How to combine data sets to do a pearson correlation


I want to calculate the pearson correlation between 2 variables, namely the expression level of 2 genes measured in biological samples. The problem is that the sample size is small (n=5). However, I also have 2 other similar data sets obtained previously with another methods. So I would like to combine the three data sets in order to reach n=15 and get a more accurate correlation. Does anybody know how to do that ? Can I simply combine the three data sets and do the correlation across all samples? I'm worried about that because the values in each set are on different scale with different means and variances. Shall I apply some normalization.
I would appreciate any suggestion or help