Am I benevolently sexist?

In her blog, xykademiqz just posted Benevolently Sexist, which I excerpt part of here:

For probably several years now he has been spearheading this notion, backed by research but not in the literal form he seems to espouse, that we need to pitch our field as the haven for those people who want to help others and that we need to do it specifically so that we would attract more women students.

…

On the other hand, there are several things that are sexist about this attitude. First, it assumes that, deep down, all women want to be nurses, and that one has to appeal to a smart woman’s inner nurse in order to bring her—nay, trick her!—into the physical sciences. It also assumes that while men are naturally geeks, women could not possibly be real geeks or like the physical sciences for the same reasons as men, or for any reasons unrelated to their inner nurse.

I don’t know what one has to do to get this through people’s skulls: There are women geeks. Honestly, they exist. *raises hand to be counted* There are women who like and are very good at math, physics, chemistry, computer science; who play video games; who like science fiction and fantasy.

Go read the whole post, and the comments attached to—they are thought-provoking.

I’m a little uncomfortable responding to the post, because I have also held the view that we could get more women into engineering if we emphasized some of the useful and helpful things engineers can do, rather than just assuming that people would sign up for the coolness of the math and programming.  Am I, then, benevolently sexist?

I have no evidence that emphasizing “helping” would make any difference to the abysmal gender balance in engineering, but it is one of the few suggestions I’ve seen that might help, and as fadsklfhlfja said, it would be a good thing to do even if it had no effect on the gender balance, so I’m comfortable recommending that engineering programs pay more attention to how they can help people.

Bioinformatics and bioengineering, my current fields, attract more women than other engineering fields at our university (though still not to parity, unlike biology, for example). The worst gender balance among undergrads here is in electrical engineering, and the next worse is in computer game design (despite an almost equal gender balance on the faculty for the department that runs the game-design major).  The EE ratio may be explainable by math phobia (though I think it has more to do with the way the EE courses are taught), but the game design ratio seems most explainable by the “usefulness” theory, as game design has all the coolness and employability factors one might want, except that.

I have no interest in tricking anyone into pursuing engineering—I only want the ones who will pursue engineering diligently (and preferably passionately). If anything, I’d like to send away the students who are just in the field because their parents think they ought to be.  But I think that a lot of students go through high school with really bad stereotypes of what engineers are (Dilbert, for example) and spreading a more accurate and honest message about engineering would go a long way towards improving gender balance.

We have a couple of concentrations in bioengineering that are very close to other majors that have bad gender balances:

• the Assistive Technology: Motor concentration is very close to the Robotics Engineering major.  There are a few extra bio courses and a corresponding shortage of upper-division tech courses, but the cores are quite similar.  The main difference is that assistive technology stresses the application of robotics to helping people with movement disabilities.  Once this concentration has existed long enough for statistics to be meaningful, I’d be interested in comparing the gender balances in the concentration with gender balances in robotics engineering.
• the Bioelectronics concentration is close to the Electrical Engineering major.  Again there are chemistry and bio courses that the EE students don’t take, and a corresponding shortage of some of the more esoteric upper-division EE courses.  The application is interfacing biological systems to computers.  Again, I’d like to see how the gender balances compare in a few years, when there have been enough students through the concentration for the statistics to be meaningful.

From what I’ve seen of the statistics so far, the bioengineering program here is doing a reasonable job at retaining women and under-represented minority students, but recruitment is still a problem—the ratios for our majors (juniors and seniors) are essentially the same as for our proposed majors (freshmen and sophomores), so we need to get better at attracting women and minority students to the field. If putting more emphasis on how the engineering we do helps people has any positive effect on recruitment, we should definitely do it.

Male- and female-dominated fields

In Percentage of Bachelor’s degrees conferred to women, by major (1970-2012), Randal S. Olson posted the following image:

History of gender balance in different fields in college.

He makes the point that there is no “STEM” gender gap. Indeed, the sciences and math are doing fine on gender balance. There are, however, large gender gaps in the engineering and computer science on one side and health professions, public administration, education, and psychology on the other. The post with this graph talks mainly about the computer science and engineering gender imbalance, which is somewhat larger than the gender imbalance on the other side (particularly if you take into account that about 60% of bachelor’s degrees now go to women).  He talks about the other side of the gender imbalance in The double-edged sword of gender equality, though without shedding much more light on the subject.

