another one that I think is interesting. And we're looking at this for using ethnicity and natural
born. So basically there's the three values for each college. Asian, as a percentage of the white
salaries, other nonwhites. So this would be like hispanic and blacks, et cetera, as a percentage of white
salaries, and foreign born as a percentage of the native born.
So in engineering, it's looking honestly pretty good. The Asian salaries had 99.3 percent of the white
salaries. The others are 99.4. And basically parity here for the native versus foreign born.
Some of the other colleges are a little bit worse. But overall these numbers look pretty good. But they
didn't always look like this. These numbers have improved over the last three years really because of
these analyses.
This is the one that I think is more important, honestly, than the averages. Because the numbers are
really big. So in engineering, for example, we have around 400 faculty. So I think that's, from what
I've heard, that's about the same size as like to Microsoft Research number of Ph.D. researchers, so on
the same order. So those are numbers, honestly big enough, that averages aren't always that useful.
But another thing that's done is individuals are flagged. And this provides a report to the deans and the
department heads identifying faculty whose monthly salaries are not in line with that predictive model.
Now, remember, that model does not account for research productivity. It's not quality. It's just age,
number of years at rank, et cetera.
For each one of those individuals that's flagged, the department head has to either provide a
justification or address it. So the justification can be really simple. It's warranted by productivity.
Then the dean must review that and submit that report to the dean of faculties.
Now, I had to do this as an interim department head and there were individuals in our department that
were flagged. None of them caused me to actually make a salary adjustment for the tenure, tenuretrack faculty, but I did get another report, which they've started running, on the teaching focus faculty.
I did make one adjustment based on that. We had -- this is all recorded. We had someone who was
significantly underpaid based on the statistics. And by 20 percent. And I was able to justify putting
forward a 20 percent raise, when we were given a 2 percent raise pool. But it was basically because of
these statistics that I was given that I was able to do that.
Now when you do -- you do have to be careful. You can't just write "warranted because of
productivity," because then the dean is going to look at your annual evaluations that you're putting
forward, and if that person is not getting -- if they're not getting a negative evaluation and they're
underpaid, that's going to be a problem. So these are taken very seriously. But I found this really
useful. And I had no idea what's being done honestly. Although it's very public in the reports on the
website, it's just something they didn't really talk about too much in terms of this, but it has had some
impact.
So here's just to show you a kind of percentage of individuals that get flagged. So in engineering last
year in the 24 one, so these numbers -- so in engineering there were 5.3 percent of the female faculty
that were flagged because their salary was lower than predicted by this model, and 5.8 percent of the
men. And there were 5.3 percent of women that were flagged because they were higher, and 8.7
percent of the men were flagged. And for the other colleges too.
You know, business was kind of interesting, because they didn't have any that were flagged to be too
low, but they had the number that were flagged because they were very high. But it is, even after the
third year of doing this, there are still individuals that get flagged through this process.
And initially when they started doing this it was met with distrust by the deans and the department
heads. They basically did not want to have to deal with this. They thought it was going to be bad data
and just another thing that they were going to have to deal with.
The PIs who ran this study, they basically went around and met with all the department heads and the
deans and they made sure that they spent serious time with the departments who ended up with a lot of
faculty that were initially flagged. They realized where those problems were going to be and spent
time with them. And they got feedback on the model. And in fact they brought in other variables to the
model. So other things that they ended up adding to the model that they originally hadn't was the rank
at hire. So that when people came in, like an associate or full professor rank, their salaries tended to be
higher. So that needed to be adjusted for appropriately. And that came from the College of
Architecture, for example. And then that helped them basically kind of improve the model. And also
they could explain it to the faculty and the department heads so they knew where it was coming from.
