Wikipedia is not an experiment - the scientific method does not apply.
This is a bizzarre statement. Wikipedia may not "be" an experiment, but it can certainly conduct experiments to find out the best way of operating.
If Wikipedia were a type of science then experiments would be the only way to find out the best way of operating. However, we are not a type of science, our aim is to operate as well as possible, not just to find out what the best way is. That means doing the experiment could be extremely damaging - this isn't the kind of proposal that can be tested on a test wiki, it must be tested on the real thing, and it could be horribly wrong. In situations were it isn't practical to test a proposition, we have to use logical arguments to work out which is best without actually doing it. It's not ideal, but in some cases it's the only way of finding out without the risk of it going wrong - I feel this is one of those cases.