This is one of my favorite weeks in DCR. It can be hard and frustrating to figure out how to do it, and to grapple with the results of a sample size calculation (which sometimes demonstrates infeasibility of your prior plan), but at least it is very useful for focusing the mind on how to define a testable hypothesis. Even if you're not enjoying it, I hope you are getting some useful information from the process!
Some complications arise every year. Here are a couple:
1) Some people have questions about whether to use the continuity-corrected version or not for a comparison of proportions. YES, you should generally use the continuity-corrected version - it corrects for some assumptions made about the distribution of a proportion which are not technically correct. So better to use it.
2) Some people have a hypothesis which is really about "non-inferiority" rather than superiority. That is, you want to prove something is similar to something else (similar enough), rather than proving that it is different. You'll have specify the "non-inferiority margin" (HOW similar does it have to be), and some other things. There are calculators for this as well - here is one: https://www.sealedenvelope.com/power/binary-noninferior/. If you think you have one of these, talk to your section leader.
3) Other complex topics - studies with more than 2 groups, or clustered samples (e.g., a RCT where you randomize clusters rather than individuals). There are notes about these situations in the book, and I encourage you to ask your section leader about these as well.
3) Finally (not related to sample size calculations) - I just want to note that our "flipped classroom" experiment this year does rely on you watching the recorded lectures. This is where the real didactic teaching occurs (along with the book reading). NOTE that you can listen to the lecture at 1.5x speed! It usually works pretty well and will get you through the lecture much faster...
See you next week.