Hacker News Reader

Doesn’t that mean that 79% of the difference is not due to bias? So boys would still be behind girls even if the bias didn’t exist. Or am I misunderstanding this?

View on HN

Tables 1-6 show you the actual results on the page labeled 40 (PDF page 42) and the page prior shows actual numbers.

In blinded effects, boys overperform in math and underperform in literacy. In non-blinded effects, boys underperform everywhere.

I wonder how this plays with the experiments that found that that girls tended to score much closer to boys on math tests when primed with the idea that they would do better and worse when primed with the idea that they would do worse.

This is called stereotype threat and there is clear publication bias in the literature. It's not replicating well in pre registered trials: https://replicationindex.com/tag/stereotype-threat-and-women...

Reply on news.ycombinator.com