The dean of faculties finds this extremely useful, and they're planning to continue with this after this
program is -- you know, the advanced grant is over. And they said that initially the deans and
department heads were unhappy with it, but now they believe it and it's just kind of part of the annual
salary process. And they said that from the -- our dean of faculty said that every year ten to twenty
faculty get raises because they show up on that report. And it's not just women. I mean it's helping
everyone. It's better for the whole institution really. And it's something that maybe came from this
advanced grant that was focused on women, but it's been good for everyone overall.
So this study was done by Lori Taylor, who's a faculty member in our Bush School of Government and
Public Service and Jeff Freud, who's in the dean of faculties office. Freud is very happy to answer any
questions or if people have questions about it. And we have -- this report is online, and actually the
previous ones are as well. So if you have any questions about this, I'm happy to try to take them. But
if you have methodological studies or questions about this study, it would be better to go to Lori. But I
have seen it from the perspective of a department head, and it was, I think, really valuable, frankly.
So I guess I will kind of switch gears here. Should I ask for questions? If anyone has any questions
about that, I think it would be good to do something like that. I don't know if you have something like
that here. Do you have analysis of your ->>: So we do. On salary and bonus, we actually do have -- it's highly sensitive but we do have data,
and there's a system of checks and balances that HR runs on all of the Microsoft management to make
sure that outliers are flagged appropriately. One thing that I found missing, and I wonder if it's an issue
at Texas A&M, is one thing that I found was missing in the tracking is promotion velocity. And I think
it is possibly less of an issue, especially early in the academic career, because there's a clock for the
promotion to assistant professor. We don't have the same sort of timeout here. So I've been very
curious about promotion velocity and other things. But it's certainly for salary and bonus, it is an
incredibly useful tool. And I think HR at Microsoft is very hard notes on all of the management in sort
of flagging.
Nancy Amato: So do you have to justify why you have ->>: Yes.
Nancy Amato: So that's the thing that I found -- first off just making sure that people are aware of it
was helpful. And you have to justify it or do something about it. For the promotion velocity, that's not
in this salary thing, because that's only on salary. But they have collected that data and they use that
data to identify, main people get stuck at associate professor level. And that is -- women do tend to stay
longer as associate. Because it's your choice as to when you go up. That was one of my things.
>>: One thing I wonder about is at State University with total transparency of salaries, it's interesting
that people who maybe have been chronically underpaid for whatever reason, so what does that do to a
culture of a place -Nancy Amato: Morality?
>>: -- [inaudible].
Nancy Amato: Well, I think that it's in a public somehow -- it's an indication of how you're valued by
an organization. So if you're receiving a poor salary, you don't feel good about it; right. But salaries
don't -- we don't get to go and complain about salaries. The only reason why they'll give us a raise is if
you're kind of a retention issue. They will give people raises for those occasionally. But like for A&M,
they limit it actually, say, "You can have it once." Because there was a problem, people would go out
and get offers, right, and they'd do it every year.
But I think -- I don't know. We're just used to it. It's a different -- you know, it's just like awards, right.
People get awards and you feel like you're better or worse than those other people. Transparency's
always good, in my opinion. And people are less likely to let inequities get really obscene because they
know that it is public knowledge.
But as you all know we're getting a very highly paid university president very soon. So A&M's hired
the University of Washington president for -- I think he's going to get $1.4 million a year. Which is a
lot for a university.
>>: More than you and your husband combined.
Nancy Amato: Yeah. So okay, so I guess I'll move on to the next thing, which isn't so much here. But
basically diversity -- in my experience, I think that overall most people aren't evil. They are just -everyone I think kind of agrees that diversity would probably be a good thing. Sometimes they're just
not knowledgeable about it. And sometimes there's also just a general lack of sensitivity they're not
aware. So we end up with a lot of unintended difference in treatment and expectations.
So for example, from my personal experience, when I went to Texas A&M, we have an industrial
affiliates program like most universities do. And typically what would happen is the new professors
would be asked to give a research talk to the -- you know, brief research talk to the industrial affiliates
to let them know what they're up to.