Computer science is a particularly strange case, as it has seen more fluctuation both in raw numbers of students (data not shown here) and gender balance than any other field. Other fields have seen large shifts in gender balance, but they have generally been gradual and nearly monotonic—not reversing course in the early 1980s.  It seems to me that the biggest drops in the ratio of women in CS came at times when the overall number of students in CS was dropping (like after the dot-com bubble burst in the 2000).  When CS grew, the number of women grew faster than the number of men.  When CS shrunk, the number of women shrunk faster than the men.  Perhaps if CS education had had a steady growth, rather than the boom-and-bust cycles that have plagued it since the late 1970s, it would not have had such a mysterious rise and fall in proportion of women in the field. The boom-and-bust cycles are not driven by the real need for CS degrees, but by media hype about relatively small shortages or excesses of personnel.  I believe that the demand for CS degrees has been stabler than the supply (unlike most other fields, where the supply has been steady even as demand has fluctuated).  Sorry, I don’t have statistics handy for that, and I’m too lazy to spend hours going through the government databases trying to match up labor market information with degree information.

Fixing the gender gaps so that most fields can draw from the full population will be difficult. Getting more men into the health professions and education could probably be solved fairly easily by paying more—and there is no societal need for more psych and public administration majors than are currently being produced. But, because CS is already a high-paying field for which there is more demand than supply, the difficulty of getting more women to choose and complete the major is a societal problem that seems difficult to address.

Some people have suggested that eliminating H1B visas for importing temporary CS workers (who are predominantly male) might help.  I don’t think that the number of H1B visas is large enough to make that big a difference, though I support replacing the H1B visas with green cards.  If there aren’t enough American workers in a field, we should import the workers on a permanent basis, not with a temporary indentured-servitude system that just serves to export the technical expertise when the workers are sent home.

Some people have suggested that a big part of the problem is the disrespect women are treated with in some workplaces—which would help explain the “leaky pipeline” phenomenon, but not why female high-school and college students are not entering the field. Student choices in high school and college are shaped much more by peer pressure and mass media than by anything about the future workplaces—so the problem is one of changing the culture in high schools and colleges—a difficult task.  There has been some success at some smaller schools (like Harvey Mudd), but a large part of that has come from aggressive admissions policies that aim for gender balance in the field at admissions time—a route not open to public schools, who can’t apply large differences in admissions based on gender.

I’m currently in charge of a bioengineering program, whose graduating class was about 36% female (13/36), and a bioinformatics program that is so small that statistics are pretty meaningless (only 2 graduates a year, both male this year). I would like to see the number of women in majors increase, particularly in the concentrations that lead to higher paying jobs (the concentrations that are further from MCD biology).  We get a few students switching to the bioengineering from MCD biology, but not many, as those students don’t take the rigorous math and physics needed for the bioengineering degree—we really have to get our students in the first year.  I’m still trying to find ways to reach those students who would be good engineers, but don’t realize it until too late.

 

Storytelling to close the gender gap?

In Closing the Gender Gap in STEM Fields With Stories, Bethany Johnsen wrote an

Making science classes more “like that” is also the suggestion of a recent Scientific American blog post, To Attract More Girls to STEM, Bring More Storytelling to Science. Its authors, teachers at a STEM-focused high school, argue that the reason for the gender gap in the STEM fields is not a shortage of girls with ability, but the failure of our science curriculum to engage their interest and kindle their passion. The remedy they propose—telling the stories of science—could lend the STEM fields some of the allure traditionally left to the humanities.

While I agree that the shortage of women in STEM fields is not due to a shortage of girls with ability (the dominance of girls at middle school and high school science fairs is clear), I’m not convinced that a story-based approach is going to work. History of science is not science, and stories about scientists are not science. Replacing science instruction in middle and high school with stories and history would leave students less prepared to study and do real science, and more likely to choose a humanities field in college.

Note that there isn’t a gender gap in biology (at least not through grad school—there is still some gender gap in paid jobs), so the problem isn’t with “STEM” as a whole, but more specifically with the math and computation-based STEM fields.  Even among those fields, there are wide disparities, with math itself coming much closer to parity than physics or computer science.  Why?  Is it something about the field, about the way the field is taught, about the culture of the practitioners, or about the culture of the students currently majoring in those fields?

Making the science instruction more interesting is a good goal, but the suggestion of the SciAm blog post “How many engineering teachers include a fiction book like Kurt Vonnegut’s Player Piano in their syllabi?” seems to me to miss the point.  Replacing science and engineering with fiction reading will not result in more students studying engineering and science—it will result in students studying literature and thinking that they are studying science.

