You are looking at math majors at a large University and want to Simply know if students who take both CAL1. and CAL2 are earnings similar Grades in both courses Since students take both math courses you have the data you need for a paired test. But this does not impact how you set up your hypothesis. When you compare means for this course with no expectation that they're even is a difference your null hypothesis would be H0: CAL 1 > CAL 2, H0: CAL 1 not equal CAL 2, H0: CAL 1 = CAL 2, H0: CAL 1 < CAL 2 or H0: CAL 1 < or equal CAL 2