05-27-2014, 02:45 PM
(05-21-2014, 02:43 PM)Bolty Wrote: Wow. Merely having a seemingly innocent question asking what your gender is before taking a math test affects women's test scores just by reminding them that they are women, because it makes them conscious of society's pressure on women not to be good at math. This can happen even if they're fully aware that it's bull-honkey. They can swear up and down that they don't believe that for one second, but it still affects them.
Stereotype threat, along with most other priming phenomena, have not largely held up to replication and meta-analysis. It's a neat idea, but it isn't well supported by the evidence at present.
Social science in general, and experimental psychology in particular, is suffering from a pretty bad case of false positives. The misuse of very crude statistical hypothesis testing and "statistical significance" as the threshold for publication has been pernicious. Though it does generate a constant stream of attention-grabbing headlines.
-Jester