If you set up your experiment the 'right' way, there is a good chance it has.
Here is one way that I think has a good chance of working: have your subjects pay a fairly steep amount for the trial, the evaluate it by having them fill in a questionnaire.
The "pay to take part" selects for gullibility, and also makes the subjects feel stupid if they don't report improvement. The "self-reporting", of course, helps get rid of such things as objective measurements that would spoil the party.
Finally, pick a small test group. That's cheaper, and will give positive results if only a few subjects report actual improvement. And if your first test fails, run a second, this, etc. identical test.
Here is one way that I think has a good chance of working: have your subjects pay a fairly steep amount for the trial, the evaluate it by having them fill in a questionnaire.
The "pay to take part" selects for gullibility, and also makes the subjects feel stupid if they don't report improvement. The "self-reporting", of course, helps get rid of such things as objective measurements that would spoil the party.
Finally, pick a small test group. That's cheaper, and will give positive results if only a few subjects report actual improvement. And if your first test fails, run a second, this, etc. identical test.