(You may have an easier time understanding my mindset here if you're familiar with The 12 Virtues Of Rationality and/or The LessWrong Sequence On Doing Hard Things; my goal here is to *solve a problem*.)

So in a (friend locked) post over on my (fairly casual) friend Joe's FB, which I accidentally derailed (sorry about that), a study was posted (full data) which shows that applications for hard-science (physicist) jobs that are completely identical except for the name on the application, are rated lower in various ways, including pay that's likely to be offered, if the name is female(-sounding).

I commented that since this study doesn't give us any new information at all about *why* that systematic bias exists, it wasn't interesting to me (phrasing it that way is a deliberate cop-out, which I will correct later). I was then badly misunderstood, I think, and I want to try to explain my views in more detail.

The tl;dr is this: there is a *problem* here, and we *HAVE NOT SOLVED IT*. If we stop at a study like this and go "Oh, it's just that we're a pack of sexist assholes", and pat each other on the back for a job well done, even if that is actually *true* (which this study most emphatically does not prove) we *STILL* haven't *solved* the problem. More on that later. I want to *solve the problem*. I'm not sure what all the people yelling at me want? I guess to be right about how I'm an asshole? But if they've got a solution to the problem, they certainly weren't willing to tell me about it.

To get anything out of this, you're going to have to take my word for some things about me, or ask my friends. I am a highly skilled computer professional (linux systems administrator). I am *extremely* self aware, and quite capable of changing my beliefs and my behaviour in the face of feedback from those around me. I care *deeply* about women, both in the specific (my life is entirely wrapped up around my girlfriends and daughters) and in general. If you can't find a way to believe me about those things, or at least accept that I might be telling the truth, you might as well stop reading right now.

I was, as with most privleged white males of my generation with smart, wealthy, liberal parents, raised to believe that men and women were exactly equal in every way and that to think otherwise was, essentially, a sin. I remember being *shocked* the first time I encountered girls acting in an extremely stereotypically girly fashion (we were lost in the woods and they fell apart), and I felt guilty for *years* afterwards for even *noticing* that this had occured. So my experience with women in my field has been horrifying, guilt-inducing, and depressing.

First, let me say that every time I have encountered women in technical positions, I have gone out of my way to do everything I know how to do to help them feel welcome (and no, that does *not* include hitting on them or telling them how pretty they look at the time; I have actively sought feedback from the women I know on how to do this properly), to teach them everything I could to help them do their jobs, to listen to them in whatever they wanted to teach me, etc, etc. The feedback I've gotten from every woman I've worked with closely has been extremely positive, up to and including asking my boss to have me specifically be the liason to the woman in question rather than any other person on my team (this has happened at least twice).

As best I can remember, and speaking her only about actual programmers or sysadmins or direct support of same, I have worked with perhaps as many as 8 women over my 15+ year carreer. This *horrifies* and *disgusts* me, but that's not the bad part, not by a long shot.

Of those women, 3 were technical managers, and very very good at their jobs. 2 out of those 3 I am certain would fail FizzBuzz (an extremely simple programming test), and the 3rd I don't know one way or the other. So they don't really count for purposes of this discussion. I'm also going to not consider technical writers, as that would make the results even worse and seems unfair. I am also excluding my two current partners.

Of the rest, I can only remember 1 female sysadmin that I considered technically competent such that I noticed, i.e. that she could easily keep up with everything I said and I thought I could maybe learn some things from her (that last is exceptional regardless; I am *extremely* good at my job). I only knew her in passing for a few weeks, because of the nature of the job, so I can't be sure, but that's the impression I was left with. In addition, I once saw a presentation at a conference (on AFS, FWIW) by a woman who seemed obviously competent as a sysadmin.

I can remember at least 2 who were, as programmers, obviously incompetent. By this I mean primarily that explaining what I wanted, or teaching them how to do it, took *far* longer than I'm used to when dealing with *mediocre* programmers. They had obvious trouble understanding me. I checked with others (including other women if possible), and it wasn't just me. Basically, I wouldn't hire them for the jobs they were getting paid to do, given the choice.

So, of the *tiny* number of women I've known in the highly-technical end of my field, about a quarter have been, by my standards, palpably incompetent (and if you don't accept the standards of someone who has been doing this as long as I have, who *would* you accept?). I've kept a mental list of the obviously incompetent people I've worked with, and there are perhaps 4 men on the list. I assure you that I've worked with a *LOT* more than 16 men.

Again: this horrifies and scares me.

The experiences of other sysadmins I've talked to have been largely similar. I actually just checked with a couple of younger-than-me friends, and they seem to be seeing something a bit better; they can name two technical women they've ever worked with, and the best they can agree on is that they're not obviously incompetent (they work at the same place). So, much better proportion, but not exactly a great sample size.

Even if you assert that I am somehow seeing nothing but statistical outliers (i.e. that there are plenty of competent female programmers and sysadmins out there, just none anywhere near me ever), I've never even *heard of* a 10x programmer who was a woman. And I'm pretty sure that such a person would be legendary. I dunno, maybe Vi Hart programs? She's totes awesome.

