Github patches from women who don't reveal their gender more likely to be accepted than patches from identifiable women

[Read the post]

5 Likes

It’s probably just a poorly configured anti-cootie filter.

15 Likes

Maybe boys are just smarter than girls so nyeah! The researchers just didn’t know that because they’re probably girls too! And boys are better at baseball! And writing in the snow!

2 Likes

No surprise over here.

3 Likes

If you don’t implement an anti-cootie filter, it can’t be poorly configured. :wink:

8 Likes

The comments on the Ars Technica piece are the expected shitshow. I should go see what Reddit users are saying (ha ha ha).

8 Likes

Yeah, but also unknown gendered men had higher acceptance rates than known gendered men. This is not a case of gender bias, it is a case of anti-gender bias, bias against gender in general rather than one particular gender. Insider known gender women had a higher acceptance rate than insider known gendered men. Overall, women had a higher acceptance rate than men. The biggest biases seem to be against outsiders and particularly against outsiders of known gender. Github prefers androgynous clique members. edit: Patch accepted.

3 Likes

Survivor bias is probably true. As covered in Preacher:

2 Likes

Those aren’t two different explanations. There is bias, unquestionably. Women who are informed enough to know this and act on it by hiding their sex might well be more likely to write better patches. How many women who’ve been on the net for more than five minutes do not know the pitfalls, bias, and outright abuse that comes with having a female-sounding handle? Go play Call Of Duty signed in as “SoudjaGirl” or somesuch. Within ten seconds you will know why not to do that. If you haven’t been around long enough for that lesson to land, chances are you haven’t been around long enough to contribute patches.

Of course that’s a gross generalization, but we’re talking about statistics here, which are by their nature generalizations.

Anybody notice the error bars on the right bar chart are about half the size of the total data?

1 Like

I took a spin through the /. comments. Same story.

“Sure maybe some female programmers are better than men, but there is a rational explanation as to why men are still better programmers on average” was about the tone.

3 Likes

FTFY
and I’m a middle-sider non-gendered orca

2 Likes

It’s interesting how the “she was asking for it” model of sexism transfers from the real world to the internet. In some cultures, women must cover every square inch of skin because otherwise if anything happened “she was asking for it.” The internet is even worse, having some cultures where revealing a female gender at all is “asking for it.”

As always in these situations, focus your ire against the people and biases that enforce these oppressive cultural standards rather than against the women who challenge them.

2 Likes

I weep for my industry (and humanity).

7 Likes

Buh, wha?

My ire is solidly against the standards that enable and reinforce these precepts. Is your interest in focusing blame (a not unworthy goal) or in interpreting these stats? If job applicants named “Lashonda Jackson” are statistically likely to be discriminated against, doesn’t it make sense that a Lashonda Jackson who is adept and thoughtful would be more likely to change that name than one who hasn’t given that issue a moment’s consideration? The injustice of that prejudice is definitely an problem, but it’s not the one we’re discussing. Especially considering that changing your online handle from “SexyGurl69” to “GenderNeutralNameThatHasntAlreadyBeenTaken” is far less of a task than changing your, y’know, actual name.

I’m sure there are SexyGurl69s who are making a point with their usernames about sexual discrimination, but like I said before, statistics are generalizations. The majority of people who choose a gender-identifiable name aren’t making political statements by doing so.

Point is, people who are generally naĂŻve enough to name themselves SexyGurl or even StudF4rmer are generally going to be naĂŻve enough not to have worthy contributions, codewise.

ETA: This doesn’t mean that discounting the contributions of SexyGurl or StudF4rmer based on their usernames is defensible, but we are talking about statistics, not individual maintainers’ biases.

With all due trepidation about making this point, the article seems to editorialise the study in a way that the authors have contradicted elsewhere:

Their conclusion: when a woman’s gender wasn’t obvious from her Github profile, her patch would be accepted more often than a patch from a man; but when a woman’s gender was clear, her patches were less likely to be accepted by men.

Appears to contradict or at least ignore the author’s statement here:

Our analysis (not in this paper – we’ve cut a lot out to keep it crisp) shows that women are harder on other women than they are on men. Men are harder on other men than they are on women.

It’s not obviously a story about men rejecting PRs from women. Digging through the paper, it seems that ‘insiders’ - people with past contributions - have their PRs accepted more often, and this effect is stronger for women than men. For outsiders, people using non-gendered names have their PRs accepted at a higher rate than people using gendered names, with the effect again being stronger for women (the gap between female non-gendered and female gendered is higher than the gap between male non-gendered and male gendered).

The big difference that nobody mentions in their reporting is that there are far more men than women in the data set, though perhaps this is sadly no longer remarkable.

This is a great study in terms of demonstrating how we can examine these issues using open data sets. The reporting of the study has generally been of a much lower quality.

2 Likes

Study may be kind of BS.

http://slatestarcodex.com/2016/02/12/before-you-get-too-excited-about-that-github-study/

I wondered about this since the error bars in the graph were so large relative to the bars.

1 Like

why is lying about one’s identity the way to go? Why should I, or anyone, have to pretend to be something else to get basic respect? Why is that okay? Why does woman = inferior? Why am I the wrong one for being my gender?

5 Likes

When all genders are revealed among outsiders, men appear to have their requests accepted at a rate of 64%, and women of 63%.

A difference of 1% while interesting may not be worth burning down Github over. Particularly with @codinghorror’s article he posted about the flaws in the study.

Sure, but I was addressing @L_Mariachi’s point that someone named LaShonda Jackson should change their name instead of expecting fair treatment, which is a broader issue than just the discussion about github, yeah?

But they also said this, so perhaps I should read more carefully? Still, I think the point stands in general.

[ETA] Also, regarding the article posted by @codinghorror, stating that basically that insider women have a slight advantage (ie, this statement):

Among insiders, women get more requests accepted than men.

Could this possibly be in part because people are making a conscious choice to be more inclusive towards women? Like, they have listened to discussions on gender in tech, recognized it as a problem, and have sought to think about and be mindful of their behavior in general towards women?

1 Like