I suppose it all comes down to the methods to measure - just like with COVD statistics.

Because one method does not come to the same conclusion others do does not automatically prove that it really is that way.