Robert J. MacG. Dawson wrote: > KR wrote: >> >> Hi, >> >> If I have two different surveys from which I calculate values of P1 >> and P2, would Cov[P1, P2]=0? The two surveys were performed in >> different years and have different sample sizes. > > You cannot have a covariance without a pairing between the elements > of the two sets of observations; covariance is a measure of whether > bigger-than-average measurements in one set tend to be paired with > bigger-than-average measurements in the other. > > You cannot have a pairing with different sample sizes.
While the context is totally unclear, the immediately above is not right (incomplete pairing is enough to lead to correlation)... if the second sample were an expanded version of the second, with the people (items?) in the first carried over to the second, then there would be enough pairing to lead to correlation. Of course, if both samples were random samples, selected independently, then there would be no correlation even if some people (items?) appear in both samples provided that one is working with the probability space generated by the sampling mechanisms. David Jones > > > Explain what question you're trying to answer; maybe somebody can tell > you what statistics you should be using. > > Robert Dawson > . > . > ================================================================= > Instructions for joining and leaving this list, remarks about the > problem of INAPPROPRIATE MESSAGES, and archives are available at: > . http://jse.stat.ncsu.edu/ . > ================================================================= . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
