What is a simple rule to decide whether to do a chi square test or a t test in the sums given by a biostats teacher?