Thousands of school children were the field test for the polio vaccine.
Just what I was saying. Lots of people got polio. You could take a group, give some a vaccine and others a placebo. Then check the polio rates in both groups, knowing that without the vaccine some percentage of them would be expected to catch polio in the next year. Ebola isn't widespread enough to expect a statistically significant percentage of people to get it.
Let's say that the polio infection rate of people who had not yet been infected with polio was 1% per year. If the placebo group had a 1% rate of infection in the next year while the vaccinated group's rate was 0.05% you could say that the polio vaccine was 95% effective (with some confidence interval based on the number of people in the test). If the ebola infection rate is 0.001% it is much more difficult to tell if a vaccine is effective unless you intentionally infect people after they've gotten the vaccine. And if you do this you better be @($@* sure that the vaccine is effective before you use it.
Another way of doing it is testing it on groups more likely to catch it (like doctors). You could also do a test where the control group is from before the vaccine was invented, but that runs into more problems of external effects like better infection control being used at the same time the vaccine is tested, so you don't know whether a lower infection rate is from the vaccine or from the changed handling procedures.