Simple Statistics help


#1

I’m trying to do some paired T statstics comparing users brainwaves in Sensory Deprivation vs a control of lying on a couch.

Anyone have any insight on how to compare something like time in Theta?

Thanks!


#2

The null hypothesis will be that condition A = condition B, so you could test that by shuffling the condition labels between the two conditions within subjects. Compare the magnitude of the effect size for the shuffled conditions to the originals to see if there are significant differences.


#3

But how do I extract the magnitudes?


#4

You would do permutation testing here. First randomly assign each trial to condition A or B, and then get the average for condition A - condition B. Repeat this process a thousand or so times to create your null hypothesis distribution. This should have a mean of around 0.
Next evaluate the significance by

zValue = (yourObservedDiff-mean(nullDiffs))/std(nullDiffs)
or
pValue = (# of nullDiffs > yourObservedDiff)/(total # of nullDiffs)