In his 1942 article «Miracles,» C.S. Lewis notes the problem of relying on ‘experience’:
If the end of the world appeared in all the literal trappings of the Apocalypse, if the modern materialist saw with his own eyes the heavens rolled up and the great white throne appearing, if he had the sensation of being himself hurled into the Lake of Fire, he would continue forever, in that lake itself, to regard his experience as an illusion and to find the explanation of it in psycho-analysis, or cerebral pathology. Experience by itself proves nothing. If a man doubts whether he is dreaming or waking, no experiment can solve this doubt, since every experiment may itself be part of the dream. Experience proves this, or that, or nothing, according to the preconceptions we bring to it.
Although Lewis is (to some degree) using hyperbole, he has a very good point. When we experience things, we always interpret it through our preconceptions. Let me give an example. When discussing with a friend some time ago, the discussion went into materialism and theism. My friend, an agnostic, pointed to an experiment where a number of people where asked to rest their hand on a table and randomly, within one minute, lift one of their fingers. The interesting thing was that a split second before the people lifted their fingers, the brain told the nerves and muscles to get ready. My friend said that this was ‘proof’ that we have no free will, and that everything is predetermined. But how can this be a ‘proof’? This tells us nothing one way or the other, and it can be read as support of naturalism, of some kind on non-theistic non-naturalism, of theism, etc.
To take the ‘extreme positions’ (naturalism and theism), this experience can be read as a support of the theory that the brain is predetermined and that when we think that we rule ourselves, we really don’t. But it can also be read as a support for the theory that the soul, which is incorporeal, but which is connected to the body (and, of course, the brain) as its form, is the ‘determining’ (or, rather, ‘self-determining’) factor. Which means that when we think that we rule ourself, we really do.
The question, then, is not what experience tells us – even though all reasoning starts with the senses – but which precondition is true, or the most probable. And that cannot be ascertained through ‘experimentation.’
 The experiments in question are Benjamin Libet’s free will experiments.