The point is that people will discriminate in favour of women, not against.

If you weren't so absurdly politically biased, you would recognize immediately that this statement doesn't follow from the conclusion of the study you linked. People "associating more positive attributes" with women is inherently tied to discrimination against women; it means they're treated like children and not like adults.

People have the same bias towards evaluating children positively too, but we don't allow children to have any responsibility.

As for the first study you linked, there's another study that follows almost an identical methodology and comes to exactly the opposite result. It's a gross distortion to present your single study as the last word of SCIENCE!!!! on gender equality in stem. Here is a good overview of the problem. Personally, I feel that college labs at least strive to be quite gender neutral on the whole (even if sometimes they don't quite succeed), and the conclusions of both studies simply reflect the known political biases of their authors.