Hard to, but not necessarily impossible to experimentally prove or disprove. You could for example live one month where you decided everything by throwing dice, another where you deliberated every decision very carefully, and a third month where you picked the first possibility that occurred to you whenever there was a decision to be made. Then you could look at your life, and try to determine which of those months were most successful according to some criteria you set up. (Get bills paid, see friends, go to work and so on.) IF the "deliberately-month" turns out to be the most successful according to your criteria, then there is a strong case for free will, since you by "using" it obtained what you wanted. Of course all this comes down to *choosing* to do this experiment, but I don't think that this invalidates the argument.

http://www.5fingre.com/5fingre/?p=40