Being an autodidact is a way to satisfy an interest quietly and to avoid pressure, in contrast to classes. So if outside pressure hinders participation, it should hinder it less among autodidacts than it does among others. Thus, (# female autodidact coders) / (# female coders) should be higher than (# of male autodidact coders) / (# male coders).
As an example, to learn more than broken Hindi I'll take a class because there is no social pressure against it. But if I want to learn about pickup artistry (which social pressure opposes), I'm going to do it on the internet. (Both examples chosen to be subjects best learned in person.)
Anyone have data on this?
But the way women are treated relative to their male peers is a massive problem within software, as opposed to the broader STEM fields.
This is weak evidence against female treatment being the cause of underrepresentation. Math-heavy STEM fields (physics, math, electrical engineering) also have very few women.
This would seem to be a third explanation for something you said had only two explanations.
I don't think you'd say that the first one doesn't exist, and when you say discrimination starts at childhood, that sounds a lot like the second.
One needs to be careful not to equivocate between the two, because they're different problems that have different causes and different solutions that need to be implemented by different people.
When you describe the two of them together, as a single possible cause, you seem to be making this mistake.
Where are the _____ engineers? As a rule, they're in the huge fucking pile of resumes in the garbage. But that answer is too simple and obvious, it makes the situation too clear. Instead, it has to be a big mystery about how it just magically happens that there isn't a black face in the office even though we are all, of course, very enlightened and not racist here on HN, and all of us are doing our very best, so sorry if you don't have enough experience for any job or are not a cultural fit anywhere.
Just a tip: the use of "begs the question" to which you refer is always intransitive. The transitive use, as by the poster you responded to, is distinct, and unambiguously a different use than the intransitive use. AFAICT, its inspired by the transitive use, which is an odd (from the point of view of the meanings of the individual words in modern English) translation of a Latin phrase into English, and -- unlike the intransitive use -- actually uses the individual words in something like their usual definitions. It also provides a convenient rationalization of the older intransitive use as a specialization of the transitive use where the question for which an answer is demanded is that of the support for the premise under discussion, so that the transitive use turns the intransitive use from an idiom that is detached from the meanings of the words in the language outside of the idiom to one connected to it in a clear way.
Accepting the transitive use, which is widely established and whose meaning is immediately clear to anyone who encounters it and which is completely unambiguously differentiated from the intransitive use, does nothing but make life better for everyone.
Opposing it isn't merely pedantry, its pedantry that ignores clear grammatical distinctions and seeks to find confusion where there is none to be found.
A lack of good candidates is a serious problem (this even more obvious racially than by gender - where are the black and latino engineers?). But the way women are treated relative to their male peers is a massive problem within software, as opposed to the broader STEM fields. Don't believe me? Sit down and have a heart-to-heart talk with experienced female software engineers about it sometime. Ask them what sexism they've experienced at work, how it's hampered their careers and their job satisfaction.