I used SPSS's cross-correlation formula to calculate the correlation coefficient. In the image 1-3, I used natrual log tranform, and there are some warning messages like the third image (FYI, the series in this reslut has no zeros.); in the image 4-5, I also did it with excluding series that with zeros. but the result is totally different. In my opinion, reslut could only be different in the series that with zeros. but truth is that the result of those series with no zero also changed. only when i exclude series with 0, the final result stays the same in those with no zeros. why and how can this happen? shouldn't the reslut stay independent among those series? 1 2 3 4 5
I wonder what happened in this prosess and which result is reliable?