Well, the day before -- I showed up in January, which is kind of unusual. So normally it's in the fall
when they have the professors do the -- the new professors give their talks. But I came in January.
And the day before this meeting, our department head, who had hired me, right, so he obviously
thought I was a good addition to the faculty, he asked me to come and talk to our industrial affiliate
about women in computing. And not about my research. And honestly until that point I hadn't really
done anything with women in computing. I didn't know anything about it at all. So I knew that they
were starting some women in computing group at the University of Illinois when I left, but I honestly
hadn't been too involved in it. So I didn't really know anything about it at all, so I was frantically trying
to find information, and I did that, and then I started a women in computing group in our university in
our department, which is going strong today and I'm really glad I did that. But that was not how -- it
was kind of different that he treated me that way.
And often we're asked to do extra service. And I like to -- I still do it a lot today, and I'll talk to you a
little bit more later on about CRA-W. But then when I was trying to go for promotion for full
professor, because I did all this service, that's how he kind of pattern matched me, "Nancy does
service," and didn't think about me in the same way as he did about the other people who maybe
weren't doing service, because he didn't know about those other things back then.
So no one ever met me ill, but it just kind of colored the way they thought about me. So I think that
that's -- we don't have -- there still are some things that are really done for bad reasons or that are really
prejudicial. But I think for most cases it's not that. It's more benign, but ends up hurting a lot of
people. So we really need to improve awareness.
Here's something that happened at Texas A&M just in terms of our faculty. When I started in 1995, I
was a second women in the department. And then I was still the second woman up until about 2001.
And then at that point we had -- we hired the next woman who, many of you know, Valerie Taylor. We
hired her as our department head. Then we hired -- since then we've been kind of steadily growing.
And now today we're about 25 percent women, which is quite high for a research department. And so
we also have ethnic diversity among those women. We've got two African-American, one hispanic,
and two Asian women.
So how did we do it? Well, I think really requires grass roots buy-in. Institutional practices, we have
required training for all of our search committees. You probably have things like that here too. But it
used to be originally they just made the chair of the search committee go. But now they make everyone
go and you have to go every two years. It's frankly pretty annoying. But you have to sit here and listen
to all that stuff and it reminds you things are important and it reminds you like those questions you're
not supposed to ask and those kind of things. And it just makes the whole process, I think, work better.
And then we've also been really good about sending our faculty and our students to external programs
that helps them build a community. Because even if you don't have that locally, they can build that
community nationally or internationally or within their research group. And I'll talk a little bit more
about that later on. I want to make sure that you guys all hear about this CRA-W and the great
programs that we have. But they've been really, really helpful for me and many of the others in our
department.
And then I think something that's really kind of key is that basically leadership. In our department,
since basically from 2001 up until now, we've had female department heads except for two years. In
2001/2002, Jennifer Welch, the other woman in our department, she was interim department head. And
that's the year that we hired Valerie Taylor as our department head. So then Valerie was our department
head for eight or nine years -- nine years I think, two four-year terms plus that last [indiscernible]. And
then we had a male department head for two years. Then I was the interim department head for a year,
and we hired Dilma DaSilva, who just started as our department head last fall. So I think that this
really does impact your ability to kind of have a good culture within your organization, and also to
recruit and retain people.
Then the last thing I wanted to talk about is to make sure you're all aware of the CRA committee on the
status of women in computing research, or CRA-W. How many of you have never heard of it? You all
have? I find that hard to believe. Or maybe that's because that's why you're here.
>>: The last lecture.
Nancy Amato: Okay. So okay, but that doesn't mean you all came. So you all know about these
things. Well, this is like one of the best groups I have been involved in in my career, honestly. I've
been involved with it as a board member since 2000. And every committee member, in order to be on
that committee, you have to have a project and you get evaluated, and you are removed -- maybe not
immediately, but it doesn't take too long if you're not doing your job.
