This past Sunday, The New York Times ran an op-ed I wrote about why so many girls still shy away from taking classes in computer science. As a special bonus to those of you who check out my blog, here’s the long version of that essay, which I delivered as a lunch-hour talk last week at Google in Cambridge, MA. Some people who’ve read only the short version think I’m somehow saying that we should persecute nerds and geeks who already are working in tech. Would I say such a thing? Of course not. I was a nerd myself. I think this version makes much clearer that I’m saying we just need to broaden and diversify the image of who belongs in a computer science class or a tech company so that we keep the people who thrive there now but bring in people who don’t currently feel comfortable.

So, here’s the original version:

The largest high-tech companies finally have begun to devote serious resources to attracting and promoting more women and minorities. Unfortunately, if young women and students of color continue to shy away from classes in computer science and engineering, these firms will compete for the same limited pool of employees who aren’t white or male.

Figuring out why people who choose not to do something choose not to do it is like trying to interview the elves who live inside your refrigerator but only come out when the light is off. Women already working for a company might tell you what makes them unhappy. But these complaints won’t necessarily pinpoint the factors that frighten girls away from studying computer science in the first place.

As unlikely as this might seem, recent research shows that many young women avoid studying computer science or engineering simply because they are afraid they won’t fit in.

For the past six years, Sapna Cheryan, a psychology professor at the University of Washington, has been studying why girls in high school are significantly less likely than boys to sign up for a class in computer science, take the Advanced Placement exam in that subject, or express interest in computer science as a career, and why female college students are four times less likely than men to major in computer science or engineering, even though they test extremely well in math.

What Cheryan’s research reveals is that many girls don’t think they would feel comfortable in such a class. This shouldn’t be too surprising. When most of us are deciding whether to move to a new town, join a club, or take up a hobby, we are as perceptive as Sherlock Holmes in picking up clues as to whether we might belong.

Over and over, Cheryan and her colleagues have found that female students are significantly more interested in enrolling in a computer class if they are shown a classroom (whether virtual or real) decorated not with Star Wars posters, science fiction books, computer parts, tech magazines, and boxes for video games and software but rather with art and nature posters, water bottles, coffee makers, lamps, plants, pens, and general-interest magazines.

Interestingly, the willingness of most male students tends not to change significantly whether they are shown a room decorated in a stereotypical manner or a room with a more neutral décor, while about a quarter of the women reported being drawn to the more stereotypical setting. The importance of what Cheryan calls an “ambient sense of belonging” proves equally strong with female college students considering a job in an office that does not conform to their image of what a computer company might look like.

Cheryan and her colleagues also found that cultural stereotypes about computer scientists strongly influence young women’s desire to take classes in the field. At a young age, girls already hold stereotypes of computer scientists as pale, awkward young men who are socially isolated, obsessively devoted to playing video games, and intensely focused on machinery, and whose genius is inborn rather than the result of study and hard work. Given that such stereotypes run counter to the indoctrination most girls receive that they should be feminine, interested in working with other people, and modest about their abilities, as well as many girls’ beliefs that they are not innately gifted at science or math, it is no surprise that they cannot see themselves as happy, successful computer scientists.

In another experiment, Cheryan and her colleagues arranged for female undergraduates to talk to an actor pretending to be a computer science major. If the actor wore a T-shirt that said “I CODE THEREFORE I AM” and claimed to enjoy video games, the female students expressed less interest in studying computer science and less confidence in succeeding in such a class than if the actor wore a solid T-shirt and claimed to enjoy hanging out with friends … even if the actor was another woman.

Such stereotypes might seem laughably outdated. And yet, studies show that the basic stereotype of a scientist hasn’t changed since the 1950s. And these stereotypes do still have a basis in reality. In some ways, young women today might have an even harder time imagining themselves as physicists, engineers, or computer scientists than when I was a physics student in the mid 1970s. Back then, women were still rebelling against the norms of the 1950s. Most of us wouldn’t have been caught dead pretending to be princesses, even when we were girls, or wearing high heels and fancy dresses as adults. We eschewed manicures, makeup, and hair salons. Fancy weddings that required a year of planning were anathema. And we weren’t about to devote every minute to hovering over our children, the way our stay-at-home mothers hovered over us.

I’m not making a value judgment about what women should wear or how much time they should devote to their weddings or their families. But young women today clearly face more pressure to dress in a feminine, sexy manner than they did when I was young. Which means that the gap between the way women see themselves and the way they see scientists has only grown wider in the past few decades, not to mention that the messages young women continue to receive from their teachers, parents, and classmates convince them this gap is real.

For instance, one girl I know was recently told by her high school teacher that she was too pretty to be good at math. And research demonstrates that female undergraduates who have watched commercials in which women focus excessively on their appearance demonstrate less interest in careers in high tech than students who haven’t viewed such commercials.

Cheryan postulates a similar theory to explain why the percentage of women studying computer science has fallen since the mid 1980s. Citing evidence that high school students get their ideas about scientists from popular culture and the media, she wonders if there might be a causal link between the decline in women’s interest in computer science and the rise of movies and TV shows such as Revenge of the Nerds and The Big Bang Theory, which portray scientists and engineers as white (or Asian) male geeks. The media’s intense focus on the start-up culture and geniuses such as Steve Jobs and Bill Gates might also have inspired more young men than women to enter the field, Cheryan theorizes, even though very few jobs are located in start-up companies and very few computer scientists are Steve Jobs or Bill Gates.

