The problem with economists is that they have no way conduct scientific experiment on monetary policies. Yes, there are data, but interpretation is everything.
Economists can have interpretations that fit the data but it turns out to be bollock.
Yes, Economists may not like to hear this, but Economics is at best a proto-science. Beside the practical difficulty of performing controlled experiments with economic policies (as you pointed out), there's also the problem that Economics is not yet grounded on any meaningful substrate in the "great stack of science". Unlike Chemistry, for instance, which is firmly grounded on the Physics substrate.
Nevertheless, my opinion is that when eventually Economics does become a science it is more likely to resemble Keynesianism than the Austrian school so favoured by this forum's hivemind.