Salik: Let's do that! Then for not paired data we just have to give each person the homework once and tell them to either listen to music or not when they're doing the assignment. Maya: But then some people may just do better on the homework because they've studied more or are doing better in the class. I think it would be really hard to know what the effect of the music is instead of just differences between people. Salik: I guess that is a disadvantage of setting it up this way? What is a good thing about doing it this way