And you guys here at Microsoft Research would be really proud of this too, because the last two -- it's
governed by co-chairs, which serve two-year terms, and the last pair of co-chairs included Catherine
McKinley, and she was co-chair with Tracy Camp from Colorado School of Mines. And now I'm a cochair with AJ Brush. So we've been doing this for -- it started in September as a co-chair.
So well, then, I really won't spend too much time if you already have all had this. But basically CRAW has programs for undergrads, which are research experiences. The DREU program, I've been
running that since 2000. I think it's a fabulous program. That takes undergrads from all over the
country and matches them with faculty at universities for some research. And AJ was -- when it was
called DMP. But it's really effective. And then the grad cohort, this is something that Microsoft
Research is one of our main sponsors and we're really grateful to them. But that program, it gets
almost all of the graduate women in the country basically participate in grad cohort at some point. And
I think almost all the Ph.D. students. And it brings them together for a day-and-a-half workshop once a
year. And it's been really effective. And then we have CREU mentoring workshops for people that are
early in their career, like senior grad students, beginning researchers and professors, and then mid
career.
>>: And we know at least one person that's coming into the mid career in June in Portland. But I -- we
can get you in. We have space.
Nancy Amato: No, it's a great program, really good. So it's at FCRC, and you can still apply, I believe.
I'm not sure.
>>: [Inaudible]
Nancy Amato: AJ can get you in. She has an in.
But the other thing I want to make sure you knew about was CERP. CERP is the Center for Evaluating
the Research Pipeline. And this is something that CRA-W started as part of our grants a few years
back. And it's really effective. What it tries to do, so we always want to evaluate these programs to see
if their effective. And it's kind of obvious, right, you would imagine that undergraduate research
experience should be good; right? But we wanted to evaluate and compare like our programs versus
students that participate in other types of programs versus students that didn't participate at all. And
that was difficult to do. And NSF, who funds many of our programs, wanted us to do that.
So you may have heard of the data buddies project. This is where many universities are participating in
this. They allow their faculty and their students to be surveyed annually. So we can capture data about
students that participate in undergraduate research, experience, or who didn't, and so then we can
compare effectiveness of like those programs, like a standard RU to the RU programs run by CRA-W.
And here's some things that we found out, for example -- we have more later data, but these ones
showed that basically if we looked at the graduates of our programs, the green were graduates of our
programs, the orange is they were -- they did research experiences as undergraduates but not through
our programs. And the blue is they didn't do any at all. So if they were going to be enrolled in a
program, our graduates were about twice as likely. 81 percent of them were in Ph.D. programs versus
37 percent who did like a regular REU, and 18 percent of non-participants. And generally we have
good results showing the effectiveness of our grad cohort programs as well.
But this is an evaluation that shows that these programs are really effective and we need to do more of
them. And we're really grateful to Microsoft Research's support of these, and Rane in particular. So I
wish she was here.
So anyways, these are kind of the main points I wanted to make is that I think this salary analysis
institution-wide was good for everyone. It helped identify anyone who was getting inequitably paid
and made their manager have to explain it or fix it.
And increasing the number of female faculty, it took time. I mean it's taken like twenty years. But it's
really changed the culture of our department and it's made it a better place for everyone, again. You
know, fixing these problems makes it better for everyone, not just for the woman. And the CRA-W
programs have been an important part of at least our department's progress in this regard.
That's it.
[Applause]
Peter Lee: So we have time for one or two questions and then I'll switch back to the third part of the
talk today.
>>: So I'm Meredith [indiscernible]. I'm a researcher here. And I think I sent an email a couple years
ago about the distributed research mentor program for undergraduates to find out whether mentors from
research perhaps instead of universities participated in answers. We've never done that before. So I
guess I was just wondering whether -- I imagine there is a shortage of mentors in universities is
whether they're expanding to include mentors in research labs is something you're thinking about and is
that something you're going to talk with the lab leadership about while you're here visiting, whether
Microsoft is going to participate in such a program?
