Hello,
I have performed a Pearson correlation on 2 separate samples of about 2000
responses, and in both cases the correlations are around 0.76. I have then
redone the analysis on all 4000 cases and the result seems to be 0.82.
So
sample A = 0.76
sample B = 0.76
A + B together = 0.82
I would have thought that this was impossible, that the result from the
joint sample would be approximately the mean of the 2 subsamples.
But I have checked and rechecked these results, using more than one piece
of software.
Could someone clarify this for me please. I'm a bit confused!
Dewi
|