Some researchers wish to compare how students at two different schools perform on an IT test. They randomly select 40 students from each school and obtain their test scores. They pair the first score from school A with the first school from school B, the second score from school A with the second school from school B and so on. They then perform a hypothesis test for matched pairs. Is this approach valid? Justify your answer. If it is not valid, how should the researchers have proceeded?