Fascinating. What was the empirical measure for 'greater natural melody' in women's voices?
Without seeing the research, my guess would be that the've determined that hearing imaginary voices resembles hearing male voices more than it does hearing female voices, and then either they or the journalist made the reasons up out of fantasy and whole cloth.
They could just has easily have concluded that male voices are more similar as they carry more authority, had they felt like throwing the bias the other direction.