Here is an article that discusses why these limitations might "introduce biases whose effects are hard to estimate" https://t.co/9P8E4i5BS2
3,349 followers
5,053 followers
@aarontay This is an example of the many challenging issues associated with bibliometric gender statistics https://t.co/ggv05mE3Wl
24,624 followers
This is a paper about how nobody should be building any models that try to infer the gender of a writer from their name or the stuff they wrote. Just not an area of research that should exist, b/c gender is an inner felt thing that you cannot infer. https:
2,377 followers
Just discovered a very thoughtful article about issues with gender inference methods, e.g. via names or images: https://t.co/fQJHPBooG8 I have used such methods before and while this ofc depends on application specifics, I'm unsure whether pro's in general