I think there are a lot of opportunities to use independent schools to test pilot programs in education. It's also possible to use independent educational programs to test new theories or practice. Large scale adoption must come out of serious research and investigation, rather just theoretical speculation.
There are countless programs that run summer sessions or weekend workshops that are built upon non-traditional educational models. I think the larger problem to consider is cost. I work for an educational program that invests over $5000 per student per year; however, most of which is expended on an eight week residential component.
Governments sometimes equate cost effective education with "good" education... sometimes meaning that new theories are tested in the field becuause that's what budgets dictate.
Really, your question hits at the key problem of anything with people involved, from governmental policy, to a mother admonishing her child to try something before saying they don't like it. Predicting how people respond to stimulus is a hard business. It's not quite as impossible as the short form of my answer might indicate, but it is difficult and fraught with all kinds of bias and error. So the best method is to subject something to a test run.
Here in the States for examples, by the time something become required or even suggested by the federal Department of Education, it's been tried in various state and municipal school systems. This gives us a far better idea of how a program is likely to work, if not a garuntee.