Jim Logajan writes:

Unfortunately you continue to use data sets that have already been
pointed out as unreliable basis for normalization.

Which datasets are you using?