Comparability of selection instruments in studies of heterosexual-social problem behaviors