Nancy Amato: Well, we have thought about this over the years. And we have recently been thinking
about it. But I didn't have any particular plans to talk about it while I'm here. But it's perhaps a good
idea.
So the thing is, I just want to make you aware that program has been restricted to send the students to
universities, to professors and research groups at universities. Part of the reason for that was the goal is
to let them kind of experience a university environment and a research environment and to see what
graduate school is like so that they could kind of mentally see that they could succeed there.
>>: Undergrads.
Nancy Amato: They're undergrads, yes. So and we're trying to get them to go on to grad school. So
these would be undergraduate researchers that we're talking about. And we want to find ways to do
that. But frankly, honestly, one of the concerns that comes up when we talk about these things is that
Microsoft is going to steal these students and they're not going to go to graduate school. Now if they
come to Microsoft Research, that's hopefully much less of an issue; right. But that always is something
that people worry about.
>>: We wait to steal them until after they get their Ph.D.
Nancy Amato: Well, that's fine. That's what we want. Yeah.
>>: This one point of data since Mary brought it up, I mean that was precisely my experience actually
is I did an undergraduate internship at [indiscernible], and I was like, "What do I have to do to work
here permanently?" And they're like, "You have to get a Ph.D." And I was like, "Why do I have to get
a Ph.D. to work here?" They're like, "That's just what you have to do." And that's actually what
brought me to do it. And so I think it could definitely be inspiring to young students to spend time in a
place like this and realize how often and fun applied research can be.
>>: I'll add that when I was an undergrad I heard a couple of researchers from MSR give a talk and I
was like, "Oh, what do I have to do to do what they do?" They're like, "You have to get a Ph.D." I was
like, "All right."
Nancy Amato: So can undergraduates intern here in the summer?
>>: That's why I think you have to talk to the lab leadership to make a special exception.
Nancy Amato: [Laughter]
>>: So we have about 1200 interns in Microsoft Research labs every year. And out of the 1200, fewer
than fifty 50 not Ph.D. students. They would be high school or undergraduate students. We do host
them. Actually out of that fifty the bulk of those are in our India and China labs. But the third largest
concentration of non-Ph.D. student interns is actually here in Redmond. So I think it would be very
easy for us to expand that number.
Nancy Amato: Well, I think it would be worthwhile working on, frankly. Because that's when you
need to capture them. If you wait until their -- we're not getting them into grad school. So if you want
to get the women and under-represented groups to show up to go to grad school, hooking them as
undergrads is really the right time. And especially if this could be I think a very attractive alternative to
them versus going to just a coding internship. They'll still feel like they're going to a company, and we
can trick them into the research.
>>: I will say that it's the stories that [indiscernible] and Scott told about being inspired to pursue
Ph.D. studies about spending time in Microsoft Research. That is the number one motivation for us in
hiring undergraduate interns in our India lab. Because there are very strong motivations for bright
undergrads to not go to graduate school in India. And so -Nancy Amato: Here too though. I mean it's getting worse and worse.
>>: So I guess I just wanted to add that we also have opportunities for high school students. Right
now I'm mentoring Madelyn, who is a high school student at [indiscernible] Academy, and she's kind of
like getting a sense of how research works. And hopefully she says that she wants to go to grad school.
>>: [inaudible] [laughter]
>>: She is going to undergrad this fall. But we have talked and she seems interested.
Nancy Amato: I'm completely with you. I started working with undergraduates for the first time since
like two years ago. The very first undergrad I worked with won the Intel -- high school, right? The
very first high schooler that I worked with, he won the Intel -- the national Intel competition for all high
schoolers. And he got a $100,000 scholarship to college. And unfortunately, though, he used to want
to be a professor. He went to Google this summer after his freshman year, now he's not sure he wants
to go to grad school.
Peter Lee: All right. Let's move on to the last part of the talk.
[Applause]
Peter Lee: So let's invite Catherine back up here. She has just a few more things to tell us about.
