I find it entirely believable that women would have a higher failure rate. A top employer such as QF can easily reject 95% of male applicants and only take the top percentile however because the number of female applicants is much lower, the number of females with top percentile scores would also be much lower. The females with top percentile scores would be just as good as the males with top percentile scores but far fewer in number therefore lower scoring females need to be taken on to make up the numbers.
On average then, the female cohort would be of a lower standard than the male one as it would be comprised of a mixture of top scorers and lower scores bought in to make up the quota, where as the male one would be composed of only high scorers.
This applies to the hiring of any minority group, if QF was found not to employ enough male pilots with red hair and had to increase their numbers then the ginger tops would have a higher fail rate.