Are demographics good criteria to screen research participants?