Catherine Hill: In fact I'm going to go through all my -- no, I'm not going through all my stats and stuff
here. So what we'll do is we're going to skip ahead to the end. Because this is a really important study
that I want to leave you with. So we're just going to talk about one more thing. And if you want to see
a bunch of stats about where women are in computing and engineering, you can always download the
report on AAUW's website.
But this is a study by [indiscernible] and her colleagues. And she did a study on engineering. But I
think this study on engineering is very relevant to what you're all doing here in the computing field. So
this is about retention. And what she looked at was a survey of women who had gotten a BA in
engineering at some point. And she looked at where they have gone and what they were doing. And if
you look at women and men, right in the beginning, right after college, you see that about 60 percent,
about the same, went on. The people that got the degree went on and got a job in industry with their
bachelor's degree. Others may have gone to graduate school; they may have gone to a different field.
We don't know what they did.
So these are the years since they graduated. So this is zero to four. So you graduate here. And now
we're at ten years after graduation, and here we're at twenty years after graduation. And we start to see
a little bit of a shift. And then when we get down to 30 to 35 years, 34 years, we see a very large
differences between women and men. So the yellow is women and the blue is the men.
So at the end of that time we still see these people in the workplace. And what's interesting is you ask
them -- you don't even look at these two groups. The women who stayed in and were in the work force
with engineering degrees, and women who had those engineering degrees but had left within the last
three to five years.
Question?
>>: Yes. If you're looking at zero to four years, you're culturally looking at a very different kind of
clique of people. Now and you're looking at 34 years, which is like 1980. So I mean is there some way
that these differences in gender can be kind of normalized against this?
Catherine Hill: Right. But the reason we -- it would be ideal to do that over time, to look at the same
people. But obviously you can't. So that is actually just an artifact of this study. You're going to see
that the people at the end are different people than the people at the top. So there may be differences.
We may see that these folks who came in at 65, 64, just stay in. Yeah, we might. But this is people
who are currently working, currently not working.
The other interesting part about this, I'll just do the next slide. But that's a very good question. Thanks
for clarifying that.
So she looked at these two groups. And what she found was the women with low job satisfaction had
experienced -- sorry, this color is really bad. They're the pink, and purple is for high. So women who
did not observe -- who did observe sexist behavior, who experienced undermining behaviors and
undermining behaviors by the co-workers, okay, those were the things that caused people to leave
engineering. The two groups were similar in terms of their age, in terms of their parent status, in terms
of their confidence in their engineering skills and their interest in engineering.
The difference that came out really strongly was the workplace. That the people who stayed in
engineering had good work places, they felt supported, all those other good things. And the ones who
left the field had not experienced that.
So I think that the importance of this study is that the workplace is making a big difference. And when
we want to talk about retention, we should be talking about how these workplaces can better support
everyone working there.
One of the places that I often site for people to take a look at -- okay, we see these things that
employers can do. One of the things I want to mention that's knocking me on this is the ADA initiative,
ADA initiative. How many of you have heard of that? Okay. They have some great and interesting
valuable things to do with how you grapple with sexual harassment in the workplace, how you grapple
with cultural misunderstandings in the workplace, and they -- really I recommend them highly.
So what we found in terms of what the [indiscernible] study came up with, what can employers do?
Okay, people want to have clear roles and responsibilities. Training and development, feeling like they
were growing was very important. Many, many people mention that. Being acknowledged for their
contributions. These aren't any -- I know this isn't anything surprising. But it's just a good thing to
know that these things really made a difference in retention. And rooting out uncivil behavior. One of
the things that I show on a slide that we're not going to look at is women and men in engineering and
computing tend to have more wide differences in their stereotype of gender. So in engineering and
computing you see different kinds of stereotypes about gender between women and men. Women
being much less stereotyped in that sort of traditional role, and men being more -- tending to have more
stereotypical views.
So a particular challenge, you talk about uncivil behavior, even benevolent things like trying to protect
a female co-worker, which is a nice thing to do and should be nice, is actually not necessarily helpful in
the long run.
