ConclusionsAt this point, I know someone will ask: "what could computer science departments do differently?" Well, more involved advising and mentoring would be a great start! That is, it isn't enough for an advisor to just say "come by if there's anything you need", because if you have impostor syndrome, you may not know what you need and you certainly won't want to admit that you need help. What if departments expected advisors to be ready to support all grad students, not just the ones who look exactly like themselves? This isn't to say that every faculty member can or should try to be an expert on every identity, but knowing what they do and don't know would be a start. Any outright acknowledgment of impostor syndrome would be a great start too. At Berkeley, there was nobody who stood up and said that most of the time when people look like they know what they're doing, they don't. I'm not sure I would have believed it even if they'd said it. Oh, sure, other people might be fumbling, but not as fumbling as me. We did have a required class on teaching techniques at Berkeley, since all grad students were required to TA for at least one semester -- in my head, I called the class "Geek Support Group", but it was actually really helpful because it was one time during the day when we got to put aside the pretense that we were all rational beings made of pure logic. So maybe a required class on how to be a grad student would have been helpful (required because I suspect the very people who needed it the most would have brushed it off if it was optional.)
Encouraging socialization in a way that includes everyone would also be helpful. Of course, most departments already have social events. In my department at Berkeley, when I was there, the CS grad students' group organized a weekly reception. However, faculty members rarely attended; the professor who I saw there most frequently seemed to stay just long enough to snag some free food. I was part of the CS grad students' group at Portland State, and over time, students stopped attending our events, even when we offered free food; it's not clear why. In contrast, in my ex-partner's department at Berkeley -- mathematics -- the department had a tea/coffee hour every afternoon, which a department assistant organized (the job wasn't pushed onto students) and was very well-attended by both students and faculty. Just having social events is not a be-all and end-all, since some students won't feel comfortable in large groups and some people always get left out, but it's a start. Of course, offering free food can help, and provides an excuse to go for someone who is reluctant to socialize.
Some people have proposed that single-sex or single-gender education might help with impostor syndrome (at least when experienced by women). It's not clear where people like me who were assigned the wrong sex at birth, or people who don't fit into the gender/sex binary, fit into such a plan. And even though I went through my own supposedly-women-only undergrad education mostly thinking I was a woman, that education didn't really have the effect of immunizing me against impostor syndrome. Maybe it was because I wasn't actually a woman. I suspect, though, that there are cis women who also had the experience of not really fitting in or being socially accepted at a women's college. I wonder whether single-sex education favors those who conform the best to the gender expectations placed on them.
I already mentioned that I think the concept of "meritocracy" is a lie and a cover for inequality, I must say that if you do believe in meritocracy -- at least as a goal to aspire to -- you should realize that meritocracy doesn't work if some people get systematically denied awareness of their own merit, while others are encouraged to have an inflated self-perception of their merit. What may look like "self-sabotage" from the outside may well be the result of structural oppression: as much as you might want to blame somebody for not going to the study group, or not getting to know professors outside of classes, or anything else that appears to be a choice, blaming doesn't accomplish as much as asking what their reasons might be for not feeling like a certain course of action would be safe.
If you're a professor or teacher and want to know how you can help students with impostor syndrome on an individual rather than structural basis, I have a few ideas. First, consider changing the environment and not just changing the person. I hope I've conveyed in this essay that impostor syndrome is at least partly a variation of Dysfunctional Environment Syndrome (a hypothetical diagnosis that might well subsume most of the contents of the DSM-V). The biggest way in which I learned to let my extroversion manifest itself and got to be able to talk to people was using text-based communication. The relationships with people developed online are what I give the most credit to as far as being able to interact with people face-to-face now. So consider how you can adapt the environment to the person and not always force people to conform to their environment.
Second, show your humanity. Talk about the history of ideas when you can fit it in, and the mistakes people made along the way. One of the things that intensified my impostor syndrome was seeing ideas presented in their perfect, completed form and assuming that they sprung that way from the heads of their creators -- I couldn't imagine myself as anything like the mathematicians or computer scientists who created what I was studying. If I'd known earlier on how much frustration and false starts every one of those people actually experienced, my life might have been different. Talk about your own past and times when you had a hard time learning something. Talk about times when you procrastinate and get distracted and can't take criticism. Talk about that seminar talk you couldn't comprehend in the slightest. Some of the moments that had the most positive effect on me were times when professors and job supervisors talked about these things. When you get shut out socially, you see anyone above you (even older grad students) as god-like because you can't see how anyone gets from where you are to where they are. Even the slightest hint of humanity from an authority figure can make a lot of difference to someone.
