I am doing differential network analysis. However; I have different sample sizes for the different conditions. I am observing that since I am using Pearson's correlation in wgcna to measure the weights- the weights of correlation depend on the sample size( where larger samples created lower weights). Since I am calculating different centrality measures individually- to compare between the networks -- I am seeing the difference in centrality I am seeing is because of the differing sample size and not the condition. Do I have to keep the sample size the same if I have to compare different network properties between different networks or conditions.
Any inputs help!