Of course, the easy way to find the answer to the OP's question would be for the AB to publish their statistics. I often wonder about how much analysis of marks there is - a certain amount is released (such as numbers taking each grade and passing etc.), a certain amount is mentioned (such as checking individual examiners' averages), but I suspect that with such a large number of entries, and so much data in numerical form, it would be easy to extrapolate extremely detailed statistics on all sorts of things.
In the aural side alone, you could see if particular parts of the aural tests are performed better by girls - but you could also see if keyboard players find other bits easier ('Listening to Music with Understanding' at Grades 6 to 8?), or if certain instruments achieve higher or lower marks than others for scales, sight reading, etc.
Now, do you think AB look at and analyse that data (and yes, I am treating 'data' as a singular)? If they do, wouldn't it be interesting to know? The dilemma would be (and why such statistics wouldn't be released) if it demonstrated that, yes, girls are better than boys at the singing parts of the so-called 'aural' tests. Would it indicate just that, erm, girls are better (
on average) than boys at the singing bits of exams, or that the tests are flawed?
And we only have to think of the example of GCSE maths to show how such statistics can influence exams: with the
elimination of coursework from GCSE maths, all of a sudden the exam statistics show that boys have outperformed girls. Does this mean that, erm, boys are better than girls at maths?
I still can't see for the life of me why singing is a compulsory part of instrumental exams. And before anyone says "because it's helpful to make you a better musician", yes, of course it is, but so too is an ability to dance all the regular dances that we play. But I don't hear anyone clamouring to include dancing in practical music exams. Of course we can test people's aural skills without making them sing. And it would eliminate the possibility of gender bias if we did.