Third, don't assume students will be proactive. If someone takes the initiative to come to you for help without nudging, they probably don't need the help that much. In college, I hardly ever attended professors' office hours. When professors (mostly in first-year courses), required meetings a few times per semester with each student, though, of course I would show up. It wasn't even a fear-inducing thing, either, it was just, well, we had an appointment, so of course I had a reason to be there. If I thought about just going to office hours of my own accord, I worried about not knowing what to say when I got there or not being able to articulate a question about whatever it was I was having difficulty with (the latter is still hard for me even now). At least based on my experience, creating structure is a way of including people, and leaving things unstructured is a recipe for leaving people behind. You could argue that requiring meetings is coercive, but I wouldn't argue that, since part of attending college is submitting to the discretion of people who (hopefully) know more than you.
In my opinion, one thing that is not the answer is trying to change curricula to be more appealing to "women's interests", whatever those are meant to be. This is something that seems to have happened mostly in an undergraduate context, and is mentioned in the book Unlocking the Clubhouse. This book, and others, have claimed that women in CS are more interested in applications -- "people stuff" as opposed to theory, which is more of a boy thing. This argument was debunked, though: the claims were based on the authors' experiences at Carnegie Mellon, but a later study at CMU showed that once the CS undergrads were more gender-balanced, women students were both more interested in and better at theory than the men were, on the average. (This claim is from a talk by Orit Hazzan that I attended -- I can try to dig up the citation if you're really interested.) So I think that "women are more interested in applications, men in theory" is such an appealing line (which still gets repeated despite being untrue) because it deflects attention from the ways in which women actively get driven out of the field. You don't have to talk about rape and harassment at technical conferences if you can convince yourself (and as many other people as possible) that women's interests just aren't as pure as men's interests, not as centered in truth, beauty and freedom. I know plenty of women who are interested in theoretical computer science for its own sake, actually, and when I was an undergrad, I didn't know a lot of other CS majors who wanted to code to save the world -- in fact, they seemed to be overwhelmingly interested in CS for the fun of it, in puzzles and problem-solving for the sheer joy of solving problems. I don't think there's anything inherently masculine about coding for fun. Claiming ownership over fun is just another way in which men keep women out.
This is all supposing that departments really do want to attract as many talented people as possible, as opposed to prioritizing the maintenance of a white-male-dominated club. My personal experience tells me that many people in academia -- especially the ones who say they aren't sexist or racist (as opposed to, well, doing anything anti-sexist or anti-racist) -- like being in white-male-dominated clubs. So I'm not too optimistic about things changing, but I've written down my thoughts in case I get proven wrong. It's possible that not all departments are as bad as the ones I've been in, but given my experiences, it's hard for me to see beyond the wall of well-meaning white men who say they really want to change things but don't want to spend even a minute working to do so. At least in my experience, you're not likely to find an actively racist or misogynist faculty or staff member in academia, but you will find a lot of people who just don't care, which of course is exactly the definition of letting racism and sexism continue (and therefore being racist and sexist).
In retrospect, I can see that the process by which departments admit students from underrepresented groups, then allow those students to self-select out of the program by not doing anything to address impostor syndrome, reflects the deeply sexist and racist nature of academia. I can also see that while affirmative action in admissions is a step in the right direction, it doesn't address sexism and racism at their roots. But that doesn't blunt the painful memories of feeling I'd been told that my intellectual ability (the only thing I'd valued in myself) wasn't valid. Perhaps there are some positive aspects to impostor syndrome: at least once I recognized it, maybe I'm more empathetic as a result, less willing to judge other people and write them off as being "stupid." (Though I still have to check myself to not lapse into habits like laughing at people for bad spelling and grammar, habits that were founded in insecurity and a need to make the things I was good at seem more important.) If I ever take a teaching job, maybe I'll be more sympathetic to students who have insecurities and fears around learning than would a teacher who had always been a star student and never faced setbacks. Still, I think that universities and graduate programs need to stop burying their collective head in the sand, stop blaming the earlier stages of the pipeline and do something to make it possible for their marginalized students who are there and interested right now to fully develop their talents.
I'll say it one more time: if you want to take steps towards making it better, a great (and easy -- if you have more money than time) way to do that is to donate to the Ada Initiative. While they're not addressing higher education directly at this time, the work they do will change the culture in ways that have far-reaching effects.