[–] onegin 0 points 8 points 8 points (+8|-0) ago
Can the world just stop paying attention to social sciences? It is impossible to conduct a meaningful study, and every report on one ends up reading as a biased editorial that ends up cherry-picking the results by pointing out potential flaws only in aspects of the study which go against their desired narrative.
[–] guy231 0 points 3 points 3 points (+3|-0) ago
Lol, article manages to conclude no-bias, despite the results, by honing in on a subset of a subset of the data that might have suggested no-bias if they had been statistically significant. The article's bias is shown when they try to assert a predetermined conclusion using the sharpshooter fallacy, and still fails.
[–] ForgotMyName ago
They choose a subset that doesn't care about their anonymity, assume they're being honest about their gender (and that the profile is even real, lol), take that set, then somehow sample it to find "male"/"female" usernames? I don't see any mention of instances where their github name's gender is the opposite of their G+ gender (which people do all the time). The whole thing is a giant mess.
[–] sakuramboo 1 point 3 points 4 points (+4|-1) ago
Considering that usernames can be anything, photo's are not needed, how are they determining who are male and female?
Are they just basing it off what people put for their accounts? How are they validating the authenticity of the sex of the account owner?
[–] ForgotMyName 0 points 1 point 1 point (+1|-0) ago
It's based on this:
Specifically, we extract users’ email addresses from GHTorrent, look up that email address on the Google+ social network, then, if that user has a profile, extract gender information from these users’ profiles. Out of 4,037,953 GitHub user profiles with email addresses, we were able to identify 1,426,121 (35.3%) of them as men or women through their public Google+ profiles. We are the first to use this technique, to our knowledge.
Though I think it's fair to say that its accuracy is still debatable. They've also now got a strong sample bias toward people that clearly don't care about their anonymity online. Also, even though they determined that the user was female, that provides no proof that users reviewing their PRs have any idea that they're female/male.
I have no idea why you would use your real name/picture/email on GitHub. Use some random screen name and some random avatar (that's what I do). No one on there should care about whether I'm a guy, a girl, or a really smart dog. Code is code. Get good or gtfo, no one cares about your gender.
[–] noblefool 0 points 1 point 1 point (+1|-0) ago (edited ago)
This article is HORRIFICALLY disingenuous . . . For one, the male stats are 15-30x larger in scale of magnitude than the female stats. Which means that if you look at things on terms of raw percentages, assuming relatively similar competence, of COURSE the smaller subset is going to have better percentages. There isn't as many events in total. I mean, it's like looking at one success data point vs. a split of fifty successes and fifty failures, and going "Oh, well subset 1 is 50% better than subset 2!" That's completely absurd, you can't compare them 1:1 because what if subset 1 fails the next two data points? Effectively what the article writer is doing here is taking 5 data points vs. 100, and going "Women are better." (I mean this in a literal sense too, men have 20x more raw data in the first table they provide)
The article then goes on to draw a bunch of averaging conclusions from the dataset (Which, once again, when your sample size is smaller assuming similar competence you can't draw any conclusions from). All of this is prefaced with a (possibly a pair of) boldfaced lie(s) - Women have to prove themselves more often and earn lower salaries. The second assertion (the wage gap) is disprovable by the fact that, once again, there exists no company that's just made up completely of women and is crushing the competition economically on the basis of said wage gap. I can't disprove the first assertion because sample size, but in my experience the 8 girls I took com sci with never had to prove their chops any more than any of the guys did. We did not have like a "Women only" screening test for whether they could write the part of the code project we assigned to them. We only cared whether people could write code.
[–] EarlPoncho 0 points 15 points 15 points (+15|-0) ago
not surprising. if you are a woman on the internet the betas will come. most people writing code are beta as hell and can't believe a woman would interact with them