Selecting controls is not selecting "normals": design and analysis issues for studying the etiology of polycystic ovary syndrome.

Selecting controls is not selecting "normals": design and analysis issues for studying the etiology of polycystic ovary syndrome.