The basic idea—to use a more story-telling approach to teaching STEM—is a good one, but I think that the stories have to be intrinsic to the science and math, like Dan Meyer’s The Three Acts Of A Mathematical Story, not stories about science, which seems to be what both blogs are advocating.

I don’t know how successful approaches like “Storytelling Alice” have been—it is no longer available though the web page claims it was successful:

A study comparing middle school girls’ experiences with learning to program in Storytelling Alice and in a version of Alice without storytelling features (Generic Alice) showed that:

• Users of Storytelling Alice spent 42% more time programming than users of Generic Alice.
• Users of Storytelling Alice were more than three times as likely to sneak extra time to work on their programs as users of Generic Alice (51% of Storytelling Alice users vs. 16% of Generic Alice users snuck extra time to program).
• Despite the focus on making programming more fun, users of Storytelling Alice were just as successful at learning basic programming concepts as users of Generic Alice.

Of course, Alice is not the most fun programming environment for middle schoolers (I think that Scratch beats it hands down), so the storytelling component may just have made it a bit better.  Has anyone ever attempted a Storytelling Scratch class? (I wasn’t able to find any equivalent to Storytelling Alice using Scratch in a very brief web search.)

The newest version of Scratch (2.0) runs as a Flash program in the browser, and has some new media-related features (like being able to interact with the video from the computer’s camera).  My son has played with it a bit, but I’ve not had time to explore the new features.  The Flash-based Scratch means that no installation is necessary to run programs, but that Scratch will not run on iOS devices (like iPads), which could be a limitation at many schools.  I understand that an iPAD app or HTML5 implementation of Scratch is planned, now that Scratch 2.0 has been released.

A better approach than stories about science may be to have more hands-on science and engineering, where students learn the science and engineering in order to accomplish something, not just to pass a course and get into college.  So far, most attempts along those lines have favored stereotypically “boy” goals (robot sports, for example, and video games), and so have not served to shrink the gender gap.

Harvey Mudd College

The NY Times published an article by Katie Hafner, Giving Women the Access Code, which is a mixture of a puff piece for the president of Harvey Mudd College (Maria Klawe) and a discussion of the partial success they’ve had in raising the number of women in their computer science program so that “this year, nearly 40 percent of Harvey Mudd’s computer science degrees will go to women”.  That is still not at the goal of 50%, but it is a lot better than most colleges are doing (32% at CMU, 30% at MIT, 10% at UCSC).

Note: the game-design major at UCSC, which dwarfs the traditional CS program, has almost exactly the same gender ratio as the traditional program.

The article by Hafner attributes the success of getting women to continue in CS at Harvey Mudd largely to a new intro to computer science course intended for students who have not had prior programming experience, taught in Python instead of Java and concentrating on applications of computer science across science.  If that were all it took, UCSC would have much better ratios, as we’ve had CMPS 5P  and BME 60 classes for several years, which look superficially very similar to Harvey Mudd’s “gold” intro to CS. They don’t seem to be resulting in many conversions to CS majors, though.

I think one major difference is that a course in computer science is required at Harvey Mudd, so more students who are initially reluctant to try it end up discovering that they enjoy the thrill of creating their own programs and debugging them.

Another major difference is that Maria Klawe, the president of Harvey Mudd, is herself a computer scientist, and so serves as a very visible role model.  At UCSC, the computer science faculty page lists 37 faculty, including lectures and adjunct faculty, only 4 of whom are women.  Computer engineering is not doing much better with 4 out of 30. Biomolecular Engineering (which is where bioinformatics is taught) has 3 women out of 16 on the faculty page, which is slightly better, but very low for a bio-related science. None of the women in the computer sciences at UCSC are in particularly visible leadership positions.

The faculty gender ratios in computer engineering and computer science used to be much better at UCSC (say 15–20 years ago).  It looks like recent hiring has been almost all male.

Hafner’s article was distributed to the UCSC School of Engineering faculty by one of the (male) computer science professors.  I think it was intended to convey the message that it is possible to do a hell of a lot better than UCSC has been doing on gender ratios in computer science.  It caught my eye both because of my own interest (I’m currently scheduled to teach BME 60 next year and hope to covert a couple of students from biology to bioinformatics), and because Harvey Mudd is one of the colleges that has come up in discussions of places that might be a good fit for my son.