Men sometimes scoff that if young women let such nebulous factors deter them from careers in physics or computer science, the women are simply exercising their free choice, and if girls were tough enough, such exaggerated stereotypes and feelings of discomfort wouldn’t discourage them. Yet the stereotypes are still pervasive. Just yesterday, I was standing in line at my corner pharmacy when I realized that the covers of the magazines I had been staring at showed a god-like portrait of Steve Jobs (the subject of a recent movie) at the top, then, beneath him, a big-busted female model in her underwear, then a heroic, clear-eyed Mark Zuckerberg, then, at the very bottom, another scantily clad female model.

All of us are bombarded by such images. How can we not absorb the lesson that computer geniuses are men, while the chief goal of a woman’s life should be to look fabulous in her underwear? I wonder how many young men would choose to major in computer science if they suspected they might need to carry out their coding while sitting in a pink cubicle decorated with posters of Sex and the City and Thelma and Louise and copies of Vogue and Cosmo scattered around the lunch room. In fact, Cheryan’s research shows that young men tend not to major in English for the same reasons women don’t major in computer science: they compare their notions of who they are to their stereotypes of English majors and decide they won’t fit in.

All this meshes with my own experience. Even though I felt more comfortable in a T-shirt and jeans than a dress or a stylish suit, after four years of majoring in physics I experienced so much pressure to dress and act like a man that I began to feel extremely uncomfortable about my identity as a woman. Working as an intern at Oak Ridge National Laboratory in Tennessee or attending a conference on theoretical physics in Austin, Texas, I felt out of place among my mostly male colleagues because I hated drinking beer or eating barbeque and didn’t like getting teased about reading novels.

When I confided to a professor that my dream was to attend graduate school in theoretical physics at Princeton, he warned me that “those guys” were so competitive my ego would take a beating. Suddenly, I imagined myself spending the rest of my life in a stark conference room with a bunch of ill-kempt guys shouting at each other, arguing, and ignoring me. By contrast, the first time I took a fiction-writing workshop, I was astonished at how beautiful the wood-paneled seminar room was, how well groomed and attractive the men and women sitting around the mahogany table seemed. I felt ugly and out of place, but I remember thinking I would someday love to look and act the way these writers looked and acted.

I have heard similar stories from countless women. Even the female students at Olin College of Engineering in Massachusetts complained to me about feeling out of place among male classmates who were always “engaging in pissing contests” or “burping the periodic table,” or male coworkers at summer jobs who continually cursed and made vulgar jokes.

The good news is that women’s stereotypes of scientists can be surprisingly malleable. In one of Cheryan’s experiments, young women who read a fake article claiming that many computer scientists no longer are geeks demonstrated a significantly greater interest in computer science than their counterparts who read a fake article saying that the stereotype remains true.

To make computer science more attractive to women, we might try to change how women think about themselves. But we might also diversify the images of scientists that women see in the media, along with the décor of classrooms in which they might want to study (this would be especially easy in the case of virtual classrooms for on-line courses) and the offices in which they might apply to work. In this way, we could retain the men and women who are drawn to these images and environments while attracting women—and men—who might previously have shied away because they felt they wouldn’t fit in.

After all, a high tech company isn’t like the army, where the stereotype of a warrior is inextricably linked to the requirements of the job. A stylishly dressed woman in a cheerful room with a beautiful painting on the wall is just as likely to write a brilliant line of code as a man in an I CODE THEREFORE I AM T-shirt sitting in a dingy cubicle decorated with a Star Wars poster and littered with soda cans and pizza boxes. Even small changes can make a big difference: a thoughtful male professor once wrote me to suggest that changing the name of the major from Computer Engineering to Computer Arts would double the number of female students overnight.

Of course, luring young women into a field by showing them images of offices that don’t exist and letting them read fake articles that claim computer scientists are more suave and socially adept than they really are might be unfair. But as Cheryan points out, the stereotypes are at best only partially true, and women who actually take classes in computer science don’t hold the same negative views as women who get their ideas from movies and TV shows. Besides, if companies do succeed in attracting more women and minorities, the stereotypes will change, as will, presumably, the grubby décor.

This is why Mayor Bill de Blasio’s recent announcement that within ten years all of New York City’s public schools must offer at least some training in computer science to all students is so important. As Cheryan herself suggests, requiring all students to take computer science not only would change the stereotype as to who does or does not take such classes, it would provide girls and students of color with the opportunity to experience the material and decide for themselves if they like the work.

At the college level, some fairly simple changes could allay the anxieties of female and minority students. At Harvey Mudd College, strategies such as renaming courses (“Introduction to programming in Java” became “Creative approaches to problem solving in science and engineering using Python”), creating separate introductory classes for students with zero programming experience and students who already know how to code, discouraging macho or show-off behavior, exposing students to non-geeky computer scientists, and allowing them to work on projects that fit their preferences led to an increase in the percentage of computer science majors who are female from ten to forty percent in only four years.

As an added benefit, Cheryan’s research on the effect of stereotypes and “an ambient sense of belonging” on career choices might have important lessons to teach us about how we might persuade more young men to enter fields they traditionally avoid, such as nursing, education, early childcare, and yes, even English and creative writing.