by Cate Montana
WTB - So often when reading about psi experiments, the results seem small, but when meta-analysis is applied, the significance seems to soar. For example, the Sherwood and Rose studies of the dream psi experiments at Memoneides resulted in a 59.1% hit rate, and that is a 9.1% increase over the statistical norm of 50% and an odds against chance potential, as you write, rated at 22 billion to one. It kind of boggles the mind that a 9.1 variance- well, that's kind of significant, but when looked at in comparison to a 22 billion to one chance of happening, it puts it into a whole different light, could you explain this?
Radin – Well, it's all devolves back into a very basic concept in statistics: the more data that you have, the more confident you can be in your overall averages. It's that simple. So, the reason why historically, Mickey Mantle is considered an expert baseball player, is because his overall base hit rat was approximately 30%. So the reason why Mickey Mantle is considered one of the best, is because his career, over 18 years gave him about a 30% hit rate.
WTB – Right. Whereas if you have just looked at his batting average in one game…
Radin - If you look at one game, maybe he struck out. So, it's only in the long-term average, where you can get a sense of high confidence about whether somebody is any good. The same goes for somebody like Michael Jordan. You look at his overall statistics, year after year, clearly he's superior to other basketball players, who on any given game might have actually done better than him. So the same is true in experiment where you're dealing with noise and measurement error in an experiment. If you do one experiment and get a really interesting result, than you might be excited. But, someone could validly say, well, maybe it was a fluke. And you have no retort for that, there's no way to respond to the possibility that maybe you were just lucky that day. So you have to run many, many experiments, and then you start building up an overall average for that. And as you build that up, your confidence in what the actual size effect is, begins to get better and better.
So the reason why a 9% effect, as in the dream telepathy study, is monstrously significant is because you have a very high confidence that that 9% is real. It's not just a chance effect. That's where the meta-analysis people [can streamline things] where the goal is to create essentially a very, very large scale experiment that would simply cost too much and require too much time before any one group to do.
WTB - which is why the global consciousness Project at Princeton with the random number generators around the world is so effective? Because it's taking this huge sampling around the world of events that create coherence? Maybe that's not such a good example. Next > 1 2 3 4 5 6 7 8