Ah, lovely; in response to a side discussion about this essay, one of the younger friends I mentioned above mentioned Allison Randal, who looks like she probably *is* a 10x programmer, which makes me extremely fucking happy and is in fact improving my whole afternoon, but unfortunately *still doesn't solve the problem*, so moving on.

So I find myself, despite trying *really* hard, responding to women I meet in technical contexts with a big beaming smile and a "Welcome, it's so nice to have you on board! Where did you work before this?" or whatever friendly smalltalk, but inside I'm going "Oh fuck this is going to be *so* frustrating".

Whaaaaat? That's not *me*. That's not how I relate to the world. I'm a kind, considerate person (with a bad habit of being a bit of a dick online, but I'm working on it, and have been for many years), I don't prejudge people, I *certainly* don't think that way about women, what the *fuck*?

The problem is, it's basically a correct response given the experiences I've actually had in my real, actual life as an actual professional in my field.

So, what actually happened in Joe's thread, this time without the cop-out I mentioned, is that I said something like "OK, from this study you're concluding that there's systematic *sexism*, but the study doesn't demonstrate that, it demonstrates that there's a systematic *bias*. What if women are actually, on the whole, worse at physics, and all the people involved are behaving rationally given that reality?".

People took this as meaning "lol let's keep the womenz down cuz they suck!", but that was *not* my point. It was a *question*. It was *NOT* a *rhetorical* question.

I really want to know!

Something is going on here, something very very wrong, and we *don't* know what it is. Certainly not from studies like *that*. Maybe there's a bunch of studies I don't know about that *do* explain the problem and how to fix it; that wolud be fantastic, tell me about them!

As far as I know, all we do know is that in early grade school there is no obivous difference between girls and boys in the subjects we teach them, and yet somehow there are almost no women programmers and essentially zero exceptionally good ones. We also know that women get paid less in technical fields. Maybe there's a connection, maybe there isn't. Maybe the entire issue really *is* simple sexism.

But you know what? If that's the case, *fix it*. Show me some school that turns out great female programmers. Show me that home-schooled children of geeks have no obvious biases in which careers they choose. If we already know what the problem is, then we should know how to fix it. At the very least, you should be able to point me to groups of people that are trying to fix what you say is the problem, and we should see measurable differences when we look at them.

I'm not aware of any such study or group or system or curriculum.

The closest thing I've ever heard of is Medina's collection of research (see Brain Rules for citations) that seems to say that the way to fix the problem with girls falling behind in math and science in school is to have sex segregation. Well, there are still sex segregated private schools out there; show me the data, please.

I want to solve this problem. I want *desperately* to solve this problem. Because if we can't solve this problem, some day one of my daughters is going to look at me and say "Robin, I want to learn computers like you!", and I'm going to know in my heart that unless I'm *extremely* lucky that interest will fade by the time she's 12 or so, and I can barely even bear to *imagine* that, let alone experience it.

Help me help my daughters. Tell me how to fix this. Tell me with actual data in the actual real world.

As far as I can tell, the only value that anyone got out of the original study that Joe posted was being able to self-congratulate about how sexist everybody is. That *doesn't solve the problem*. You're not *helping*, you're just *giving up*.

I will change anything I need to change about my life, my attitudes, my beliefs, whatever is needed (ask my therapist, Myles Downes, if you think that's an idle boast; I can do this, and I will) if you just tell me *what is actually going to help*. But you don't know. Because these studies aren't asking the right questions. At all. And it's making me very, very sad.

Answering some specific objections from the thread over at Joe's FB (paraphrased):

1. "We already know there's no aptitude difference". We do? The only research I've seen shows that we know there's no aptitude difference *in early school age children*. If we really know this is true, show me a college, or even high school, graduating class in which the bull curve, in all subjects, ignores sex differences, please.

2. "The study compared people with identical qualifications". So what? I've met people with fantastic qualifications that couldn't code their way out of a wet paper bag. Just because two people both have a CS degree from CMU does *not* mean they are identical clones. It doesn't mean they're identically competent in the workforce. Hell, I'd be surprised if there was even a very strong correlation; school is not like work. So, how strongly do the qualifications this study used correlate with long-term job competence? What happens if you control for sex? Again, I'm not saying I know the answer, I'm saying I *DO NOT* know the answer, and I desperately want to, because until we know the answer we can't solve the problem.

3. This one is not paraphrased: "It makes the entirely credible assumption that candidates with identical credentials will have identical qualifications (more precisely, that gender is not predictive of the relative qualifications of two candidates with identical credentials). I think that's the proper default assumption, and someone who wishes to challenge it should have the burden of proving that it's incorrect.". Well, I can't speak to physics, but in my field, I've not observed that assumption to hold. I bow out of running the study in question; scientific studies are not, in any respect, my field. I would be *more* than happy to throw money at someone who *is* studying this, though.

ADDENDUM:

A friend pointed me to a German survey, in which 46% of German women from 14-29 years old claimed to know at least one programming language. Even if it's crap in terms of actual competence (i.e. "can you pass fizzbuzz?"), that *claim* is so starkly different from what I'd expect to see in the US that it gives me real hope that at least *some* of the problem is both social and solvable. I'm smiling right now as I type this, in fact.