Oh, and the last thing I'll say is this is our report. It is on the web now and we're very excited about it.
We have all of our slides, many of which you didn't see but you can download them for yourself and
use them and do whatever you like with them.
So thank you.
[Applause]
Peter Lee: [Inaudible] for questions. We have a few minutes left. Why don't we take a few questions.
>>: I'm wondering in your studies if you had a chance to dig a little deeper into the kinds of uncivil
behaviors. Because I feel like there's always a spectrum, and sometimes it's the ones that are more
subtle, that are worse, like the woman doesn't get invited out to lunch with the rest of the people, they
don't have the say and they don't get invited out for drinks. Like you hear a lot of that in the sales
professions and things like that. Do you have a sense of like a relative impact of the -Catherine Hill: You know what we can say with some certainty is that explicit bias is really declined
enormously. It's very unlikely that someone is going to say women can't do this kind of work. Which
they used to say thing likes that. That's what they believe. But there's much more of this implicit bias,
which I know you all had a talk on already, but I think it's a critically important issue, because I do
agree with you. Yeah.
Yes?
>>: Have you done some studies based -- detecting if there is this type of varied behavior toward
women in engineering? Because I know a lot of schools; they are very science oriented. A lot of
schools like they're building these robotics clubs, science [indiscernible]. Have you taken a time to
look at studies seeing that whether this effects a girl's wanting to into a STEM field? Because I feel
like there's a lot.
Catherine Hill: Yes, I think you're absolutely right. One of the most important things about biases and
stereotypes that we take them in on ourselves. And women are much harder on themselves. And
women are getting lower grades, getting Bs and Cs.
One of the studies that we looked at in engineering students in this case, men were getting Bs and Cs,
women were getting Bs and Cs; the men were like, "I'm doing about average. I'm going to do better. I
can improve." The women were like, "Oh, I have to quit. I can't do this. So women can be very hard
on themselves. And they have higher expectations for what they think -- there's a wonderful study by
Shelly corral. It's a little old now. But she looks at -- comes up with a fake skill and basically tells the
participants this is a mail skill, spatial -- basically it's telling which is black and white imagines. You
have to estimate how much of an image is black and how much of it is white. So spatial contrast, that's
what it was called.
Anyway, everyone got this little quiz and the women, they gave them all back the same grades. And
the women thought you had to do even better. They thought they did worse. Their estimates both on
themselves and what they needed to do to be good at this were higher. So I think that's something that
we see happening.
Yes?
There's so many studies out there in this area it's just overwhelming. And what I also think is
interesting, and I think all of you probably already know this, but this topic of getting women involved
in computing and engineering is incredibly important. And people are paying so much attention to it in
Washington D.C. We just had the American Competes act is going up. And that's going to be
addressing this issue. So it's really everywhere. We find an enormous amount of interest in it. So you
guys are going to be in the limelight. And that's good.
Nancy Amato: So just actually in regards to that, I have a little plug for -- Peter mentioned that ICRA,
the International Conference on Robotics and Automation, the conference is going to be in Seattle on
the last week in May. On the Saturday, I think is that May 30#th#? Whatever that Saturday is, we're
going to have an event which we're calling "Go Girl Go," which is basically for girls, you know, middle
school and high school girls. And we're going to have about STEM careers. And we're doing it
together with Washington first and the national first robotics.
And we will need mentors. We need female professional mentors. And from any level actually. We're
interested in undergrads through senior people. So if people are interested to do that, I was actually -that was one of the things I wanted to talk to you about is I need to recruit table topic mentors.
Catherine Hill: And men can be wonderful mentors for women too.
Nancy Amato: We're hoping that we're going to have somewhere between 200 and 1,000 girls there.
It's going to be in the Convention Center in Seattle.
Peter Lee: All right. Let's thank our speakers one more time.
[Applause]
Download