A belief is not the same as a hypothesis. You can't test beliefs. For example - you might 'believe' that Santa delivers everyone's presents on christmas eve. But you can't test that in science without first theorising how he might physically deliver several hundred million presents to several hundred million locations around the globe in only a matter of a few hours. Alternatively, you might 'believe' in the resurrection - but you can't test it scientifically without first developing a theory of how someone might return from the dead. It's not the belief that's being tested - it's the theory of the known or unknown physical events which might explain it. Let's leave 'belief' out of it?
You're jumping ahead of yourself, you don't need to start building convoluted models straight off the bat. It's perfectly reasonable to establish whether something works before you start thinking about how it might work. In your Santa example you could try and figure out how he might do it, the magic reindeer model or whatever, and come to a conclusion about how likely or unlikely it is, but that still doesn't tell you if he comes at all. He may well do it by some as yet unknown mechanism. In the first instance, all you've got to do is sit in a kid's bedroom on Christmas Eve and observe what happens. Then you can start building a model that you think explains your observations.
In the paper experiment that I'm attempting to construct around these bracelets, one plausible outcome is even that the subject's belief that they work is both the result, and the explanation for the result. Belief is part of what we're investigating.
