The last one is always so. "People will elect Trump by a narrow margin based on a sample of 10000 people with a % error of 2%" is a typical statement.
A probability experiment will always count the number of times an event occurs. That's the whole point of an experimental probability. How many people respond favorably to a new Cancer Treatment is a typical question asked by someone conducting an experiment in probability.
C is kind of iffy. Perhaps it is too expensive, or it takes place over too long a period. Ideally repeat trials should take place whenever possible. Bone Marrow Transplants are very expensive because of the strict isolation conditions. You're likely only to get one study done this, especially initially. To check to see if Cancer is a generational thing would be an example of a study that takes too long to repeat many times.
B is simply not true. They are not the same. What theoretical model would you use to mimic a survey? You actually have to go do the survey.
A is wrong. It is the other way around.
Answer C is a maybe but check it. D and E are both true.
