But that's exactly what i explained above. It's not that also using women's voices would've been harder, it's that doing so would've been a worse study design, as explained in my initial comment and in the quote from the paper that you copied here.

A good study narrows down factors to as few as possible.