It's likely that an arbitrary metric can vary among groups. This as such is not problematic. The problem with the quoted 'research' is that they are effectively

It's likely that an arbitrary metric can vary among groups. This as such is not problematic. The problem with the quoted 'research' is that they are effectively just p-mining data sets and trying to find correlations between IQ and other metrics.