Which statistics in which study? Given the current system any sampling from college/university would be cherry picking vs general populous (unless you also sample general population with similar constraints to ensure a like for like comparison) so can't really be trusted.