It seems to me that the key question is whether the presentiment effect varies with the length of the trial. That 2002 paper explains (very clearly) why the bias occurs, and it basically depends on the interaction between the arousal model and the sampling error (at least as I understand it). They have a graph showing how the bias tends to zero as the trial length increases.
Of course, this is made more complicated by the fact that subjects may get bored over longer trials!
Before everyone throws out presentiment as bunk, I think it is important to remember that strange timing effects to do with consciousness are nothing new (Libet). Some of the conventional explanations for that behaviour are pretty way out!
David |