You are viewing a single comment's thread.

view the rest of the comments →

0
3

[–] guy231 0 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.

0
0

[–] 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.