How would you scale down an organic chemistry experiment? Would you have to know an expected percent yield to scale it down?

To give a definitive answer we need more information. However, as a rule of thumb, I just scale them down according to what I want. Since most organic reactions are not at 100%, you might want to include a factor (like 65% or so) for the yield. You are right, however, that the knowing the yield will help in scaling down IF you are looking for a better outcome.