The question said a "random" sample which means that she doesn't know if the curfew of the sample tells her father that other people's curfews are earlier (eg. 8 o'clock), thus persuading her father to change her curfew an hour earlier which is not good for her.
She should organize specific data telling here father that other people have curfews later than 9.
