Originally Posted by binip
There are differences amongst the highest performers, which are consistent. Believe me, it bugs me more than anyone because I was that girl--the one in all those high math classes--who couldn't believe how nonchalantly and confidently the boys took the tests. I'm sure their confidence (aloofness? unemotionality as teens? I don't know) affected their scores positively, and my insecurity, my scores negatively. I hate that it's a trend, because I'd like to think it's just my insecurity. But if you look it up, the top 5% or 2%, you will see that it is very consistently that the boys outscore the girls, and that differences between LA and math are greater, the further right you go on the bell curve.
Some may say it is not wise to declare a test "biased" because some participants come to it with personal insecurity. Effective strategies may include coaching for growth mindset and a sense that one is fulfilling their own potential as an individual, not competing with other test takers or representing their gender by their performance.

Here is an interesting article regarding test performance, however the results are NOT by gender (which may provide a worthy follow up research study): APA - 2007 - Seeing red impairs test performance. I'd also be curious to know how they controlled for other variables (breakfast, sleep, etc).