Originally Posted by
Limbo
Robin, regarding the paper by Dean Radin in your OP. I was wondering if you could highlight the specific flaws in this section:
OK, well first and foremost, he is reducing the sample that he is using to test for the effect.
The anticipation effect would come from the cumulative effect of a number of small outliers distributed across
the entire sample on the aggregated data, there is no reason to suppose this effect would be stronger in some rather than others.
So all Radin is doing is reducing the statistical power for his test of anticipation. In my simulations I found 450 trials threw up all sorts of conflicting results.
He is ignoring the runs of "emotional" pictures which can also add to the effect.
He is looking at 13 calm trials prior to an emotional trial. How many runs of 13 calm photos would there be in 450 randomly selected photos, even at a 1:3 ratio?
In any case he does find the positive linear trend that would indicate the presence of an anticipation effect. The low co-efficient would be due to noise in the tests - the same is true of my simulations, since I add in noise.
I am not sure what he means by the next part. If there were a positive linear trend then the mean PSCL values relative to the local baseline would be positive.
All in all he seems to be going out of his way to avoid finding an anticipation effect in his data, which could be easily detected by various analyses on the entire dataset of each experiment.
[edit]And note that in his figure 8, the observations closer to the emotional event (which as he points out will be more representative and have a smaller error) has a quite strongly positive trend, I would suggest this pretty much supports anticipation effect as being the culprit[/edit]