This is easy to falsify by having someone else evaluate whether a musician's performance improves with practice. This is very easy to double-blind: just record

This is easy to falsify by having someone else evaluate whether a musician's performance improves with practice. This is very easy to double-blind: just record "before" and "after" performances and don't tell the evaluator which is which. I would be willing to bet that the effects of practice on a musical performance are very easily measurable.